Revised 2023 and Later Model Year
LightDuty Vehicle GHG Emissions
Standards:
Regulatory Impact Analysis
SEPA
United States
Environmental Protection
Agency

-------
Revised 2023 and Later Model Year
Light-Duty Vehicle GHG Emissions
Standards i
Regulatory Impact Analysis
This technical report does not necessarily represent final EPA decisions
or positions. It is intended to present technical analysis of issues using
data that are currently available. The purpose in the release of such
reports is to facilitate the exchange of technical information and to
inform the public of technical developments.
Assessment and Standards Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
NOTICE
&EPA
United States
Environmental Protection
Agency
EPA-420-R-21-028
December 2021

-------
Table of Contents
List of Tables	v
List of Figures	x
EXECUTIVE SUMMARY	XI
Revisions to Light-duty GHG Emissions Standards	xii
Compliance Incentives and Flexibilities	xiv
Summary of Costs and Benefits	xvi
Summary of the Analysis of Alternatives to the Final Rule	xvm
Description of Alternatives	xviii
Summary of Costs and Benefits of the Alternatives	xx
Summary of the Costs and Benefits of the Final Revised Standards Compared to the Alternatives	xxii
RIA Chapter Summary	xxiv
Chapter 1: Background	xxiv
Chapter 2: Technology Feasibility, Effectiveness, Costs, and Lead-time	xxiv
Chapter 3: Economic and Other Key Inputs	xxiv
Chapter 4: Modeling GHG Compliance	xxv
Chapter 5: Projected Impacts on Emissions, Fuel Consumption, and Safety	xxv
Chapter 6: Vehicle Program Costs and Fuel Savings	xxv
Chapter 7: Non-GHG Health and Environmental Impacts	xxv
Chapter 8: Vehicle Sales, Employment, and Affordability and Equity Impacts	xxv
Chapter 9: Small Business Flexibilities	xxv
Chapter 10: Summary of Costs and Benefits	xxv
CHAPTER 1: BACKGROUND	1-1
1.1	Summary of 2012 Final Rulemaking	1-2
1.1.1	Light-duty Vehicle GHG Emissions Standards	1-2
1.1.2	Flexibilities	1-5
1.2	2016-2018 Midterm Evaluation of 2021-2025 Light-duty Vehicle GHG Standards	1-5
1.2.1	Updated EPA 2018 MTE Analysis	1-6
1.2.1.1	Updated Base Year Fleet Data:	1-7
1.2.1.2	Updated Fuel Price and Fleet Projections	1-7
1.2.1.3	Other Updates to the ALPHA Vehicle Model	1-7
1.2.1.4	Updates to the Technologies Considered and Technology Effectiveness	1-8
1.2.1.5	Updates to Cost Analysis	1-8
1.2.1.6	Updated Sensitivity Analyses	1-9
1.2.2	Comparison of Analytical Results Between the 2012 FRM and the MTE	1-9
1.3	Agency Actions, March 2017 - April 2020	 1-13
1.3.1	2017 Reconsideration of the MTE Final Determination and 2018 MTE Final Determination	1-13
1.3.2	SAFE	1-13
1.3.2.1 New GHG Compliance Flexibilities Established Under SAFE2	1-14
References for Chapter 1	1-15
CHAPTER 2:TECHNOLOGY FEASIBILITY, EFFECTIVENESS, COSTS, AND LEAD-TIME	2-1
2.1	Final Standards	2-1
2.1.1 Revised Final Compliance Incentives and Flexibilities	2-4
2.2	Light-duty Vehicle Technology Feasibility	2-4
2.2.1	Feasibility of the Revised Final Standards	2-4
2.2.2	Alternatives to the Revised Standards	2-6
1

-------
2.3	Vehicle Technologies	2-8
2.3.1	Recent Advances in Internal Combustion Engines	2-10
2.3.2	Changes to Engine Technologies Represented in the Analysis for the Final Rule	2-11
2.3.3	Vehicle Electrification	2-12
2.3.4	Automotive Li-ion Battery Costs	2-14
2.4	Analysis of Manufacturers Generation and Use of GHG Credit	2-14
References for Chapter 2	2-17
CHAPTER 3:ECONOMIC AND OTHER KEY INPUTS	3-1
3.1	Rebound	3-1
3.1.1	Accounting for the Fuel Economy Rebound Effect	3-1
3.1.2	Summary of Historical Literature on the LDV Rebound Effect	3-2
3.1.3	Review of Recent Literature on LDV Rebound.	3-5
3.1.4	Basis for Rebound Effect Used in this Final LDV Rule	3-13
3.2	Energy Security Impacts	3-17
3.2.1	Review of Historical Energy Security Literature	3-18
3.2.2	Review of Recent Energy Security Literature	3-19
3.2.3	Cost of Existing U.S. Energy Security Policies	3-22
3.2.4	U.S. Oil Import Reductions from this Final Rule	3-24
3.2.5	Oil Security Premiums Used for this Final Rule	3-26
3.2.6	Energy Security Benefits of the Final Rule	3-29
3.3	Social Cost of Greenhouse Gases	3-30
3.4	Drive Surplus, Congestion and Noise	3-46
References for Chapter 3	3-48
CHAPTER 4:MODELING GHG COMPLIANCE	4-1
4.1	Compliance Modeling, Analytical Updates, and Analytical Revisions	4-1
4.1.1	Changes made to the Model Inputs since the Proposed Rule	4-6
4.1.1.1	Off-Cycle Credit Cost and changes since the Proposed Rule	4-6
4.1.1.2	Battery Costs and Changes since the Proposed Rule	4-7
4.1.1.3	Restricting HCR2 Technology from the Available Technologies	4-11
4.1.1.4	Shifting of Input File Years due to the Updated Baseline Fleet	4-12
4.1.2	GHG Targets and Compliance Levels	4-12
4.1.2.1	Final Standards	4-12
4.1.2.2	Alternatives	4-17
4.1.3	Projected Compliance Costs per Vehicle	4-21
4.1.3.1	Final Standards	4-21
4.1.3.2	Alternatives	4-24
4.1.4	Technology Penetration Rates	4-26
4.1.4.1	Final Rule	4-26
4.1.4.2	Alternatives	4-30
4.1.4.3	Fleet Mix	4-32
4.1.5	Sensitivities	4-32
4.1.5.1	Compliance Costs per Vehicle and Technology Penetration	4-34
4.1.5.2	How Battery Costs Impact Non-BEV Vehicle Costs in our Modeling	4-36
4.2	Estimates of Fuel Economy Impacts	4-38
4.2.1	Final Rule	4-38
4.2.2	Alternatives	4-40
References for Chapter 4	4-44
CHAPTER 5:PROJECTED IMPACTS ON EMISSIONS, FUEL CONSUMPTION, AND SAFETY	5-1
ii

-------
5.1	Projected Emissions Impacts	5-1
5.1.1	Greenhouse Gas Emissions	5-1
5.1.1.1	Final Rule	5-1
5.1.1.2	Alternatives	5-4
5.1.2	Non-Greenhouse Gas Emissions	5-7
5.1.2.1	Final Rule	5-7
5.1.2.2	Alternatives	5-12
5.2	Projected Fuel Consumption	5-15
5.2.1	Final Rule	5-15
5.2.2	Alternatives	5-17
5.3	Projected Safety Impacts	5-18
References for Chapter 5	5-21
CHAPTER 6:VEHICLE COSTS, FUEL SAVINGS AND NON-EMISSION BENEFITS	6-1
6.1	Costs	6-1
6.1.1	Final Rule	6-2
6.1.2	A Iternatives	6-3
6.2	Fuel Savings	6-5
6.2.1	Final Rule	6-5
6.2.2	Alternatives	6-6
6.3	Non-Emission Benefits	6-8
6.3.1	Non-emission Benefits of the Final Rule	6-9
6.3.2	Non-emission Benefits of the Proposal and Alternative	6-10
CHAPTER 7:NON-GHG HEALTH AND ENVIRONMENTAL IMPACTS	7-1
7.1 Health and Environmental Impacts of Non-GHG Pollutants	7-1
7.1.1	Background on Non-GHG Pollutants Impacted by the Final Standards	7-1
7.1.1.1	Particulate Matter	7-1
7.1.1.2	Ozone	7-2
7.1.1.3	Nitrogen Oxides	7-2
7.1.1.4	Sulfur Oxides	7-2
7.1.1.5	Carbon Monoxide	7-3
7.1.1.6	Air Toxics	7-3
7.1.2	Health Effects Associated with Exposure to Non-GHG Pollutants	 7-3
7.1.2.1	Particulate Matter	7-3
7.1.2.2	Ozone	7-7
7.1.2.3	Nitrogen Oxides	7-8
7.1.2.4	Sulfur Oxides	7-9
7.1.2.5	Carbon Monoxide	7-10
7.1.2.6	Air Toxics	7-11
7.1.2.6.1	Health Effects Associated with Exposure to Benzene	7-11
7.1.2.6.2	Health Effects Associated with Exposure to Formaldehyde	7-12
7.1.2.6.3	Health Effects Associated with Exposure to Acetaldehyde	7-12
7.1.2.6.4	Health Effects Associated with Exposure to Naphthalene	7-13
7.1.2.6.5	Health Effects Associated with Exposure to 1,3-Butadiene	7-13
7.1.2.6.6	Health effects Associated with exposure to other air toxics	7-14
7.1.2.7	Exposure and Health Effects Associated with Traffic	7-14
7.1.3	Environmental Effects Associated with Exposure to Non-GHG Pollutants	7-16
7.1.3.1	Visibility	7-16
7.1.3.2	Ozone Effects on Ecosystems	7-17
7.1.3.3	Deposition	7-18
7.1.3.4	Environmental Effects of Air Toxics	7-19
111

-------
7.2 Non-GHG Monetized Health Benefits	7-19
7.2.1 Uncertainty	 7-28
References for Chapter 7	7-31
CHAPTER 8:VEHICLE SALES, EMPLOYMENT, ENVIRONMENTAL JUSTICE, AND
AFFORD ABILITY AND EQUITY IMPACTS	8-1
8.1	Sales Impacts	8-1
8.1.1	Conceptual Framework	8-1
8.1.1.1	Existence of the Energy Efficiency Gap	8-1
8.1.1.2	Potential Explanations for the Existence of the Energy Efficiency Gap	8-4
8.1.2	How Sales Impacts were Modeled	8-7
8.1.3	Sales Impacts	8-10
8.2	Employment Impacts	8-12
8.2.1	Conceptual Framework	8-12
8.2.2	How Employment Impacts were Modeled	8-14
8.2.3	Employment Impacts	8-14
8.3	Environmental Justice	8-16
8.3.1	GHG Impacts	8-17
8.3.1.1 Effects on Specific Populations of Concern	8-18
8.3.2	Non-GHG Impacts	8-20
8.4	Affordability and Equity Impacts	8-21
8.4.1	Effects on Lower-Income Households	8-23
8.4.2	Effects on the Used Vehicle Market	8-26
8.4.3	Effects on Access to Credit	8-27
8.4.4	Effects on Low-Priced Cars	8-28
8.4.5	Effects of Electric Vehicles on Affordability	8-29
8.4.6	Summary of Affordability and Equity Effects	8-29
References for Chapter 8	8-30
CHAPTER 9: SMALL BUSINESS FLEXIBILITIES	9-1
References for Chapter 9	9-3
CHAPTER 10: SUMMARY OF COSTS AND BENEFITS	10-1
10.1	Final Rule	10-1
10.2	Comparison to Proposal (Less Stringent Alternative)	10-4
10.3	Comparison to the More Stringent Alternative 2 Minus 10	10-8
10.4	Sensitivities	10-12
iv

-------
List of Tables
Table 1: Projected Industry Fleet-wide C02 Compliance Targets (grams/mi)*	xm
Table 2: Projected Industry Fleet Average Target Year-Over-Year Percent Reductions	xiv
Table 3: Final EPA's Extension of Credit Carry-forward Provisions	xv
Table 4: Monetized Discounted Costs, Benefits, and Net Benefits of the Final Program for Calendar
Years through 2050 (Billions of 2018 dollars)AB C,D'E	xvii
Table 5 Car, Light Truck and Fleet Average Cost per Vehicle Relative to the No Action Scenario
(2018 DOLLARS) 	XVII
Table 6 GHG Reductions Through 2050	xvii
Table 7: Applicability of Program Provisions to the Final Standards, and the Proposal and
Alternative 2 minus 10 Standards	xvm
Table 8: Projected Fleet Average Target Levels for Final Standards and Alternatives (CO2
grams/mile) *	XIX
Table 9: Monetized Discounted Costs, Benefits, and Net Benefits of the Proposal Standards for
Calendar Years through 2050 (Billions of 2018 dollars)AAC>D'E	xxi
Table 10: Monetized Discounted Costs, Benefits, and Net Benefits of Alternative 2 minus 10 for
Calendar Years through 2050 (Billions of 2018 dollars)AAC-D-E	xxii
Table 11: Present Value Monetized Discounted Costs, Benefits, and Net Benefits of the Final Program
and Alternatives for Calendar Years through 2050 (Billions of 2018 dollars)AAC'D'E	xxm
Table 12: Annualized Monetized Discounted Costs, Benefits, and Net Benefits of the Final Program
and Alternatives for Calendar Years through 2050 (Billions of 2018 dollars)a,b,c,d,e	xxiv
Table 1-1: Projected Fleet-Wide Emissions Compliance Targets under the Footprint-Based CO2
Standards in the 2012 FRM	1-3
Table 1-2: Projected vs. Actual Car/Truck Sales Share, 2016-2019 Model Years	1-3
Table 1-3: Comparison of technology penetrations into the light-duty fleet and per vehicle costs in
2025 (INCREMENTAL TO 2021) FOR THE 2012 FRM COMPARED TO ANALYSES CONDUCTED BY EPA UNDER THE
MTE. All per vehicle costs are shown in 2018$ to maintain consistency with other analyses
WITHIN THIS RIA	 1-11
Table 1-4: Comparison of fuel price, percentage of cars and trucks in the fleet, and C02 fleet average
EMISSIONS TARGETS WHEN TAKING INTO CONSIDERATION THE CAR AND TRUCK FLEET MIX FOR THE 2012 FRM
COMPARED TO ANALYSES CONDUCTED BY EPA UNDER THE MTE	 1-12
Table 1-5: Comparison per vehicles costs for passenger cars, light-duty trucks and the combined
LIGHT-DUTY VEHICLE FLEET IN 2025 (INCREMENTAL TO 2021) FOR THE 2012 FRM COMPARED TO ANALYSES
CONDUCTED BY EPA UNDER THE MTE. PER VEHICLE COSTS ARE SHOWN IN 2018$ TO MAINTAIN CONSISTENCY
WITH OTHER ANALYSES WITHIN THIS RIA	 1-12
Table 2-1: Final Footprint-based C02 Standard Curve Coefficients	2-1
Table 2-2: Estimated Fleet-wide C02 Target Levels Corresponding to the Final Standards	2-4
Table 2-3: Projected Fleet Average Target Levels for Revised Standards and Alternatives (C02
grams/mile)*	2-7
Table 2-4: Production Share by Engine Technologies for MY 2015-2020	 2-9
Table 2-5: Production Share by Transmission Technologies for MYs 2015-2020	 2-9
Table 2-6: Distribution of 2021 MY Vehicle Models and Number of Vehicles Which Generate Credits
vs. 2023 MY Standards (All Vehicles)	2-15
Table 2-7: Distribution of 2021 MY Vehicle Models and Number of Vehicles Which Generate Credits
vs. 2023 MY Standards (Gasoline ICE and Hybrid Vehicles)	2-16
Table 2-8: Distribution of 2021 MY Vehicle Models and Number of Vehicles Which Generate Credits
vs. 2022 MY Standards (All Vehicles)	2-16
Table 3-1: Estimates of the Rebound Effect Using U.S. Aggregate Time-Series Data on Vehicle Travel
	3-2
V

-------
Table 3-2: Estimates of the Rebound Effect Using U.S./State and Canadian/Province Level Data	3-3
Table 3-3: Estimates of the Rebound Effect Using U.S. Household Survey Data	3-3
Table 3-4: Studies Given Significant Weight in Developing an Estimate of the VMT Rebound Effect for
this Final Rule	3-15
Table 3-5: Projected Trends in U.S. Oil Exports/Imports, Net Oil Product Exports, Net Crude
Oil/Product Exports, Oil Consumption and U.S. Oil Import Reductions Resulting from the Final
LDV GHG rule from 2023 to 2050 (Millions of barrels per day (MMBD))*	3-26
Table 3-6: Macroeconomic Oil Security Premiums for Selected Years from 2023-2050 (2018$/Barrel)*
	3-29
Table 3-7: Annual Energy Security Benefits of the Final LDV GHG/Fuel Economy Rule for Selected
Years 2023-2050 (in Billions of 2018$)	3-29
Table 3-8: Interim Global Social Cost of Carbon Values, 2020-2070 (2018$/Metric Tonne C02)99	 3-34
Table 3-9: Interim Global Social Cost of Methane Values, 2020-2070 (2018$/Metric Tonne CH4)99.... 3-35
Table 3-10: Interim Global Social Cost of Nitrous Oxide Values, 2020-2070 (2018$/Metric Tonne N20)99
	3-35
Table 3-11: Estimated Global Climate Benefits from Changes in CO2 Emissions 2023 - 2050 for the Final
Rule (Billions of 2018$)	3-40
Table 3-12: Estimated Global Climate Benefits from Changes in CH4 Emissions 2023 - 2050 for the Final
Rule (Billions of 2018$)	3-41
Table 3-13: Estimated Global Climate Benefits from Changes in N20 Emissions 2023 - 2050 (Billions of
2018$)	3-42
Table 3-14: Estimated Global Climate Benefits from Changes in GHG Emissions 2023 - 2050 (Billions of
2018$)	3-43
Table 3-15: Estimated Global Climate Benefits from Changes in GHG Emissions 2023-2050 for the
Proposal Standards (Billions of 2018$)	3-44
Table 3-16: Estimated Global Climate Benefits from Changes in GHG Emissions 2023 - 2050 for
Alternative 2 minus 10 (Billions of 2018$)	3-45
Table 3-17: Costs Associated with Congestion and Noise (2018 dollars per vehicle mile)	3-47
Table 4-1: Changes made to SAFE FRM CCEMS Inputs for NPRM Model Runs	4-3
Table 4-2: Changes made to EPA NPRM CCEMS Inputs for Final Rule Model Runs	4-4
Table 4-3 Vehicle Sales Available for Refresh and Redesign	4-5
Table 4-4: Cost per Off-Cycle Credit used in the NPRM (2018 dollars)	4-6
Table 4-5: Cost per Vehicle relative to No Action at different levels of Zero-Cost Off-cycle Credit
(2018 DOLLARS) 	 4-6
Table 4-6: Incremental Off-cycle Credit Cost ($/gram) for Different Levels of Off-cycle Credit (2018
DOLLARS) 	 4-6
Table 4-7. Cost Changes for a 60 kWh BEV Battery	4-9
Table 4-8: Car Targets (C02 gram/mile)	4-14
Table 4-9: Light Truck Targets (CO2 gram/mile)	4-14
Table 4-10: Sales Weighted Fleet Targets (CO2 gram/mile)	4-15
Table 4-11: Car Achieved (CO2E gram/mile)	4-15
Table 4-12: Light Truck Achieved (CO2E gram/mile)	4-16
Table 4-13: Sales Weighted Fleet Achieved (C02e gram/mile)	4-16
Table 4-14: Car Targets under the Proposal and Alternative 2 minus 10 Standards (C02 gram/mile) .. 4-
18
Table 4-15: Light Truck Targets under the Proposal and Alternative 2 minus 10 Standards (C02
gram/mile)	4-18
Table 4-16: Fleet Targets under the Proposal and Alternative 2 minus 10 Standards (CO2 gram/mile) 4-
19
vi

-------
Table 4-17: Car Targets Achieved under the Proposal and Alternative 2 minus 10 Standards (CO2E
gram/mile)	4-19
Table 4-18: Light Truck Targets Achieved under the Proposal and Alternative 2 minus 10 Standards
(CO2E gram/mile)	4-20
Table 4-19: Fleet Targets Achieved under the Proposal and Alternative 2 minus 10 Standards (CO2E
gram/mile)	4-20
Table 4-20: Car Costs/Vehicle Relative to the No Action Scenario (2018 dollars)	4-21
Table 4-21: Light Truck Cost per Vehicle Relative to the No Action Scenario (2018 dollars)	4-21
Table 4-22: Fleet Average Cost per Vehicle Relative to the No Action Scenario (2018 dollars)	4-22
Table 4-23 Costs per Vehicle Projected through 2050 for the Final Standards (2018 dollars per
vehicle)	4-24
Table 4-24: Car Average Cost per Vehicle for the Proposal and Alternative 2 minus 10 Standards
Relative to the No Action Scenario (2018 dollars)	4-25
Table 4-25: Light Truck Average Cost per Vehicle for the Proposal and Alternative 2 minus 10
Standards Relative to the No Action Scenario (2018 dollars)	4-25
Table 4-26: Fleet Average Cost per Vehicle for the Proposal and Alternative 2 minus 10 Standards
Relative to the No Action Scenario (2018 dollars)	4-26
Table 4-27 BEV+PHEV Penetration Rates under the No Action Scenario	4-27
Table 4-28: Car BEV+PHEV Penetration Rates under the Final Standards	4-27
Table 4-29: Light Truck BEV+PHEV Penetration Rates under the Final Standards	4-28
Table 4-30: Fleet BEV+PHEV Penetration Rates under the Final Standards	4-28
Table 4-31: Fleet ICE Technology Penetration Rates under the Final Standards	4-29
Table 4-32: Impact of Advanced Technology Multipliers on the Penetration of BEV and PHEV
Technology	4-30
Table 4-33: Car BEV+PHEV Penetration Rates under the Proposal and Alternative 2 minus 10
Standards	4-30
Table 4-34: Light Truck BEV+PHEV Penetration Rates under the Proposal and Alternative 2 minus 10
Standards	4-31
Table 4-35: Fleet BEV+PHEV Penetration Rates under the Proposal and Alternative 2 minus 10
Standards	4-31
Table 4-36 Fleet Mix Projections for the Final Standards, Proposal and Alternative 2 minus 10	4-32
Table 4-37: Costs per Vehicle for the Final Standards and Sensitivities relative to their No Action
Scenarios (2018 dollars)*	4-34
Table 4-38: MY 2026 Technology Penetration Rates for the No-Action and Final Standards in the AEO
High, AEO Low, Allow HCR2 and No Hybrids Sensitivities	4-35
Table 4-39 MY 2026 Technology Penetration Rates for the No-Action and Final Standards in the
Battery Costs Higher and Lower, Demand Elasticity of -0.15 and -1.0 and the Perfect Trading
Sensitivities	4-36
Table 4-40 Cost Contributions Comparing our Primary Battery Costs to the Higher Battery Costs
(2018 DOLLARS PER VEHICLE)	 4-37
Table 4-41 Cost Contributions Comparing our Primary Battery Costs to the Lower Battery Costs
(2018 DOLLARS PER VEHICLE)	 4-37
Table 4-42: Fuel Economy (MPG) Estimates based on the GHG Standards*	4-39
Table 4-43: Fuel Economy (MPG) Estimates assuming full use of AC Leakage Credits*	4-39
Table 4-44: Fuel Economy (MPG) Estimated "Label Value"*	4-40
Table 4-45: Fuel Economy (MPG) Estimates based on the GHG Standards for the Proposal Standards*
	4-40
Table 4-46: Fuel Economy (MPG) Estimates Assuming Full Use of AC Leakage Credits for the Proposal
Standards*	4-41
Table 4-47: Fuel Economy (MPG) Estimated "Label Value" Under the Proposal Standards*	4-41
vii

-------
Table 4-48: Fuel Economy (MPG) Estimates Based on the GHG Standards of the Alternative 2 minus 10
Standards*	4-42
Table 4-49: Fuel Economy (MPG) Estimates Assuming Full Use of AC Leakage Credits in the
Alternative 2 minus 10 Standards*	4-42
Table 4-50: Fuel Economy (MPG) Estimated "Label Value" Under the Alternative 2 minus 10
Standards*	4-43
Table 5-1: Impacts on GHG Emissions under the Final Standards Relative to the No Action Scenario5-3
Table 5-2: Impacts on GHG Emissions under the Proposal Standards Relative to the No Action
Scenario	5-5
Table 5-3: Impacts on GHG Emissions under the Alternative 2 minus 10 Standards Relative to the No
Action Scenario	5-6
Table 5-4: Impacts on Upstream Non-GHG Emissions Under the Final Standards Relative to the No
Action Scenario	5-10
Table 5-5: Estimated Non-GHG Emission Impacts of the Final Standards Relative to the No Action
Scenario	5-11
Table 5-6: Impacts on Upstream Non-GHG Emissions under the Proposal Standards Relative to the No
Action Scenario	5-12
Table 5-7: Impacts on Non-GHG Emissions under the Proposal Standards Relative to the No Action
Scenario	5-13
Table 5-8: Impacts on Upstream Non-GHG Emissions under the Alternative 2 minus 10 Standards
Relative to the No Action Scenario	5-14
Table 5-9: Impacts on Non-GHG Emissions under the Alternative 2 minus 10 Standards Relative to the
No Action Scenario	5-15
Table 5-10: Impacts on Fuel Consumption for the Final Standards Relative to the No Action Scenario
	5-16
Table 5-11: Impacts on Fuel Consumption for the Proposal Standards Relative to the No Action
Scenario	5-17
Table 5-12: Impacts on Fuel Consumption for Alternative 2 minus 10 Relative to the No Action
Scenario	5-18
Table 6-1: Costs Associated with the Final Program Relative to the No Action Scenario ($Billions of
2018 DOLLARS)	 6-2
Table 6-2: Costs Associated with the Proposal Relative to the No Action Scenario ($Billions of 2018
DOLLARS) 	 6-3
Table 6-3: Costs Associated with Alternative 2 minus 10 Relative to the No Action Scenario ($Billions
OF 2018 DOLLARS)	 6-4
Table 6-4: Fuel Savings Associated with the Final Program ($Billions of 2018 dollars)	6-5
Table 6-5: Fuel Savings Associated with the Proposal ($Billions of 2018 dollars)	6-6
Table 6-6: Fuel Savings Associated with Alternative 2 minus 10 ($Billions of 2018 dollars)	6-7
Table 6-7: CCEMS Inputs used to Estimate Refueling Time Costs	6-8
Table 6-8: CCEMS Inputs used to Estimate Electric Refueling Time Costs	6-8
Table 6-9: Benefits from Non-Emission Sources under the Final Rule ($Billions of 2018 dollars)	6-9
Table 6-10: Benefits from Non-Emission Sources Under the Proposal ($Billions of 2018 dollars) .... 6-10
Table 6-11: Benefits from Non-Emission Sources under Alternative 2 minus 10 ($Billions of 2018
DOLLARS) 	 6-11
Table 7-1: Health Effects of Ambient PM2 5	7-21
Table 7-2: Health Effects of Ambient Ozone	7-22
Table 7-3: Additional Unquantified Health and Welfare Benefits Categories	7-23
Table 7-4: PM-related Benefit-per-ton Values (2018$)a	7-25
Table 7-5: Undiscounted Stream, Present and Annualized Value of PM2.5-Related Benefits from 2023
THROUGH 2050 FOR THE FINAL RULE (DISCOUNTED AT 3 PERCENT AND 7 PERCENT; $BlLLIONS OF 2018$)A. 7-26
viii

-------
Table 7-6: Undiscounted Stream, Present and Annualized Value of PM25-related Benefits from 2023
THROUGH 2050 FOR THE PROPOSAL (DISCOUNTED AT 3 PERCENT AND 7 PERCENT; $BlLLIONS OF 2018$)A.... 7-27
Table 7-7: Undiscounted Stream, Present and Annualized Value of PM25-related Benefits from 2023
THROUGH 2050 FOR ALTERNATIVE 2 MINUS 10 (DISCOUNTED AT 3 PERCENT AND 7 PERCENT; $BlLLIONS OF
2018$)A	7-28
Table 8-1: Policy Elasticities Corresponding to Selected Demand and Scrappage Elasticities	8-9
Table 8-2: Sales Impacts, 2.5 Years of Fuel Savings in Net Price, Demand Elasticity -0.4	8-10
Table 8-3: Sales Impacts, 2.5 Years of Fuel Savings in Net Price, Demand Elasticity -0.15	8-11
Table 8-4: Sales Impacts, 2.5 Years of Fuel Savings in Net Price, Demand Elasticity -1	8-11
Table 8-5: Employment Impacts, Based on Sales Estimates in Table 8-2 (Demand Elasticity -0.4)	8-14
Table 8-6: Employment Impacts, Based on Sales Estimates in Table 8-3 (Demand Elasticity -0.15)	8-15
Table 8-7: Employment Impacts, Based on Sales Estimates in Table 8-4 (Demand Elasticity -1)	8-15
Table 9-1: Primary Vehicle SBA Small Business Categories	9-1
Table 9-2: Small Business Entities	9-2
Table 10-1: Costs Associated with the Final Program ($Billions of 2018 dollars)	10-1
Table 10-2: Fuel Savings Associated with the Final Program ($Billions of 2018 dollars)	10-2
Table 10-3: Benefits from Non-Emission Sources for the Final Rule ($Billions of 2018 dollars)	10-2
Table 10-4: PM25-related Emission Reduction Benefits of the Final Rule ($Billions of 2018 dollars) 10-
3
Table 10-5: Climate Benefits from Reduction in GHG Emissions ($Billions of 2018 dollars)	10-3
Table 10-6: Net Benefits (Emission Benefits + Non-Emission Benefits + Fuel Savings - Costs) Associated
with the Final Program ($Billions of 2018 dollars)	10-4
Table 10-7: Costs Associated with the Proposal ($Billions of 2018 dollars)	10-5
Table 10-8: Fuel Savings Associated with the Proposal ($Billions of 2018 dollars)	10-5
Table 10-9: Benefits from Non-Emission Sources Associated with the Proposal ($Billions of 2018
DOLLARS) 	 10-6
Table 10-10: PM25-related Emission Reduction Benefits Associated with the Proposal ($Billions of
2018 DOLLARS)	 10-6
Table 10-11: Climate Benefits from Reduction in Greenhouse Gas Emissions Associated with the
Proposal ($Billions of 2018 dollars)	10-7
Table 10-12: Net Benefits (Emission Benefits + Non-Emission Benefits + Fuel Savings - Costs) for the
Proposal ($Billions of 2018 dollars)	10-8
Table 10-13: Costs Associated with the Alternative 2 minus 10 ($Billions of 2018 dollars)	10-9
Table 10-14: Fuel Savings Associated with Alternative 2 minus 10 ($Billions of 2018 dollars)	10-9
Table 10-15: Benefits from Non-Emission Sources Associated with Alternative 2 minus 10 ($Billions of
2018 DOLLARS)	 10-10
Table 10-16: PM2 5-related Emission Reduction Benefits Associated with Alternative 2 minus 10
($BlLLIONS OF 2018 DOLLARS)	 10-10
Table 10-17: Climate Benefits from Reduction in Greenhouse Gas Emissions Associated with
Alternative 2 minus 10 ($Billions of 2018 dollars)	10-11
Table 10-18: Net Benefits (Emission Benefits + Non-Emission Benefits + Fuel Savings - Costs)
Associated with Alternative 2 minus 10 ($Billions of 2018 dollars)	10-12
Table 10-19 Monetized Discounted Costs, Benefits, and Net Benefits of the Proposed Program and
each Sensitivity for Calendar Years through 2050 ($Billions of 2018 dollars, 3 percent Discount
Rate)abcd	10-13
Table 10-20 Monetized Discounted Costs, Benefits, and Net Benefits of the Proposed Program and
each Sensitivity for Calendar Years through 2050 ($Billions of 2018 dollars, 7 percent Discount
Rate)abcd	10-14
IX

-------
List of Figures
Figure 1: Final Fleet-Wide C02-Equivalent g/mi Compliance Targets (solid black line), Compared to
2012 FRM, SAFERule, and Proposal	xm
Figure 2: Final Standards Fleet Average Targets Compared to Alternatives	xix
Figure 1-1: Regulatory Timeline for the Model Year (MY) 2017 and Later Light-duty Vehicle (LDV)
Greenhouse Gas (GHG) Emissions Standards, the Midterm Evaluation, and Safe Rule. The top
ROW REPRESENTS AGENCY ACTIONS THAT USED EPA ANALYSES AS THE BASIS	 1-2
Figure 1-2:2012 FRM Footprint Curves for Passenger Car CO2 (g/mile) Standards	1-4
Figure 1-3:2012 FRM Footprint Curves for Light-duty Truck CO2 (g/mile) Standards	1-4
Figure 1-4: Comparison of fleet average (car and truck), per-vehicle technology costs in 2025 from
THE 2012 FRM TO SUBSEQUENT ANALYSES CONDUCTED BY EPA DURING THE MTE (2018$). VERTICAL LINES
ON TOP OF THE BARS REPRESENT THE RANGE OF SENSITIVITY ANALYSES CONDUCTED	 1-10
Figure 2-1: CO2 vs Footprint Compliance Curves for Cars	2-2
Figure 2-2: CO2 vs Footprint Compliance Curves for Trucks	2-2
Figure 2-3: Final Fleet-Wide C02-Equivalent g/mi Compliance Targets (solid black line), Compared to
2012 FRM, SAFE Rule, and Proposal	2-3
Figure 2-4: Final Rule Fleet Average Targets Compared to the Proposal and Alternative 2 minus 102-7
Figure 3-1: Frequency Distribution of SC-C02 Estimates for 2030	 3-36
Figure 3-2: Frequency Distribution of SC-CH4 Estimates for 2030l	3-37
Figure 3-3: Frequency Distribution of SC-N20 Estimates for 2030l	3-37
Figure 4-1: Final Fleet-Wide C02-Equivalent g/mi Compliance Targets (solid black line), Compared to
2012 FRM, SAFE Rule, and Proposal	4-13
Figure 4-2: Final Rule Fleet Average Targets Compared to the Proposal and Alternative 2 minus 10 4-
17
Figure 4-3. Battery cost sensitivity cases in terms of $/kWh DMC for a representative 60 kWh battery
	4-33
Figure 4-4 Gasoline and Electricity Prices used in the Primary and Sensitivity Analyses	4-34
Figure 5-1 Electricity Generating Unit (EGU) and Refinery Emission Factors for CO2	5-2
Figure 5-2 Electricity Generating Unit (EGU) and Refinery Emission Factors for CH4	5-2
Figure 5-3 Electricity Generating Unit (EGU) and Refinery Emission Factors for N20	5-2
Figure 5-4 Cumulative CO2 Reductions relative to the No Action scenario (Million Metric Tons CO2) 5-
7
Figure 5-5 Electricity Generating Unit (EGU) and Refinery Emission Factors for NOx	5-8
Figure 5-6 Electricity Generating Unit (EGU) and Refinery Emission Factors for PM2 5	5-8
Figure 5-7 Electricity Generating Unit (EGU) and Refinery Emission Factors for SOx	5-8
Figure 5-8 Electricity Generating Unit (EGU) and Refinery Emission Factors for VOC	5-9
Figure 5-9 Tailpipe Emission Factors for MYs 2023,2026 and 2035 for NOx and PM2 5	5-9
x

-------
Executive Summary
This Regulatory Impact Analysis (RIA) contains supporting documentation to the
Environmental Protection Agency (EPA) final rulemaking and addresses requirements in Clean
Air Act Section 317. The preamble to the Federal Register notice associated with this document
provides the full context for the EPA final rule, and it references this RIA throughout.
The EPA is establishing revised, more stringent national greenhouse gas (GHG) emissions
standards for passenger cars and light trucks under section 202(a) of the Clean Air Act (CAA).
Section 202(a) requires EPA to establish standards for emissions of air pollutants from new
motor vehicles which, in the Administrator's judgment, cause or contribute to air pollution which
may reasonably be anticipated to endanger public health or welfare. 42 U.S.C. 7521(a).
This program further revises the light-duty vehicle GHG standards previously revised by the
SAFE rule and builds upon earlier EPA actions and supporting analyses that established or
maintained stringent light-duty vehicle GHG emissions standards. For example, in 2012, EPA
issued a final rule establishing light-duty vehicle GHG standards for model years (MY) 2017-
2025,a which were supported in analyses accounting for compliance costs, lead time and other
relevant factors.b That rule and its analyses also accounted for the development and availability
of advanced GHG emission-reducing technologies for gasoline-fueled vehicles, which
demonstrated that the standards were appropriate under section 202(a) of the CAA.C This final
rule relies upon additional analysis that consider updated data and recent developments. Auto
manufacturers are currently implementing an increasing array of advanced gasoline vehicle GHG
emission reduction technologies at a rapid pace throughout their vehicle fleets. Vehicle
electrification technologies are also advancing rapidly, as battery costs have continued to decline,
and automakers have announced an increasing diversity and volume of zero-emission vehicle
models. Additionally, in 2019, several auto manufacturers voluntarily entered into agreements
with the State of California to comply with GHG emission reduction targets through MY 2026
across their national vehicle fleets (the "California Framework Agreements") that are more
stringent than the previous EPA standards as revised by the SAFE rule. These developments
further supported EPA's decision to reconsider and revise the previous EPA standards and to
establish more stringent standards, particularly in light of factors indicating that more stringent
near-term standards were feasible at reasonable cost and would achieve significantly greater
GHG emissions reductions and public health and welfare benefits than the previous program.
EPA has conducted outreach with a wide range of interested stakeholders, including labor
unions, states, and industry as provided in E.O. 13990, as part of our regulatory development
process for the revised light-duty GHG emissions standards.
a EPA's model year emission standards also apply in subsequent model years, unless revised, e.g., MY 2025
standards issued in the 2012 rule also applied to MY 2026 and beyond.
b 77 FR 62624, October 15, 2012.
c Id.
xi

-------
The revisions to the standards are limited to MYs 2023-2026, consistent with lead time
considerations under the CAA.d We designed the program based on our assessment that the
revised standards are reasonable and appropriate and will achieve a significant level of GHG
reductions for MYs 2023-2026 vehicles, with the expectation that a future, longer-term program
for MYs 2027 and later will build upon these near-term standards.
Revisions to Light-duty GHG Emissions Standards
As with EPA's previous light-duty GHG programs, EPA has finalized footprint-based
standards curves for both passenger cars and trucks. Each manufacturer has a unique standard for
the passenger cars category and another for the truck category6 for each MY based on the sales-
weighted footprint-based CO2 targetsf of the vehicles produced in that MY. Figure 1 shows
EPA's revised standards, expressed as average fleetwide GHG emissions targets (cars and trucks
combined), projected through MY 2026. For comparison, the figure also shows the
corresponding targets for the SAFE final rulemaking (FRM) and the 2012 FRM. The final fleet
targets pick up from the existing SAFE FRM targets for model years 2021 and 2022, but then
ramp down considerably in model year 2023, nearly reaching the 2012 FRM targets for that
model year. The final fleet targets approximately parallel the 2012 FRM targets for model years
2023 and 2024, are approximately equivalent to the 2012 FRM target in model year 2025 (the
last year of stringency increases in the 2012 FRM), and then decrease at a more stringent year-
over-year downward slope for one additional model year, to model year 2026 (which is also the
last year of stringency increases in the SAFE FRM). As with all EPA light-duty GHG rules, the
targets in the last year of stringency increases would then remain at the same level for all
subsequent model years unless changed by a subsequent rulemaking. Figure 1 and Table 1
present the estimates of EPA's final revised standards, again in terms of the projected overall
industry fleetwide CCh-equivalent emission compliance target levels. The industry fleet-wide
estimates in Table 1 are projections based on modeling EPA conducted for the final rule, taking
into consideration projected fleet mix and footprints for each manufacturer's fleet in each model
year. Figure 1 and Table 2 present the projected industry fleet average year-over-year percent
reductions (and cumulative reductions from 2022 through 2026) comparing the previous
standards under the SAFE rule and the final, revised standards. See Chapter 2 for a full
discussion of the revised standards.
d Note that while only the 2023-2026 Light-duty GHG Standards are revised, as with all EPA vehicle emissions
standards, the MY 2026 standards will remain in place for all subsequent MYs, unless and until the standards for
future MYs are revised in a subsequent rulemaking.
e Passenger cars include cars and smaller cross-overs and SUVs, while the truck category includes larger cross-overs
and SUVs, minivans, and pickup trucks.
f Because compliance is based on the full range of vehicles in a manufacturer's car and truck fleets, with lower-
emitting vehicles compensating for higher-emitting vehicles, the emission levels of specific vehicles within the fleet
are referred to as targets, rather than standards.
Xll

-------
260
240
220
jy
£
JiS
rsi
o
u

• • • SAFE FRM

— •2012 FRM

•Proposal
N_ X* • • - .
^^^Final Standards
• • •
• •
• •
200
180
160
140
2020 2021 2022 2023 2024 2025 2026 2027
Model Year
Table 1: Projected Industry Fleet-wide CO2 Compliance Targets (grams/mi)*
Model Year
Cars
Light Trucks
Fleet

CO2 (g/mile)
CO2 (g/mile)
CO2 (g/mile)
2022 (SAFE reference)
181
261
224
2023
166
234
202
2024
158
222
192
2025
149
207
179
2026 and later
132
187
161
Total change 2022-2026
-49
-74
-63
Percent change 2022-2026
27.1%
28.4%
28.1%
*The combined car/truck CO2 targets are a function of projected car/light truck shares, which have been
updated for this final rule (MY 2020 is 44 percent car and 56 percent light trucks while the projected
mix changes to 47 percent cars and 53 percent light trucks by MY 2026).
Xlll

-------
Table 2: Projected Industry Fleet Average Target Year-Over-Year Percent Reductions

SAFE Rule
Proposed Rule
Final Rule

Cars
Light
Trucks
Combined
Cars
Light
Trucks
Combined
Cars
Light
Trucks
Combined
2023
1.7%
1.7%
2.1%
8.4%
10.4%
9.8%
8.4%
10.4%
9.8%
2024
0.6%
1.5%
1.4%
4.7%
5.0%
5.1%
4.8%
4.9%
5.1%
2025
2.3%
1.7%
2.2%
4.8%
5.0%
5.0%
5.7%
7.0%
6.6%
2026
1.8%
1.6%
1.9%
4.8%
5.0%
5.0%
11.4%
9.5%
10.3%
Cumulative
6.3%
6.3%
7.4%
20.9%
23.1%
22.8%
27.1%
28.3%
28.3%
* Note the percentages shown for the SAFE rule targets have changed slightly from the proposed rule, due to the
updates in our base year fleet from MY 2017 to MY 2020 manufacturer fleet data.
** These are modeled results based on projected fleet characteristics and represent percent reductions in projected
targets, not the standards (which are the footprint car/truck curves), associated with that projected fleet (see Section
III for more detail on our modeling results).
Compliance Incentives and Flexibilities
The existing Light-duty GHG program established in the 2010 and 2012 rules includes several
key flexibilities, such as credit programs and technology incentives, including:
•	Credit Averaging, Banking, and Trading (ABT) with credit carry-forward, credit
carry-back, transferring of credits between a manufacturer's car and truck fleets, and
credit trading between manufacturers (see Chapter 2.1.1)
•	Off-cycle credits for GHG emissions reductions not captured by the test procedures
used for fleet average compliance with the footprint-based standards
•	Air conditioning credits for system efficiency improvements and reduced refrigerant
leakage or use of low global warming potential refrigerants
•	Multiplier incentives for advanced technology vehicles including electric vehicles,
fuel cell vehicles, and plug-in hybrid-electric vehicles
•	Full-size pick-up incentives for hybridization or GHG improvements equivalent to
hybridization
EPA has finalized a limited, targeted set of extended or additional compliance flexibilities and
incentives that we believe are appropriate given the stringency and lead time of the revised
standards. There are four types of flexibilities/incentives, in addition to flexibilities/incentives
that are already available for these MYs and that are carried over from EPA's previous
regulations:
1.	A limited extension of carry-forward credits generated in MYs 2017 and 2018;
2.	An extension of the advanced technology vehicle multiplier credits for MYs 2023 and
2024 with a cumulative credit cap;
3.	Restoration of the 2012 rule's full-size pickup truck incentives for strong hybrids or
similar performance-based credit for MYs 2023 and 2024 (provisions which were
removed in the SAFE rule); and
4.	An increase of the off-cycle credits menu cap from 10 g/mile to 15 g/mile for MYs 2023
through 2026.
We summarize these flexibilities and incentives below and provide further detail within
Sections I.B.2 and II.B.4 of the Preamble to this final rule.
xiv

-------
The GHG program includes existing provisions initially established in the 2010 rule, which
set the MY 2012-2016 GHG standards, for how credits may be used within the program. These
averaging, banking, and trading (ABT) provisions include credit carry-forward, credit carry-back
(also called deficit carry-forward), credit transfers (within a manufacturer), and credit trading
(across manufacturers). These ABT provisions define how credits may be used and are integral
to the program. The previous SAFE program limited credit carry-forward to 5 years. EPA has
revised this to include a limited extension of credit carry-forward for credits generated in MYs
2017 and 2018. This revision changes the credit carry-forward time limitation for those MY
credits from five to six years as shown in Table 3. For all other credits generated in MY 2016
and later, credit carry-forward remains unchanged at five years.
Table 3: Final EPA's Extension of Credit Carry-forward Provisions
MY in
which
Credits
are
Banked
MYs Credits Are Valid Under EPA's Proposed and Final Extension

2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2016

X
X
X
X
X




2017


X
X
X
X
X
+



2018



X
X
X
X
X
+


2019




X
X
X
X
X


2020





X
X
X
X
X

2021






X
X
X
X
X
Notes:
x = Current program.
= Additional years proposed but not finalized.
+ = Additional years finalized.
The previous GHG program also included temporary incentives through MY 2021 that
encouraged the use of advanced technologies such as electric, hybrid, and fuel cell vehicles, as
well as incentives for full-size pickups using strong hybridization or technologies providing
similar emissions reductions to hybrid technology. The full-size pickup incentives were
originally available through MY 2025, but the SAFE rule removed these incentives for MYs
2022 through 2025. When EPA established these incentives in the 2012 rule, we recognized that
they would reduce the effective stringency of the standards. However, we believed that it was
worthwhile to have a limited near-term loss of emissions reduction benefits to increase the
potential for far greater emissions reduction and technology diffusion benefits in the longer term.
Our rationale was that the temporary regulatory incentives would help bring low emission
technologies to market more quickly than an efficient market would in the absence of
incentives.8 With these same goals in mind for this program, we revised the multiplier incentives
from MY 2023 and 2024 with a cap on multiplier credits and reinstated full-size pickup
incentives for MYs 2023 and 2024 that had been removed from the program by the SAFE rule.
These incentives are intended as a temporary measure supporting the transition to zero-emission
vehicles and to provide additional flexibility in meeting the revised MY 2023-2026 standards,
g The 2020 EPA Automotive Trends Report. EPA-420-R-21-003 January 2021.
xv

-------
especially in the earlier years. For further details, see Sections I.B.2 and II.B.4 within the
Preamble to the final rule.
The previous program also included credits for real-world emissions reductions not reflected
on the test cycles used for measuring CO2 emissions for compliance with the fleet average
standards. There were credits in place for using technologies that reduce GHG emissions that
aren't captured on EPA tests ("off-cycle" technologies) and improvements to air conditioning
systems that increase efficiency and reduce refrigerant leakage. These credit opportunities did
not sunset under the previous regulations and remain a part of the program through MY 2026 and
beyond. EPA has revised an aspect of the off-cycle credits program to provide additional
opportunities for manufacturers to generate credits by increasing the pre-defined menu credit cap
from 10 to 15 g/mile for MYs 2023 through 2026. EPA has also modified some of the regulatory
definitions used to determine whether a technology is eligible for the menu credits. EPA did not
change the air conditioning credit elements of the light-duty GHG program.
Summary of Costs and Benefits
We estimate that this rule will result in significant present value net benefits of $120 billion to
$190 billion (annualized net benefits of $6.2 billion to $9.5 billion) - that is, the total benefits far
exceed the total costs of the program. Table 4 below summarizes EPA's estimates of total
discounted costs, fuel savings, and other benefits. The results presented here project the
monetized environmental and economic impacts associated with the revised standards during
each calendar year through 2050. The program will have significant social benefits including
$130 billion in climate benefits (with the average SC-GHGs at a 3 percent discount rate) and fuel
savings of $150 billion to $320 billion exclusive of fuel taxes. For American drivers, who
purchase fuel inclusive of fuel taxes, the fuel savings will total $210 billion to $420 billion in
present-value through 2050 consisting of $51 billion to $100 billion in savings in the form of fuel
taxes. With these fuel savings, consumers will benefit from reduced operating costs over the
vehicle lifetime.
The benefits include climate-related economic benefits from reducing emissions of GHGs that
otherwise contribute to climate change, reductions in energy security externalities caused by U.S.
petroleum consumption and imports, the value of certain particulate matter-related health
benefits (including premature mortality), the value of additional driving attributed to the rebound
effect, and the value of reduced refueling time needed to fill up a more fuel-efficient vehicle. The
analysis also includes estimates of economic impacts stemming from additional vehicle use, such
as the economic damages caused by crashes, congestion, and noise (from increased rebound
driving). See Chapter 10 for more information regarding these estimates.
xvi

-------
Table 4: Monetized Discounted Costs, Benefits, and Net Benefits of the Final Program for Calendar Years
through 2050 (Billions of 2018 dollars)a'b'c'd'e

Present Value
Annualized Value

3% Discount Rate
7% Discount Rate
3% Discount Rate
7% Discount Rate
Costs
$300
$180
$15
$14
Fuel Savings (exclusive of
$320
$150
$16
$12
taxes)




Benefits
$170
$150
$8.6
$8.1
Net Benefits
$190
$120
$9.5
$6.2
Notes:
a Values rounded to two significant figures; totals may not sum due to rounding. Present and annualized values are based on the stream of
annual calendar year costs and benefits included in the analysis (2021 - 2050) and discounted back to year 2021.
b Climate benefits are based on reductions in CO2, CH4 and N20 emissions and are calculated using four different estimates of the social cost
of each greenhouse gas (SC-GHG model average at 2.5 percent, 3 percent, and 5 percent discount rates; 95th percentile at 3 percent discount
rate), which each increase over time. For the presentational purposes of this table, we show the benefits associated with the average SC-GHGs
at a 3 percent discount rate but the Agency does not have a single central SC-GHG point estimate. We emphasize the importance and value of
considering the benefits calculated using all four SC-GHG estimates and present them later in this RIA. As discussed in Chapter 3.3 of the
RIA, a consideration of climate benefits calculated using discount rates below 3 percent, including 2 percent and lower, are also warranted
when discounting intergenerational impacts.
cThe same discount rate used to discount the value of damages from future GHG emissions (SC-GHGs at 5, 3, and 2.5 percent) is used to
calculate the present and annualized values of climate benefits for internal consistency, while all other costs and benefits are discounted at
either 3 percent or 7 percent.
dNet benefits reflect the fuel savings plus benefits minus costs.
eNon-GHG impacts associated with the standards presented here do not include the full complement of health and environmental effects that,
if quantified and monetized, would increase the total monetized benefits. Instead, the non-GHG benefits are based on benefit-per-ton values
that reflect only human health impacts associated with reductions in PM2.5 exposure.	
EPA estimates the average per-vehicle cost to meet the standards to be $1,000 in MY 2026, as
shown in Table 5 below. We discuss our cost analysis in more detail in Preamble Section III and
RIA Chapter 4.
Table 5 Car, Light Truck and Fleet Average Cost per Vehicle Relative to the No Action Scenario (2018
dollars)

2023
2024
2025
2026
Car
$150
$288
$586
$596
Light Truck
$485
$732
$909
$1,356
Fleet Average
$330
$524
$759
$1,000
The final standards will achieve significant reductions in GHG emissions. As seen in Table 6
below, through 2050 the program will achieve more than 3.1 billion tons of GHG emission
reductions, which is 50 percent greater emissions reductions than EPA's proposed standards.
Table 6 GHG Reductions Through 2050
Emission Impacts relative to No Action
Percent Change from No Action
C02
(Million metric tons)
CH4
(Metric tons)
N20
(Metric tons)
C02
CH4
N20
-3,125
-3,272,234
-96,735
-9%
-8%
-8%
xvii

-------
Summary of the Analysis of Alternatives to the Final Rule
Description of Alternatives
Along with the finalized standards, we analyzed both a more stringent and a less stringent
alternative. For the less stringent alternative, referred to as "Proposal" or "Proposed Standards",
EPA assessed the coefficients of the standards proposed in the NPRM, including the advanced
technology multipliers consistent with those proposed. Given the increased stringency of the
final standards compared to the proposal for MYs 2025 and 2026, EPA believes the proposal
represents an appropriate less stringent alternative for comparison.
For the more stringent alternative, referred to as "Alternative 2 minus 10", EPA assessed
Alternative 2 from our proposed rule with an additional 10 grams/mile increased stringency in
MY 2026, per our request for public comment on this option. This alternative is more stringent
than the final standards in MYs 2023 and 2024. For this alternative, EPA used the coefficients
from Alternative 2 in the proposed rule for MYs 2023 through 2025, with the standards
increasing in stringency by an additional 10 grams/mile compared to Alternative 2 standards in
MY 2026. The Alternative 2 minus 10 standards are the same as the final standards for MYs
2025 and 2026 and differ from the final standards in MYs 2023 and 2024.
EPA is finalizing several changes to program flexibilities. Further details regarding program
flexibilities can be found in Sections I.B.2 and II.B.4 of the Preamble to this Final Rule.
Flexibility changes, for the purpose of analyzing alternatives, are applied to Alternative 2 minus
10 as well as the final standards, as shown in Table 7 below, including the applicability of
flexibilities to the final standards and alternatives being analyzed.
Table 7: Applicability of Program Provisions to the Final Standards, and the Proposal and Alternative 2
minus 10 Standards
Provision
Final Standards
Proposal
Alternative 2
minus 10
Extension of credit carry-forward
MYs 2017 and
2018
MYs 2016-2020
MYs 2017 and
2018
Advanced technology incentive multipliers
MYs 2023-
2024, with cap
MYs 2022-2025
with cap
No
Increase of off-cycle menu cap from 10 to 15 g/mile
Yes, for MYs
2023-2026
Yes, beginning in
MY 2020
Yes, for MYs
2023-2026
Reinstatement of full-size pickup incentive for
strong hybrids or equivalent technologies
Yes, for MYs
2023 and 2024
Yes, for MYs
2022-2025
Yes, for MYs
2023 and 2024
Note:
EPA's technical analysis, presented in Chapter 4, consists of model runs using a model capable of reflecting some
but not all of these provisions. The modeling includes consideration of advanced technology incentive multipliers
for the proposed and final standards but not for the Alternative. The model runs also include the 15 grams per
mile off-cycle menu cap as appropriate given the standards or targets to which a fleet being modeled is
complying. Not included in the model runs are the full-size pickup truck technology incentive credit or the
extension of the emissions credit carry-forward.
The fleet average targets for the two alternatives compared to the final standards are provided
in Table 8. As discussed in detail in Chapter 2.3.3, there has been a proliferation of recent
announcements from automakers signaling a rapidly growing shift in investment away from
internal-combustion technologies and toward high levels of electrification. EPA has also heard
from a wide range of stakeholders over the past several months, including but not limited to the

-------
automotive manufacturers and the automotive suppliers, that the significant investments being
made now to develop and launch new EV product offerings and in the expansion of EV charging
infrastructure could enable higher levels of EV penetration to occur in the market place by the
model year 2026 time frame than we have projected in this rule for the revised model year 2026
standards.
Table 8: Projected Fleet Average Target Levels for Final Standards and Alternatives (CO2 grams/mile) *
Model Year
Final Standards
Projected Targets
Proposal Projected
Targets
Alternative 2 minus
10 Projected
Targets
2021**
229
229
229
2022**
224
224
224
2023
202
202
198
2024
192
192
189
2025
179
182
180
2026
161
173
161
* Targets shown are modeled results and, therefore, reflect fleet projections impacted by the
underlying standards. For that reason, slight differences in targets may occur despite equality
of standards in a given year.
** SAFE rule targets included here for reference.
260
240
220
jy
E
200
r\i
O
u
180
160

SAFE FRM
	2012 FRM
	Proposal
• • • 'Alternative 2 minus 10
Final Standards
x-	x
		
•
¦
/
/
/
/
/
/
/
1
140
2020 2021 2022
2023 2024
Model Year
2025
2026
2027
Figure 2: Final Standards Fleet Average Targets Compared to Alternatives
xix

-------
As shown in Figure 2, the range of alternatives that EPA analyzed differ from the final
standard targets in any given model year in 2023-2025 by 3 to 4 g/mile, and in model year 2026
by 12 g/mi. EPA believes the final standards are reasonable and appropriate considering the
relatively limited lead time for the standards (especially for model years 2023-2024), our
assessment of feasibility, the existing automaker commitments to meet the California Framework
(representing about 28 percent of the auto market), the standards adopted in the 2012 rule,
submitted public comments, the reasonableness of our estimated costs per vehicle, and the need
to reduce GHG emissions. EPA provides further discussion of the feasibility of the revised
standard and alternatives and the selection of the revised standards within Chapter 2.2.2. The
analysis of costs and benefits of the Proposal and Alternative 2 minus 10 standards is shown in
the Chapters 4, 5, 6, and 10.
Summary of Costs and Benefits of the Alternatives
EPA estimates that the less stringent (Proposal) alternative would result in significant present
value net benefits of $82 billion to $130 billion (annualized net benefits of $4.2 billion to $6.4
billion) - that is, the total benefits would far exceed the total costs of the program. Table 9 below
summarizes EPA's estimates of total discounted costs, fuel savings, and benefits for the
Proposal. The results presented here project the monetized environmental, public health and
economic impacts associated with the Proposal standards during each calendar year through
2050. The Proposal would have significant social benefits including $83 billion in climate
benefits (with the average SC-GHGs at a 3 percent discount rate) and fuel savings of $100 billion
to $210 billion exclusive of fuel taxes. For American drivers, who purchase fuel inclusive of fuel
taxes, the fuel savings would total $130 billion to $270 billion through 2050. With these fuel
savings, consumers would benefit from reduced operating costs over the vehicle lifetime.
The benefits include climate-related economic benefits from reducing emissions of GHGs that
otherwise contribute to climate change, reductions in energy security externalities caused by U.S.
petroleum consumption and imports, the value of certain particulate matter-related health
benefits (including premature mortality), the value of additional driving attributed to the rebound
effect, and the value of reduced refueling time needed to fill up a more fuel-efficient vehicle. The
analysis also includes estimates of economic impacts stemming from additional vehicle use, such
as the economic damages caused by crashes, congestion, and noise (from increased rebound
driving). See the Chapters 4, 5, 6, and 10 for more information regarding these estimates.
xx

-------
Table 9: Monetized Discounted Costs, Benefits, and Net Benefits of the Proposal Standards for Calendar
Years through 2050 (Billions of 2018 dollars)a'b'c'd'e

Present Value
Annualized Value
3% Discount Rate
7% Discount Rate
3% Discount Rate
7% Discount Rate
Costs
$190
$110
$9.8
$9.2
Fuel Savings
$210
$100
$11
$8.2
Benefits
$110
$96
$5.6
$5.3
Net Benefits
$130
$82
$6.4
$4.2
Notes:
a Values rounded to two significant figures; totals may not sum due to rounding. Present and annualized values are based on the stream of
annual calendar year costs and benefits included in the analysis (2021 - 2050) and discounted back to year 2021.
b Climate benefits are based on reductions in CO2, CH4 and N20 emissions and are calculated using four different estimates of the social cost
of each greenhouse gas (SC-GHG model average at 2.5 percent, 3 percent, and 5 percent discount rates; 95th percentile at 3 percent discount
rate), which each increase over time. For the presentational purposes of this table, we show the benefits associated with the average SC-GHGs
at a 3 percent discount rate but the Agency does not have a single central SC-GHG point estimate. We emphasize the importance and value of
considering the benefits calculated using all four SC-GHG estimates and present them later in this RIA. As discussed in Chapter 3.3 of the
RIA, a consideration of climate benefits calculated using discount rates below 3 percent, including 2 percent and lower, are also warranted
when discounting intergenerational impacts.
c The same discount rate used to discount the value of damages from future GHG emissions (SC-GHGs at 5, 3, and 2.5 percent) is used to
calculate the present and annualized values of climate benefits for internal consistency, while all other costs and benefits are discounted at
either 3 percent or 7 percent.
dNet benefits reflect the fuel savings plus benefits minus costs.
eNon-GHG impacts associated with the standards presented here do not include the full complement of health and environmental effects that,
if quantified and monetized, would increase the total monetized benefits. Instead, the non-GHG benefits are based on benefit-per-ton values
that reflect only human health impacts associated with reductions in PM2.5 exposure.	
We estimate that Alternative 2 minus 10 standards would result in significant present value
net benefits of $120 billion to $180 billion (annualized net benefits of $5.7 billion to $9.2 billion)
- that is, the total benefits would far exceed the total costs of the program. Table 10 below
summarizes EPA's estimates of total discounted costs, fuel savings, and benefits for Alternative
2 minus 10. The results presented here project the monetized environmental and economic
impacts associated with the final standards during each calendar year through 2050. Alternative 2
minus 10 would have significant social benefits including $130 billion in climate benefits (with
the average SC-GHGs at a 3 percent discount rate) and fuel savings of $160 billion to $320
billion exclusive of fuel taxes. For American drivers, who purchase fuel inclusive of fuel taxes,
the fuel savings would total $210 billion to $430 billion through 2050. With these fuel savings,
consumers would benefit from reduced operating costs over the vehicle lifetime.
The benefits include climate-related economic benefits from reducing emissions of GHGs that
otherwise contribute to climate change, reductions in energy security externalities caused by U.S.
petroleum consumption and imports, the value of certain particulate matter-related health
benefits (including premature mortality), the value of additional driving attributed to the rebound
effect, and the value of reduced refueling time needed to fill up a more fuel-efficient vehicle. The
analysis also includes estimates of economic impacts stemming from additional vehicle use, such
as the economic damages caused by crashes, congestion, and noise (from increased rebound
driving). See the Chapters 4,5,6, and 10 for more information regarding these estimates.
xxi

-------
Table 10: Monetized Discounted Costs, Benefits, and Net Benefits of Alternative 2 minus 10 for Calendar
Years through 2050 (Billions of 2018 dollars)a'b'c'd'e

Present Value
Annualized Value
3% Discount Rate
7% Discount Rate
3% Discount Rate
7% Discount Rate
Costs
$320
$190
$16
$15
Fuel Savings
$320
$160
$16
$13
Benefits
$170
$150
$8.9
$8.3
Net Benefits
$180
$120
$9.2
$5.7
Notes:
a Values rounded to two significant figures; totals may not sum due to rounding. Present and annualized values
are based on the stream of annual calendar year costs and benefits included in the analysis (2021 - 2050) and
discounted back to year 2021.
b Climate benefits are based on reductions in C02, CH4 and N20 emissions and are calculated using four
different estimates of the social cost of each greenhouse gas (SC-GHG model average at 2.5 percent, 3 percent,
and 5 percent discount rates; 95th percentile at 3 percent discount rate), which each increase over time. For the
presentational purposes of this table, we show the benefits associated with the average SC-GHGs at a 3 percent
discount rate but the Agency does not have a single central SC-GHG point estimate. We emphasize the
importance and value of considering the benefits calculated using all four SC-GHG estimates and present them
later in this RIA. As discussed in Chapter 3.3 of the RIA, 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 same discount rate used to discount the value of damages from future GHG emissions (SC-GHGs at 5, 3,
and 2.5 percent) is used to calculate the present and annualized values of climate benefits for internal consistency,
while all other costs and benefits are discounted at either 3 percent or 7 percent.
dNet benefits reflect the fuel savings plus benefits minus costs.
e Non-GHG impacts associated with the standards presented here do not include the full complement of health
and environmental effects that, if quantified and monetized, would increase the total monetized benefits. Instead,
the non-GHG benefits are based on benefit-per-ton values that reflect only human health impacts associated with
reductions in PM2 5 exposure.	
Summary of the Costs and Benefits of the Final Revised Standards Compared to the
Alternatives
Table 11 through Table 12 provide summaries of the final rule's costs and benefits compared
to the costs and benefits of the two alternatives that were analyzed. The benefits include climate-
related economic benefits from reducing emissions of GHGs that otherwise contribute to climate
change, reductions in energy security externalities caused by U.S. petroleum consumption and
imports, the value of certain particulate matter-related health benefits (including premature
mortality), the value of additional driving attributed to the rebound effect, and the value of
reduced refueling time needed to fill up a more fuel-efficient vehicle. The analysis also includes
estimates of economic impacts stemming from additional vehicle use, such as the economic
damages caused by crashes, congestion, and noise (from increased rebound driving). See
Chapters 4, 5, 6 and Chapter 10 for more information regarding these estimates. Net benefits for
the Final Revised Standards exceed those of either the Proposal or Alternative 2 minus 10 when
using a 3 percent discount rate. At a 7 percent discount rate, the net benefits for the Final
Standards are approximately equivalent to Alternative 2 minus 10 and exceed those of the
Proposal.
XXll

-------
Table 11: Present Value Monetized Discounted Costs, Benefits, and Net Benefits of the Final Program and
Alternatives for Calendar Years through 2050 (Billions of 2018 dollars)a'b'c'd'e

3% Discount Rate
7% Discount Rate

Final
Standards
Proposal
Alternative 2
minus 10
Final
Standards
Proposal
Alternative 2
minus 10
Costs
$300
$190
$320
$180
$110
$190
Fuel
$320
$210
$320
$150
$100
$160
Savings






Benefits
$170
$110
$170
$150
$96
$150
Net
$190
$130
$180
$120
$82
$120
Benefits






Notes:
a Values rounded to two significant figures; totals may not sum due to rounding. Present and annualized values
are based on the stream of annual calendar year costs and benefits included in the analysis (2021 - 2050) and
discounted back to year 2021.
b Climate benefits are based on reductions in CO2, CH4 and N20 emissions and are calculated using four different
estimates of the social cost of each greenhouse gas (SC-GHG model average at 2.5 percent, 3 percent, and 5
percent discount rates; 95th percentile at 3 percent discount rate), which each increase over time. For the
presentational purposes of this table, we show the benefits associated with the average SC-GHGs at a 3 percent
discount rate but the Agency does not have a single central SC-GHG point estimate. We emphasize the
importance and value of considering the benefits calculated using all four SC-GHG estimates and present them
later in this RIA. As discussed in Chapter 3.3 of the RIA, 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 same discount rate used to discount the value of damages from future GHG emissions (SC-GHGs at 5, 3,
and 2.5 percent) is used to calculate the present and annualized values of climate benefits for internal consistency,
while all other costs and benefits are discounted at either 3 percent or 7 percent.
dNet benefits reflect the fuel savings plus benefits minus costs.
e Non-GHG impacts associated with the standards presented here do not include the full complement of health
and environmental effects that, if quantified and monetized, would increase the total monetized benefits. Instead,
the non-GHG benefits are based on benefit-per-ton values that reflect only human health impacts associated with
reductions in PM2 5 exposure.	
XXlll

-------
Table 12: Annualized Monetized Discounted Costs, Benefits, and Net Benefits of the Final Program and
Alternatives for Calendar Years through 2050 (Billions of 2018 dollars)a,b,c,d,e

3% Discount Rate
7% Discount Rate

Final
Standards
Proposal
Alternative 2
minus 10
Final
Standards
Proposal
Alternative 2
minus 10
Costs
$15
$9.8
$16
$14
$9.2
$15
Fuel
$16
$11
$16
$12
$8.2
$13
Savings






Benefits
$8.6
$5.6
$8.9
$8.1
$5.3
$8.3
Net
$9.5
$6.4
$9.2
$6.2
$4.2
$5.7
Benefits






Notes:
a Values rounded to two significant figures; totals may not sum due to rounding. Present and annualized values
are based on the stream of annual calendar year costs and benefits included in the analysis (2021 - 2050) and
discounted back to year 2021.
b Climate benefits are based on reductions in CO2, CH4 and N20 emissions and are calculated using four different
estimates of the social cost of each greenhouse gas (SC-GHG model average at 2.5 percent, 3 percent, and 5
percent discount rates; 95th percentile at 3 percent discount rate), which each increase over time. For the
presentational purposes of this table, we show the benefits associated with the average SC-GHGs at a 3 percent
discount rate but the Agency does not have a single central SC-GHG point estimate. We emphasize the
importance and value of considering the benefits calculated using all four SC-GHG estimates and present them
later in this RIA. As discussed in Chapter 3.3 of the RIA, 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 same discount rate used to discount the value of damages from future GHG emissions (SC-GHGs at 5, 3,
and 2.5 percent) is used to calculate the present and annualized values of climate benefits for internal consistency,
while all other costs and benefits are discounted at either 3 percent or 7 percent.
dNet benefits reflect the fuel savings plus benefits minus costs.
e Non-GHG impacts associated with the standards presented here do not include the full complement of health
and environmental effects that, if quantified and monetized, would increase the total monetized benefits. Instead,
the non-GHG benefits are based on benefit-per-ton values that reflect only human health impacts associated with
reductions in PM2 5 exposure.	
RIA Chapter Summary
This document contains the following Chapters:
Chapter 1: Background
This chapter provides background on previous Agency actions with respect to the light-duty
vehicle GHG emissions program and summaries of previous EPA analyses.
Chapter 2: Technology Feasibility, Effectiveness, Costs, and Lead-time
This chapter summarizes the revisions to the model year 2023 and later light-duty vehicle
GHG standards. It also includes a summary of GHG compliance incentives and flexibilities and
discusses technological feasibility and manufacturer's lead-time considerations.
Chapter 3: Economic and Other Key Inputs
This chapter provides EPA's analyses of rebound effects, energy security impacts, the social
cost of greenhouse gases, and the costs associated with congestion and noise.
xxiv

-------
Chapter 4: Modeling GHG Compliance
This chapter discusses the analytical methodology used to model GHG emissions compliance
of the light-duty vehicle fleet with the standards and then summarizes the resulting estimated
compliance costs and associated technology pathways necessary to comply with the revisions to
the model year 2023 and later GHG standards.
Chapter 5: Projected Impacts on Emissions, Fuel Consumption, and Safety
This chapter documents EPA's analysis of the emission, fuel consumption and safety impacts
of the emission standards for light-duty vehicles. Light-duty vehicles include passenger vehicles
such as cars, sport utility vehicles, vans, and pickup trucks. Such vehicles are used for both
commercial and personal uses and are significant contributors to the total United States (U.S.)
GHG emission inventory.
Chapter 6: Vehicle Program Costs and Fuel Savings
In this chapter, EPA presents our estimated costs associated with the vehicle program. This
includes summaries of the vehicle level costs associated with new technologies expected to be
added to meet the model year 2023 and later GHG standards. The analysis also provides costs
associated with congestion, noise, fatalities and non-fatal crashes.
Chapter 7: Non-GHG Health and Environmental Impacts
In this chapter we discuss the health effects associated with non-GHG pollutants, specifically:
particulate matter, ozone, nitrogen oxides (NOx), sulfur oxides (SOx), carbon monoxide and air
toxics. These pollutants will not be directly regulated by the revisions to the GHG standards, but
the standards will affect emissions of these pollutants and precursors.
Chapter 8: Vehicle Sales, Employment, and Affordabilitv and Equity Impacts
This chapter presents the methodology and analytical results for EPA's modeling of vehicle
sales and employment impacts. It also includes an analysis of affordability and the potential
impacts on lower-income households, the used vehicle market, access to credit, and the low-
priced new vehicle segment.
Chapter 9: Small Business Flexibilities
This chapter discusses the flexibilities provided to small businesses under the revisions to the
model year 2023 and later light-duty GHG standards.
Chapter 10: Summary of Costs and Benefits
This Chapter presents a summary of costs, benefits, and net benefits of the program and the
alternatives.
xxv

-------
Chapter 1: Background
The analyses for this final rule represent the sixth time that EPA has analyzed the feasibility
and cost associated with meeting stringent GHG standards in the 2021/2022 through 2025/2026
timeframe. Previous analyses include the 2012 FRM, the 2016 Draft Technical Assessment
Report (DTAR), the 2016 MTE Proposed Determination, the 2018 analysis performed to update
the MTE analyses for the previous administration, and the proposal from earlier this year.
Through these six analyses EPA has applied five different initial fleets and updated critical
inputs such as fuel costs. We have continued to develop our cost and effectiveness assessments
and we have refined our analytical tools, including the CAFE Compliance and Effects Modeling
System (CCEMS), for the final rule. For more details regarding EPA's use of CCEMS for the
final rule analysis, please see Chapter 4. As discussed below and summarized in Figure 1-1, the
results have been remarkably consistent when comparing previous EPA analyses to the analysis
for the final rule summarized in Chapter 4.
In 2012, EPA established greenhouse gas (GHG) emissions standards for model year 2017
and later new passenger cars, light-duty trucks, and medium-duty passenger vehicles.1 The
program was projected to reduce GHG emissions from model year 2025 light-duty vehicles by
50 percent relative to model year 2010 vehicles.
As part of the 2012 Final Rule, EPA made a regulatory commitment to conduct a Midterm
Evaluation (MTE) of the standards for MY 2022-2025. As a part of this process, EPA examined
a wide range of factors, such as developments in powertrain technology, vehicle electrification,
vehicle mass reduction and potential vehicle safety impacts, the penetration of fuel efficient
technologies in the marketplace, consumer acceptance of fuel efficient technologies, trends in
fuel prices and the vehicle fleet, employment impacts, and many other factors.
The 2012 Final Rule established three formal steps for the MTE process:
1.	Draft Technical Assessment Report (TAR) to be issued jointly with the National
Highway Traffic Safety Administration (NHTSA) and the California Air Resources
Board (CARB) with opportunity for public comment. This was completed in July 2016.
2.	The EPA Administrator was to make a Proposed Determination with opportunity for
public comment. The Proposed Determination was completed in November 2016.
3.	The EPA Administrator was to make a final determination with regard to whether the
standards remained appropriate or should be changed no later than April 1, 2018. The
Final Determination was completed in January 2017 and the Revised Final Determination
was completed in April 2018.
There were opportunities for public input on the Draft TAR and the Proposed Determination
and a formal Response to Comments document was issued by EPA along with the Final
Determination in January 2017.
A timeline for the 2012 final rule, the MTE, and the SAFE rule is summarized within Figure
1-1. Despite the extensive EPA economic, scientific, and engineering analyses made publicly
available as part of the MTE process through the January 2017 Final Determination, and the
availability of an updated 2018 EPA MTE Analysis completed in January 2018, these prior EPA
1-1

-------
analyses were not used as the basis of the Agency's March 2017 MTE Reconsideration, April
2018 Revised MTE Final Determination or the proposed or final SAFE rules.
March 2017
Reconsideration of
the Midterm
Evaluation Final
Determination
August 2018
SAFE Vehicles
Proposed Rule for
Model Years 2021-
2026
July 2016
The Midterm
Evaluation (MTE)
Draft Technical
Assessment Report
January 2017
Final Determination
of the 2022-2025
LDV GHG
Standards under the
MTE
November 2016
Proposed
Detenu ination of the
2022-2025 TDV
GHG Standards
under the MTE
April 2020
SAFE Vehicles
Final Rule for
Model Years 2021-
2026
April 2018
Revised Final
Determination
imder the MTE
October 2012
Final Rule for
MY2017 and Later
LDV GHG
Emissions and
CAFE Standards
Figure 1-1: Regulatory Timeline for the Model Year (MY) 2017 and Later Light-duty Vehicle (LDV)
Greenhouse Gas (GHG) Emissions Standards, the Midterm Evaluation, and Safe Rule. The top row
represents Agency actions that used EPA analyses as the basis.
1.1 Summary of 2012 Final Rulemaking
1.1.1 Light-duty Vehicle GHG Emissions Standards
The 2017 and later light-duty vehicle GHG standards were established within the 2012 Final
Rulemaking (2012 FRM) based upon CO2 emissions-footprint curves, where each vehicle has a
different CO2 emissions compliance target depending on its characteristic footprint (i.e., the area
contained within the vehicle wheelbase and track width). In general, vehicles with a larger
footprint have higher corresponding vehicle CO2 emissions standards. As a result, the burden of
compliance within this program was distributed across all vehicles and all manufacturers and
each manufacturer would have its own fleet-wide standard that reflects the vehicles it chooses to
produce. The program also provided a wide range of credit programs and flexibilities for
manufacturers to meet 2017 and later GHG standards.
Table 1-1 shows the projected fleet-wide CO2 emission targets under the footprint-based
approach used in the 2012 FRM. Passenger car CO2 emission levels were projected to increase in
stringency from 212 to 143 grams per mile (g/mi) between MY s 2017 and 2025. Similarly, fleet-
wide CO2 emission levels for trucks were projected to increase in stringency from 295 g/mi in
MY 2017 to 203 g/mi in MY 2025. EPA projected that the average light-duty vehicle (combined
car and truck) tailpipe CO2 compliance level in MY 2017 would be 243 g/mi, phasing down by
MY 2025 to 163 g/mi. These projected targets in the first three rows include the effects of credits
and flexibilities. In contrast, the final row provides the actual tailpipe emissions achieved by
manufacturers for 2016-2019 based on certification data and excludes the effects of credits and
flexibilities.
1-2

-------
Table 1-1: Projected Fleet-Wide Emissions Compliance Targets under the Footprint-Based CO2 Standards in
the 2012 FRM

2016
2017
2018
2019
2020
2021
2022
2023
2024
2025

base









Passenger Cars
225
212
202
191
182
172
164
157
150
143
(g/mi)










Light Trucks
298
295
285
277
269
249
237
225
214
203
(g/mi)










Combined Cars
250
243
232
222
213
199
190
180
171
163
& Trucks (g/mi)










Actual Tailpipe
285
284
280
282






CO2, Cars &










Trucks (g/mi)










Notes:










Actual Tailpipe C02 adapted from the 2020 EPA Automotive Trends Report.2



The difference between actual tailpipe CO2 emissions and the projected standards is due to
not only the credits and flexibilities, but also the difference between the projected car/truck sales
mix at the time of the 2012 FRM, and the actual sales mix for each model year. The 2012 FRM
projected car sales greater than 60 percent for all model years. Table 1-2 shows the projected
sales mix from the original rule, the actual sales mix achieved, and the effective increase in
industry standards (in g/mi) for years 2016-2019 due solely to the increase in truck sales share.
For example, the combined standard of 222 g/mi projected for 2019 MY increased by 17 g/mi -
to 239 g/mi - primarily due to the 44 percent and 56 percent sales shares of passenger vehicles
and light trucks, respectively.21
Table 1-2: Projected vs. Actual Car/Truck Sales Share, 2016-2019 Model Years

2016
base
2017
2018
2019
Proj. Passenger Car Share
66%
63%
64%
64%
Proj. Light Truck Share
34%
37%
36%
36%
Actual Passenger Car Share
55%
53%
48%
44%
Actual Light Truck Share
45%
47%
52%
56%
Car/Truck Shift Effect on Stds. (g/mi)
+8
+8
+13
+17
Figure 1-2 and Figure 1-3 show the vehicle footprint vs. CO2 emissions standards curves for
cars and trucks, respectively, from the 2012 FRM. For passenger cars, the CO2 compliance
values associated with the footprint curves declined on average by approximately 5 percent per
year from the MY 2016 projected passenger car industry-wide compliance level through MY
2025. To separately address GHG compliance challenges faced while preserving the utility of
light-duty trucks (e.g., towing and payload capabilities), the GHG standards in the 2012 FRM
provided a lower annual rate of improvement for light-duty trucks during the initial years of the
program. The average annual rate of CO2 emissions reduction in MYs 2017 through 2021 were
3.5 percent per year, increasing to 5 percent per year for MYs 2022 through 2025.
a While there are other factors which further increased the standards (such as slight growth in average footprint since
2012, which increases the standards by another 4 g/mi), the most significant effect is seen in the difference in
car/truck sales mix.
1-3

-------
3 250
	2016
	2017
	2018
	2019
— ¦ - 2020
— 2021
	 - - 2022
2023
	2024
	2025
Figure 1-2: 2012 FRM Footprint Curves for Passenger Car CO2 (g/mile) Standards
2016
2017
2018
2019
2020
2021
2022
2024
Footprint (sf)
Figure 1-3: 2012 FRM Footprint Curves for Light-duty Truck CO2 (g/mile) Standards
1-4

-------
1.1.2 Flexibilities
EPA created flexibilities to ensure manufacturers could comply with the light-duty vehicle
GHG standards. Manufacturers create product plans with the goal of full compliance with EPA
regulations, however these product plans span many years and the flexibility to both bank credits
some years (over comply) and create debits (under comply) is a key flexibility within many EPA
regulations. Credit programs allow manufacturer to phase in new technologies during product
redesigns and new product introductions instead of redesigning all vehicles to comply at once.
Also, when executing product plans to meet our standards, manufacturers may need to respond to
changes in fuel prices, changes in consumer demand, and parts shortages such as the recent
semiconductor chip shortage3 that directly affect what manufacturers can build. EPA has
anticipated that manufacturers would need, and would take advantage of, program flexibilities
within its light-duty GHG programs. This includes both credits and incentives, such as car/truck
credit transfers, air conditioning credits, off-cycle credits, advanced technology vehicle
multipliers, intermediate volume manufacturer lead-time provisions, and hybrid and
performance-based incentives for full size pick-up trucks. See the 2017-2025 Preamble section
III.C (EO12866 2017-2025 GHG-CAFE Standards_2060-AQ54_FRM_FRN_20120827_Final)
for an extended discussion of these credits.
1.2 2016-2018 Midterm Evaluation of 2021-2025 Light-duty Vehicle GHG Standards
The Draft Technical Assessment Report (TAR), issued jointly by EPA, NHTSA, and CARB
for public comment, was the first formal step in the MTE process.4'5 A wide range of technical
and economic issues relevant to the light-duty GHG emissions standards for MY 2022-2025
were examined and shared with the public within the Draft TAR. The analyses contained within
the approximately 1,200 pages of the Draft TAR and the subsequent public comments received
on the Draft TAR informed the EPA's development of the Proposed Determination (PD)6'7 and
the Final Determination (FD).8'9 The primary conclusions of the Draft TAR were:
•	A wider range of technologies exist for manufacturers to use to meet the MY 2022-
2025 standards at costs similar to, or lower than, those projected in the 2012 rule;
•	Advanced gasoline vehicle technologies will continue to be the predominant
technologies, with modest levels of strong hybridization and very low levels of full
electrification (plug-in vehicles) needed to meet the standards;
•	The car/truck mix reflects updated consumer trends that are informed by a range of
factors including economic growth, gasoline prices, and other macro-economic trends.
However, as the standards were designed to yield improvements across the light-duty
vehicle fleet, irrespective of consumer choice, updated trends are fully accommodated
by the footprint-based standards.
The analyses from the Draft TAR were further updated and included as part of an
approximately 700-page Technical Support Document10 (TSD) released in conjunction with the
PD and referenced within the FD. Key updates within the TSD included:
•	Use of the fuel prices, vehicle sales volumes, and car/truck mix from the 2016 Energy
Information Administration's Annual Energy Outlook (AEO2016)11
•	Use of MY 2015 for the base year vehicle fleet
1-5

-------
•	Changes to EPA's vehicle simulation model to include the most recent data on
technology effectiveness from the EPA vehicle benchmarking testing program and
other sources
•	Changes to battery costs for electrified vehicles based upon updated data from the
Argonne National Laboratories (ANL) Battery Performance and Cost (BatPaC) model
•	Building in additional quality assurance checks of technology effectiveness estimates
•	Changes to vehicle class definitions for effectiveness modeling and a greater
resolution of vehicle types to provide more accuracy and precision in representing
technology cost and effectiveness for the future vehicle fleet
•	Better accounting for tire and aerodynamic improvements in the base year vehicle
fleet
The Administrator's November 2016 Proposed Determination was the following:
1.	The MY 2022-2025 light-duty GHG standards are feasible
2.	The standards will achieve significant CO2 and oil reductions
3.	The standards will provide significant benefits to consumers and the public
4.	The auto industry is thriving and meeting the standards more quickly than required
5.	Continued reductions in CO2 emissions are essential to help address the threat of climate
change
The Administrator also determined that there was ample evidence that supported
strengthening the standards; however, she chose not to propose revising the levels of the GHG
standards finalized in 2012. Comments received on the Draft TAR were addressed as part of a
formal response to comments within the appendices of the TSD. 12 Comments on the PD and
TSD were addressed within a separate Response to Comments Document released as part of the
FD.13
The Administrator's January 2017 Final Determination was:
1.	The MY 2022-2025 standards remain appropriate under section 202(a)(1) of the Clean
Air Act
2.	The standards are feasible at reasonable cost, without need for extensive electrification
3.	The standards will achieve significant CO2 and oil reductions
4.	The standards will provide significant benefits to consumers and to the public
5.	The auto industry is thriving and meeting the standards more quickly than required
1.2.1 Updated EPA 2018 MTE Analysis
EPA completed an analysis in January 2018 that further updated the analyses from the TSD.14
Although conducted by EPA as part of the MTE and to inform an anticipated SAFE NPRM, the
updated EPA analysis was not used as part of the revised Final Determination of April 2018, the
SAFE NPRM, or the SAFE FRM. The following updates to the November 2016 TSD were
included within EPA's updated January 2018 MTE analysis:
1-6

-------
1.2.1.1	Updated Base Year Fleet Data:
The base year vehicle fleet was updated from MY 2015 to MY 2016. The EIA AEO sales
projections, car and truck percentages within the fleet, and fuel prices were updated to
AEO2017.15 The OEM and vehicle class market share projections were updated using data
purchased from IHS Markit, Ltd. Improvements were also implemented in EPA's ALPHA
vehicle model and OMEGA compliance model to better characterize technologies within the
base year fleet. The improvements included conducting ALPHA vehicle model runs for each
vehicle configuration in the base year (road loads; engine, transmission, and accessory models)
and confirming model alignment with CO2 from EPA vehicle emissions certification data.
The resolution of technology characterization was improved within EPA's ALPHA vehicle
model and OMEGA GHG compliance model via the following changes:
•	Increased number of engine maps for turbocharged/downsized engines (i.e., 3 different
engine maps vs. 1 for TSD)
•	Increased number of engine maps to represent port-fuel-injected (PFI) and gasoline
direct injection (GDI) engines (2 different engine maps each vs. 1 each for the TSD)
•	Use of fleet-wide technology characterization to characterize the GHG performance of
the 2016 fleet based primarily on certification data submitted by manufacturers to
EPA's VERIFY Database
Additional data was also obtained from EPA's Test Car Database and technical specifications
that were not available in either the EPA VERIFY or Test Car databases (e.g. curb weight,
dimensions, power steering type) were obtained via other public and commercially available
sources of vehicle data such as Edmunds.com©, Wards Automotive (Penton©) and AllData
Repair (AllData LLC©). Further details of the 2016 base year fleet characterization can be found
in Bolon et al.16
1.2.1.2	Updated Fuel Price and Fleet Projections
Future fuel prices were updated to reflect AEO2017 projections.15 Updated fleet volume and
car/truck percentage projections were based on preliminary AEO2018 projections and updated
IHS Markit forecasting.17
1.2.1.3	Other Updates to the ALPHA Vehicle Model
ALPHA modeling process improvements were put into place to implement cloud computing
and improve computational efficiency. This allowed full combinatorial modeling of vehicle
technology packages, including all combinations of engines, transmissions, accessories and road
loads. The introduction of full combinatorial modeling allowed replacement of the Lumped
Parameter Model (LPM) previously used within the OMEGA model with peer-reviewed
response surface equations (RSEs) based entirely on ALPHA modeling. Under this approach,
packages applied to future vehicles contained only the technology combinations reflected within
ALPHA runs. This also eliminated any manual calibration of the LPM.
Mass reduction (MR) was applied in predefined steps based on the amount of MR required to
move a vehicle into a new estimated test weight (ETW) bin. Mass reduction in passenger cars
was thus not constrained by lower curb weight limits as was done for the previous TSD safety
analysis.
1-7

-------
BEVs, PHEVs and HEVs were mapped into unique vehicle types rather than being mapped
into ICE vehicle types. The total number of vehicle types increased to 42 from the 29 vehicle
types used within the TSD analysis, which allowed for greater granularity in both cost and
effectiveness calculations.
1.2.1.4	Updates to the Technologies Considered and Technology Effectiveness
Additional technologies were used in the updated analysis that were not used within the TSD
to better reflect recent vehicle product introductions. Effectiveness for engine technologies was
also updated based on EPA engine and chassis dynamometer benchmarking. Updates included:
•	Addition of a new, dynamically-controlled cylinder deactivation technology
(deacFC)18 based on vehicle benchmarking of Tula's Dynamic Skip Fire system, with
greater effectiveness than traditional fixed cylinder deactivation (deacPD), although at
higher costs due to the necessity for deactivation hardware for each cylinder
•	Addition of a 2nd generation turbocharged downsized engine package based on EPA
benchmark testing of the Honda L15B7 1.5L turbocharged, direct-injection engine19
•	ALPHA modeling of 12V Start-Stop and 48V Mild Hybrids for every combination of
engine/trans/vehicle class instead of using constant effectiveness for these
technologies applied to each vehicle class within the TSD
•	Use of an engine map for Atkinson (ATK2+CEGR) technology based on EPA
benchmark testing of the MY 2018 Camry 2.5L A25A FKS engine18 in place of using
developmental engine test data and GT-POWER engine modeling within the TSD
•	Updates to both aerodynamic drag technologies and other road-load reducing
technologies20
1.2.1.5	Updates to Cost Analysis
A significant number of updates were included within the cost analysis. This included updates
to the costs of vehicle electrification and other technology, some changes to indirect costs, and
use of a 2016 dollar basis in order to be consistent with AEO2017. The changes to the cost
analysis relative to the TSD included:
•	Use of an updated ANL BatPaC model (BatPaC Version 3.1,9 October 2017) as the
basis for BEV, PHEV, HEV and mild HEV battery costs
•	The learning curves for battery costs were adjusted to ensure consistency between
BatPaC and OMEGA
•	Non-battery BEV and PHEV costs were updated based on more recent teardown data
from California Air Resources Board, UBS, and other references.21,22:23
•	Level 2 home charging costs were updated based on data provided by the California
Air Resources board on the cost of electric vehicle service equipment (EVSE).24
•	BEV/PHEV battery and non-battery integration efforts were changed within OMEGA
to a "medium complexity" as opposed to the "high complexity" used in the TSD,
resulting in application of a reduced indirect cost markup
•	Some additional cost savings were applied for BEVs since they did not need to add
additional technology to comply with light-duty Tier 3 criteria pollutant emissions
standards. Such costs were found to have been applied to BEVs within the TSD.
1-8

-------
•	Markups on emerging and future technologies remained at near-term levels through
2025 instead of using near-term levels through 2018 or 2024 as was done in the TSD
•	LUB2 & EFR2 were added as incremental technologies to LUB1 and EFR1 in
OMEGA and both LUB1 and EFR1 were included in all base year Exemplar vehicles.
•	Cooled EGR costs were changed to a single EGR loop when applied to ATK2 engines.
Previous Cooled EGR costs had assumed a higher cost low-pressure/higher pressure
dual loop system for application to highly boosted (27-bar BMEP) turbocharged
engines no longer used within the analysis.
1.2.1.6 Updated Sensitivity Analyses
The range of sensitivities analyzed within the OMEGA model for the Updated EPA 2018
MTE Analysis included:
•	AEO2016 central, high, and low fuel price scenarios11'13
•	No additional mass reduction beyond what existed in MY 2016 base year fleet
•	Technology adoption for 20 percent of trucks constrained to 2021 standards level
•	Limiting the adoption of advanced, non-turbo engine technology to 10 percent of fleet
•	No new adoption of advanced transmission technologies
•	No new adoption of advanced turbocharged/downsized engines
•	Added consideration of credit trading between manufacturers
•	No car-truck credit transfers within a manufacturer's fleet
1.2.2 Comparison of Analytical Results Between the 2012 FRM and the MTE
Table 1-3 provides a comparison of MY 2025 light-duty vehicle fleet-average technology
penetrations and per-vehicle costs for the central analytical case from the 2012 FRM and for
central analytical cases and sensitivity analyses for the Draft TAR, TSD, and EPA's Updated
2018 Analysis. Although EPA is finalizing new standards for MY's 2023 through 2026, a
comparison of the CEMMS analytical results for the final MY 2025 and MY 2026 standards (see
RIA Chapter 4.1.2) shows remarkable consistency with analytical results over the last 10 years.
Figure 1-4 shows a graphical representation comparing per vehicle costs for the same 2012
through 2018 EPA analyses. Table 1-4 compares the fuel prices, assumed car/truck fleet mix and
resulting fleet average CO2 g/mile emissions targets for each of these analyses. Table 1-5
provides per vehicle costs in 2025 broken down separately for cars and trucks in the light-duty
vehicle fleet. The CEMMS analysis in RIA Chapter 4.1.3 found fleet-level per vehicle costs of
$759 and $1000 for MY 2025 and MY 2026, respectively, and previous EPA analyses ranged
from $922 to $1228 per vehicle for a roughly comparable level of stringency.c
b The CCEMS analysis for this rule described in Chapter 4 uses AEO2021 for estimating gasoline prices. In general,
AEO2021 reference, high, and low estimates for the retail price of gasoline are lower than comparable cases within
AEO2016 and AEO 2017. For example, in 2025 the AEO2021 reference retail gasoline price in 2018$ is estimated
to be $2.44 per gallon vs. $3.13 and $3.05 per gallon for AEO2016 and AEO2017, respectively.
c Please note, however, that there are differences in the "no action" cases used for determining costs between the
final standards and the previous 2012 - 2018 EPA analyses. For a complete description of the "no action" case used
for this rulemaking, please see Chapter 4.
1-9

-------
$1,400
$1,200
S $1,000
* $800
$600
$400
$200
Si.228
5969
5922
5976
~	2012 FRM
~	Draft Tar
~	PD/FDTSD
~	2018 Analysis
| Sensitivities
2012 FRM	Draft TAR	PD/FDTSD 2018 EPA Analysis
(July 2016)	(Nov. 2016)	(January 2018)
Figure 1-4: Comparison of fleet average (car and truck), per-vehicle technology costs in 2025 from the 2012
FRM to subsequent analyses conducted by EPA during the MTE (2018$). Vertical lines on top of the bars
represent the range of sensitivity analyses conducted.
1-10

-------
Table 1-3: Comparison of technology penetrations into the light-duty fleet and per vehicle costs in 2025
(incremental to 2021) for the 2012 FRM compared to analyses conducted by EPA under the MTE. All per
vehicle costs are shown in 2018$ to maintain consistency with other analyses within this RIA
Technology [1]
2012 Final
Rule [2]
Draft TAR [3]
PD/FD TSD [4]
Updated EPA 2018 Analysis
[51,[61,[71
Primary
Analysis
Range of Sensitivities
Analyzed
Primary
Analysis
Range of Sensitivities
Analyzed
Primary
Analysis
Range of Sensitivities
Analyzed
Advanced High-
efficiency
Engines [8]
93%
81%
58% to 86%
62%
36% to 82%
67%
56% to 73%
Cylinder
Deactivation
not modeled
not
modeled
not modeled
49%
43% to 55%
28%
24% to 31%
8 speed and
other advanced
transmissions
(%) T91
91%
90%
89% to 91%
93%
92% to 94%
90%
90% to 94%
Mass reduction
(%) noi
-7%
6%
2% to 6%
8%
1% to 9%
4%
2% to 5%
Off-cycle
technology (%)
not modeled
not
modeled
not modeled
26%
8% to 53%
not
modeled
not modeled
Stop-start (%)
15%
20%
15% to 31%
15%
12% to 39%
16%
12% to 20%
Mild Hybrid (%)
26%
18%
13% to 38%
18%
16% to 27%
1%
0% to 3%
Strong Hybrid
(%)
5%
2.6%
2.0% to 3.0%
2%
2% to 3%
2%
1% to 2%
phev (%)rm
0%
1.7%
2% to 2%
2%
2% to 2%
1%
1% to 1%
BEV (%) [111
2%
2.6%
2.0% to 3.0%
3%
2% to 4%
2%
1% to 2%
Per vehicle cost
(2018$)
$1,228
$969
$938 to $1,125
$922
$840 to $ 1,175
$976
$942 to $1,242
Notes:
[1]	Technology penetrations shown are absolute and MY 2025 vehicle costs are incremental to MY 2021.
[2]	The 2012 FRM values are based on the AEO2012 Early Release "Reference Case" and analytical results were originally reported as average
per vehicle costs of $1070 in 2010$.
[3]	The Draft TAR values are based on the AEO 2015 "Reference Case" and analytical results were originally reported as average per vehicle
costs of $894 in 2013$.
[4]	The Proposed/Final Determination values are based on the AEO 2016 "Reference Case", which included a 53 percent/47 percent car/truck
mix. Analytical results were originally reported as average per vehicle costs of $875 in 2015$.
[5]	The 2018 Updated Analysis values are based on the AEO 2017 "Reference Case", which included a 42 percent/58 percent car/truck mix.
Analytical results were originally reported as average per vehicle costs of $935 in 2016$.
[6]	Advanced high-efficiency engines updated based on benchmarking of MY 2016 and MY 2017 OE engines.
[7]	Lumped parameter modeling was completely removed in favor of peer reviewed response surface equations based entirely on ALPHA
vehicle modeling.
[8]	Includes both turbocharged/downsized and Atkinson Cycle engines.
[9]	Including continuously variable transmissions (CVT).
[10]	The mass reductions are fleet average percent reduction in curb weight relative to the 'null' package.
[111 BEV and PHEV penetrations include the California Zero Emission Vehicles (ZEV) program.
As EPA analyses were updated for the MTE through 2018, projected 2025 fuel prices
decreased, the car/truck fleet mix shifted to a higher percentage of trucks, and the fleet CO2
g/mile targets increased relative to the analysis for the 2012 FRM (Table 1-4).
The MTE analyses reflect an approximate $200 decrease in per vehicle fleet costs relative to
the 2012 FRM analysis. Some of the MTE sensitivity analyses have per vehicle costs that are
approaching or approximately equivalent to that of the 2012 FRM analysis. Despite considerable
updates to the EPA analyses between 2012 and 2018, and a significant increase in the percentage
of trucks in light-duty fleet (from 33 percent in the 2012 FRM analysis to 58 percent in the 2018
analysis), per vehicle costs for either the light-duty vehicle fleet (Figure 1-4) or separately for
light-duty car or trucks (Table 1-5) have remained remarkably stable.
1-11

-------
Table 1-4: Comparison of fuel price, percentage of cars and trucks in the fleet, and CO2 fleet average
emissions targets when taking into consideration the car and truck fleet mix for the 2012 FRM compared to
analyses conducted by EPA under the MTE.

2012 Final Rule
[1]
Draft TAR [2]
PD/FD TSD [3]
Updated EPA 2018
Analysis [4]
2025 Fuel Price
$4.44
$3.20
$3.13
$3.05
2025 Car/Truck Fleet Mix[5]
67%/33%
52%/48%
53%/47%
42%/58%
2025 Fleet CO2 Target
(g C02/mi)
163
175
173
180
Notes:
[1]	AEO 2011 Reference Case, fuel price converted to 2018$
[2]	AEO 2015 Reference Case, fuel price converted to 2018$
[3]	AEO 2016 Reference Case, fuel price converted to 2018$
[4]	AEO 2017 Reference Case, fuel price converted to 2018$
[5]	Car/Truck definitions used by EPA for GHG standards differ from those used by AEO. The Car/Truck Fleet
Mix in 2025 is based upon EPA's regulatory car and truck definitions.
Table 1-5: Comparison per vehicles costs for passenger cars, light-duty trucks and the combined light-duty
vehicle fleet in 2025 (incremental to 2021) for the 2012 FRM compared to analyses conducted by EPA under
the MTE. Per vehicle costs are shown in 2018$ to maintain consistency with other analyses within this RIA.

2012 Final Rule [1]
Draft TAR [2]
PD/FD TSD [3]
Updated EPA 2018 Analysis [4]
(sensitivity range in parentheses)
Car
$1,101
$766
$790
$805
($805 -$1,021)
Truck
$1,487
$1,191
$1,073
$1,098
($1,010-$1,454)
Fleet
$1,228
$969
$922
$976
($942-$1,242)
Notes:
[1]	AEO 2011 Reference Case, converted to 2018$
[2]	AEO 2015 Reference Case, converted to 2018$
[3]	AEO 2016 Reference Case, converted to 2018$
[4]	AEO 2017 Reference Case, converted to 2018$
1-12

-------
1.3 Agency Actions, March 2017 - April 2020
1.3.1	2017 Reconsideration of the MTE Final Determination and 2018 MTE Final
Determination
On March 15, 2017 EPA announced that the final determination, issued on January 12, 2017,
would be reconsidered in coordination with NHTSA. On April 2, 2018, a new Mid-term
Evaluation Final Determination was signed, which withdrew the previous Final Determination
and found that the model year 2022-2025 greenhouse gas standards were not appropriate and
should be revised.25
1.3.2	SAFE
In April 2020, EPA published "The Safer Affordable Fuel-Efficient (SAFE) Vehicles Rule for
Model Years 2021-2026 Passenger Cars and Light Trucks," a final rule amending the 2012 FRM
beginning with MY 2021 by establishing new and substantially less stringent GHG standards for
MY 2021 and later light-duty vehicles.26
On January 20, 2021, President Biden signed Executive Order 13990, which issued the
following directives to EPA and other federal agencies:
"Section 1. Policy. Our Nation has an abiding commitment to empower our workers and
communities; promote and protect our public health and the environment; and conserve our
national treasures and monuments, places that secure our national memory. Where the Federal
Government has failed to meet that commitment in the past, it must advance environmental
justice. In carrying out this charge, the Federal Government must be guided by the best science
and be protected by processes that ensure the integrity of Federal decision-making. It is,
therefore, the policy of my Administration to listen to the science; to improve public health and
protect our environment; to ensure access to clean air and water; to limit exposure to dangerous
chemicals and pesticides; to hold polluters accountable, including those who disproportionately
harm communities of color and low-income communities; to reduce greenhouse gas emissions;
to bolster resilience to the impacts of climate change; to restore and expand our national
treasures and monuments; and to prioritize both environmental justice and the creation of the
well-paying union jobs necessary to deliver on these goals.
To that end, this order directs all executive departments and agencies (agencies) to
immediately review and, as appropriate and consistent with applicable law, take action to
address the promulgation of Federal regulations and other actions during the last 4 years that
conflict with these important national objectives, and to immediately commence work to confront
the climate crisis.
Sec. 2. Immediate Review of Agency Actions Taken Between January 20, 2017, and January
20, 2021. (a) The heads of all agencies shall immediately review all existing regulations, orders,
guidance documents, policies, and any other similar agency actions (agency actions)
promulgated, issued, or adopted between January 20, 2017, and January 20, 2021, that are or
may be inconsistent with, or present obstacles to, the policy set forth in section 1 of this order.
For any such actions identified by the agencies, the heads of agencies shall, as appropriate and
consistent with applicable law, consider suspending, revising, or rescinding the agency actions.
In addition, for the agency actions in the 4 categories set forth in subsections (i) through (iv) of
this section, the head of the relevant agency, as appropriate and consistent with applicable law,
1-13

-------
shall consider publishing for notice and comment a proposed rule suspending, revising, or
rescinding the agency action within the time frame specified.
... (ii) Establishing Ambitious, Job-Creating Fuel Economy Standards: 'The Safer Affordable
Fuel-Efficient (SAFE) Vehicles Rule Part One: One National Program,' 84 FR 51310
(September 27, 2019), by April 2021; and 'The Safer Affordable Fuel-Efficient (SAFE) Vehicles
Rule for Model Years 2021-2026 Passenger Cars and Light Trucks,' 85 FR 24174 (April
30,2020), by July 2021. In considering whether to propose suspending, revising, or rescinding
the latter rule, the agency should consider the views of representatives from labor unions, States,
and industry. "21
With respect to § 2(ii) of Executive Order 13990 for the purposes of this document, we are
referring to 84 FR 51310 as "SAFE" and 85 FR 24174 as "SAFE2". The revision of MY 2023 to
MY 2026 Light-duty Vehicle GHG standards under SAFE2 is the purpose of the Notice of
Proposed Rulemaking of which this Regulatory Impact Analysis is a part. Reconsideration of
SAFE is the subject of a separate Agency action.28
In response to this Executive Order, EPA has considered taking action under the Clean Air
Act with respect to the SAFE GHG emissions standards. As described in further detail in
Preamble Section VI and elsewhere in the preamble to this rulemaking, we are finalizing more
stringent GHG standards under our Clean Air Act authority. For more information regarding the
SAFE rule and why EPA believes that the revised final standards are appropriate, see section
I. A. 1 of the preamble to this final rule.
1.3.2.1 New GHG Compliance Flexibilities Established Under SAFE2
As part of the amendment of MY 2021 and later GHG emissions standards under SAFE2, a
small number of flexibilities related to real world fuel efficiency improvements were included.
EPA continued to allow manufacturers to make improvements related to air conditioning
refrigerants and leakage and credit those improvements toward compliance with GHG standards.
EPA made no changes to the 10 g-C02/mi off-cycle credit cap. EPA also extended the "0 g/mi
upstream" incentive for electric vehicles through 2026 beyond its original sunset of MY 2021
and established a new credit multiplier for natural gas vehicles through the 2026 model year. For
natural gas vehicles, both dedicated and dual-fueled, EPA established a multiplier of 2.0 for
model years 2022-2026.
To support easier use of certain real world fuel efficiency improvements, EPA added high
efficiency alternators and advanced A/C compressors to the off-cycle credit menu to help
streamline the program by allowing manufacturers to select the menu credit g/mi values rather
than continuing to seek credits through the public approval process. The credit levels added to
the menu were based on data previously submitted by multiple manufacturers through the off-
cycle credits application process. The high efficiency alternator credit is scalable with efficiency,
providing an increasing credit value of 0.16 grams/mile CO2 per percent improvement as the
efficiency of the alternator increases above a baseline level of 67 percent efficiency. The
advanced A/C compressor credit value is 1.1 grams/mile for both cars and light trucks. For more
information on any aspect of these changes see 84 FR 24174, April 30, 2020.26 For a summary of
the final revisions of the averaging banking and trading program and credit carry forward
provisions, please refer to Section II. A. 4. within the Preamble for the Final Rule.
1-14

-------
References for Chapter 1
1	U.S. EPA and U.S. DOT/NHTSA. 2017 and Later Model Year Light-Duty Vehicle Greenhouse Gas Emissions
and Corporate Average Fuel Economy Standards. U.S. Federal Register, Vol. 77, No. 199, pp. 62624-63200,
October 15, 2012.
2	U.S. EPA. The 2020 EPA Automotive Trends Report - Greenhouse Gas Emissions, Fuel Economy, and
Technology since 1975. EPA-420-R-21-003, January 2021.
3	LaReau, J.L. "Everything you need to know about the chip shortage that's plaguing automakers", Detroit Free
Press June 15, 2021 Last accessed on the internet 7/16/2021 URL:
https://www.freep.com/story/money/cars/2021/06/15/car-chip-shortage-2021/7688773002/
4	U.S. EPA and U.S. DOT/NHTSA. Notice of Availability of Midterm Evaluation Draft Technical Assessment
Report for Model Year 2022-2025 Light Duty Vehicle GHG Emissions and CAFE Standards. U.S. Federal Register,
Vol. 81, No. 144, pp. 49217-49220, July 27, 2016.
5	U.S. EPA, CA-EPA/ARB, U.S. DOT/NHTSA. Draft Technical Assessment Report: Midterm Evaluaton of Light-
Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average Fuel Economy Standards for Model
Years 2022-2025. EPA-420-D-16-900, July 2016.
6	U.S. EPA. Proposed Determination on the Appropriateness of the Model Year 2022-2025 Light-Duty Vehicle
Greenhouse Gas Emissions Standards Under the Midterm Evaluation. U.S. Federal Register, Vol. 81, No. 234, pp.
87927-87928, December 6, 2016.
7	U.S. EPA Proposed Determination on the Appropriateness of the Model Year 2022-2025 Light-Duty Vehicle
Greenhouse Gas Emissions Standards under the Midterm Evaluation. EPA-420-R-16-020, November 2016.
0
U.S. EPA. EPA Determination Letter from Administrator Gina McCarthy to Stakeholders Regarding Model Year
2022-2025 Light-Duty Greenhouse Gas Standards, January 12, 2017. Last accessed on the Internet on April 23,
2021 at the following URL: https://www.epa.gov/sites/production/files/2017-01/documents/mte-stakeholder-letter-
2017-01-12.pdf
9	U.S. EPA. Final Determination on the Appropriateness of the Model Year 2022-2025 Light-Duty Vehicle
Greenhouse Gas Emissions Standards under the Midterm Evaluation. EPA-420-R-17-001, January 2017.
10	U.S. EPA. Proposed Determination on the Appropriateness of the Model Year 2022-2025 Light-Duty Vehicle
Greenhouse Gas Emissions Standards under the Midterm Evaluation: Technical Support Document. EPA-420-R-16-
021, November, 2016.
11	U.S. Energy Information Administration. Annual Energy Outlook 2016 with Projections to 2040. DOE/EIA-
0383(2016). August 2016.
12	U.S. EPA. Proposed Determination on the Appropriateness of the Model Year 2022-2025 Light-Duty Vehicle
Greenhouse Gas Emissions Standards under the Midterm Evaluation: Technical Support Document, Appendices A
through D. EPA-420-R-16-021, November, 2016.
13	U.S. EPA. Final Determination on the Appropriateness of the Model Year 2022-2025 Light-Duty Vehicle
Greenhouse Gas Emissions Standards under the Midterm Evaluation - Response to Comments. EPA-420-R-17-002,
January 2017.
14	McDonald, J. Memo to Docket, No. EPA-HQ-OAR-2021-0208, EPA/OAR/OTAQ January 9, 2018 Briefing for
OAR Assistant Administrator Wehrum.
15	U.S. Energy Information Administration. Annual Energy Outlook 2017 with Projections to 2050. January 2017.
16	Bolon, K., Moskalik, A., Newman, K., Hula, A., Neam, A. (2018). Characterization of GHG Reduction
Technologies in the Existing Fleet. SAE Technical Paper 2018-01-1268, doi: 10.4271/2018-01-1268.
1	7
U.S. Energy Information Administration. Annual Energy Outlook 2018 with Projections to 2050. February 2018.
18 Kargul, J., Stuhldreher, M., Barba, D., Schenk, C., Bohac, S., McDonald, J., & Dekraker, P. (2019).
Benchmarking a 2018 Toyota Camry 2.5-liter Atkinson Cycle Engine with Cooled-EGR. SAE International Journal
of Advances and Current Practices in Mobility, 1(2), 601.
1-15

-------
19	Stuhldreher, M., Kargul, J., Barba, D., McDonald, J., Bohac, S., Dekraker, P., & Moskalik, A. (2018).
Benchmarking a 2016 Honda Civic 1.5-liter L15B7 turbocharged engine and evaluating the future efficiency
potential of turbocharged engines. SAE International journal of engines, 11(6), 1273.
20	Moskalik, A., Bolon, K., Newman, K., & Cherry, J. (2018). Representing GHG Reduction Technologies in the
Future Fleet with Full Vehicle Simulation. SAE International journal of fuels and lubricants, 11(4), 469.
21	California Air Resources Board. Advanced Strong Hybrid and Plug-In Hybrid Engineering Evaluation and Cost
Analysis, CARB Agreement 15CAR018, prepared for CARB and California EPA by Munro & Associates, Inc. and
Ricardo Strategic Consulting, April 25, 2017.
22	Hummel, P., Lesne, D., Radlinger, J., Golbaz, C., Langan, C., Takahashi, K.,... & Shaw, L. (2017). UBS
Evidence Lab Electric Car Teardown—Disruption Ahead. UBS report, Basel.
23	Safoutin, M.J. (2018) Predicting Powertrain Costs for Battery Electric Vehicles Based on Industry Trends and
Component Teardowns. Proceedings of the 31 st International Electric Vehicle Symposium & Exhibition and
International Electric Vehicle Technology Conference. Society of Automotive Engineers of Japan, 2018. ISBN:
9781510891579.
24
California Air Resources Board. California's Advanced Clean Cars Midterm Review - Summary Report for the
Technical Analysis of the Light-duty Vehicle Standards. (2017). Last accessed on the Internet on 5/6/2021 at the
following URL: https://ww2.arb.ca.gov/resources/documents/2017-midterm-review-report.
25	U.S. EPA. Mid-Term Evaluation of Greenhouse Gas Emissions Standards for Model Year 2022-2025 Light-Duty
Vehicles. U.S. Federal Register, Vol.83, No. 72, pp 16077-16087, April 13, 2018.
26	U.S. EPA and U.S. DOT/NHTSA. The Safer Affordable Fuel-Efficient (SAFE) Vehicles Rule for Model Years
2021-2026 Passenger Cars and Light Trucks. U.S. Federal Register, Vol.85, No. 84, pp 24174-25278, April 30,
2020.
27
Order, Executive. 13990. Protecting Public Health and the Environment and Restoring Science to Tackle the
Climate Crisis. U.S. Federal Register, Vol. 86., No. 14, pp 7037-7043, January 25, 2021.
28
U.S. EPA. California State Motor Vehicle Pollution Control Standards; Advanced Clean Car Program;
Reconsideration of a Previous Withdrawal of a Waiver of Preemption; Opportunity for Public Hearing and Public
Comment. U.S. Federal Register, Vol. 86, No. 80, pp 22421-22430, April 28, 2021.
1-16

-------
Chapter 2: Technology Feasibility, Effectiveness, Costs, and Lead-time
EPA is finalizing revised national greenhouse gas (GHG) emissions standards for passenger
cars and light trucks under section 202(a) of the Clean Air Act (CAA). Section 202(a) requires
EPA to establish standards for emissions of air pollutants from new motor vehicles which, in the
Administrator's judgment, cause or contribute to air pollution which may reasonably be
anticipated to endanger public health or welfare. The transport sector is currently the largest
source of anthropogenic GHG emissions in the U.S. There are technologically feasible to achieve
additional reductions for MY 2023 through MY 2026 light-duty vehicles at reasonable cost per
vehicle and without compromise to vehicle utility or safety. As in many prior EPA mobile source
rulemakings, the decision on what standards to set and on what implementation timeframe is
largely based on the availability, capability, and cost of the emissions control technology along
with the need for reductions of GHG and the benefits of doing so. This final rule will also
establish a path toward more significant reductions in the years following 2026.
2.1 Final Standards
As with the existing GHG standards, EPA is finalizing separate car and truck standards—that
is, vehicles defined as cars have one set of footprint-based curves, and vehicles defined as trucks
have a different set.1 Generally, passenger cars include cars and smaller cross-overs and SUVs,
while the truck category includes larger cross-overs and SUVs, minivans, and pickup trucks.
Because compliance is based on a sales-weighting of the full range of vehicles in a
manufacturer's car and truck fleets, the footprint based CO2 emission levels of specific vehicles
within the fleet are referred to as targets, rather than standards. In general, for a given footprint,
the CO2 g/mile target for trucks is higher than the target for a car with the same footprint. The
curves are defined mathematically in EPA's regulations by a family of piecewise linear functions
(with respect to vehicle footprint) that gradually and continually ramp down from the MY 2022
curves established in the SAFE rule. EPA's minimum and maximum footprint targets and the
corresponding cutpoints are provided below in Table 2-1 for MYs 2023-2026 along with the
slope and intercept defining the linear function for footprints falling between the minimum and
maximum footprint values. For footprints falling between the minimum and maximum, the
targets are calculated as follows: Slope x Footprint + Intercept = Target. Figure 2-1 and Figure
2-2 provide the existing MY 2021-2022 and final MY 2023-2026 footprint curves graphically for
both cars and light trucks, respectively.
Table 2-1: Final Footprint-based CO2 Standard Curve Coefficients

Car
Truck

2023
2024
2025
2026
2023
2024
2025
2026
MIN C02 (g/mi)
145.6
138.6
130.5
114.3
181.1
172.1
159.3
141.8
MAX C02 (g/mi)
199.1
189.5
179.4
160.9
312.1
296.5
277.4
254.4
Slope (g/mi/ft2)
3.56
3.39
3.26
3.11
3.97
3.77
3.58
3.41
Intercept (g/mi)
-0.4
-0.4
-3.2
-13.1
18.4
17.4
12.5
1.9
MIN footprint (ft2)
41
41
41
41
41
41
41
41
MAX footprint (ft2)
56
56
56
56
74
74
74
74
2-1

-------
230
210
190
a; 170
r\l
O
u 150
*E
130
03
oo 110
90
70
50
•2021
¦2022
2023
2024
•2025
•2026+
^HrMro^LntDr^oocno^HrMro^LntDr^oo
^^^^^^^^^LnLDLnLnLnLnLnLnLn
footprint (ft2)
Figure 2-1: CO2 vs Footprint Compliance Curves for Cars
350
300
250
E
03
00
E 200
150
100
•2021
¦2022
2023
2024
•2025
•2026+
50
-------
MY 2022 and MY 2023 curves is roughly double the gap between the curves for MYs 2024-
2026.
Figure 2-3 shows EPA's final standards, expressed as year-over-year fleetwide GHG
emissions targets (cars and trucks combined), projected through model year 2026 and beyond.
For comparison, the figure also shows the corresponding targets for the recent SAFE FRM and
the 2012 FRM. The final fleet targets start from the prior SAFE FRM targets for model year
2022, but ramp down considerably in model year 2023, nearly reaching the 2012 FRM targets
for that model year. The final fleet targets approximately parallel the downward slope of the
2012 FRM targets for model years 2023 and 2024, are approximately equivalent to the 2012
FRM in 2025 (the last year of the 2012 FRM), and then decrease at a more stringent downward
slope for one additional model year to model year 2026 (the last year of the SAFE FRM). As
with all EPA light-duty GHG rules, the standards would then remain in place at the same level
for all subsequent model years unless revised by a subsequent rulemaking. Table 2-2 presents
EPA's final standards presented in Figure 2-3, again in terms of the projected overall fleetwide
CCh-equivalent emission compliance target levels.
260
240
220
[aT
I 200
CM
O
u
180
160
140
2020 2021 2022 2023 2024 2025 2026 2027
Model Year
Figure 2-3: Final Fleet-Wide CCh-Equivalent g/mi Compliance Targets (solid black line), Compared to 2012
FRM, SAFE Rule, and Proposal.
• • • SAFE FRM
— »2012 FRM
•Proposal
Final Standards
2-3
-------
Table 2-2: Estimated Fleet-wide CO2 Target Levels Corresponding to the Final Standards
Model Year
Cars
Trucks
Fleet

CO2 (g/mile)
CO2 (g/mile)
CO2 (g/mile)
2023
166
233
202
2024
158
222
192
2025
149
207
179
2026 and later
132
187
161
2.1.1 Revised Final Compliance Incentives and Flexibilities
EPA is finalizing provisions for credit extension that are more targeted than those proposed.
EPA proposed to extend credit carry-forward for MY 2016-2020 credits to allow more flexibility
for manufacturers in using banked credits in MYs 2023-2026. Specifically, EPA proposed a two-
year extension of MY 2016 credits and a one-year extension of MY 2017-2020 credits. After
considering comments and further analyzing the need for extended credit life, EPA is adopting a
narrower approach for the final rule of only adopting a one-year credit life extension for MY
2017-2018 credits so they may be used in MYs 2023-2024, respectively. For details regarding
the averaging banking and trading program and credit carry forward provisions, please refer to
section II.A.4 within the Preamble for the Final Rule.
2.2 Light-duty Vehicle Technology Feasibility
2.2.1 Feasibility of the Revised Final Standards
Based upon the light-duty vehicle fleet compliance analysis summarized within Chapter 4 of
this RIA and the updated analytical results for the final rule in RIA Chapters 5, 6, and 10; and
consistent with the extensive public record established by EPA with its publication of the 2012
FRM, July 2016 Draft TAR, November 2016 Proposed Determination January 2017 Final
Determination, August 2021 NPRM; and taking into consideration the averaging, banking, and
trading provisions; the final MY 2023 and later light-duty GHG standards are feasible using
existing vehicle technologies that are already widely available within the current light-duty
vehicle fleet.
The feasibility of the revised standards is best understood within the context of the decade-
long light-duty vehicle GHG emissions reduction program in which the automotive industry has
innovated a wide range of GHG-reducing technologies. Over this time, the industry has had the
ability to plan for increasingly stringent GHG emissions requirements. The result has been the
widespread and continual introduction of new and improved GHG-reducing technologies across
the industry, many of which were in the early stages of development at the beginning of the
program in 2012. See Chapter 2.3 for a discussion of technological progression, status of
technology penetration, and Chapter 4.1.4 for our assessment of the continuing technology
penetration across the fleet.
The technological achievements already developed and increasing in application to vehicles
within the current new vehicle fleet (Chapter 2.3) will enable the industry to achieve the final
standards even without the development and implementation of additional technologies.
Compliance with the final standards, adjustment to the pace of technology penetration of existing
GHG reduction technologies, and adjustment to the management of both existing GHG credits
and particularly the generation of credits under the revised light-duty GHG program will occur
2-4
-------
within the full context of the revised incentives and flexibilities that will be available under the
Final Rule. As we discuss in Chapter 2.4, our assessment shows that a large portion of the
current fleet (MY 2021 vehicles), across a wide range of vehicle segments, already meets future
standards and that there are clear opportunities for automakers to focus their sales and marketing
on these more efficient products.
The multi-year nature of automotive design and engineering development also means that the
industry's product plans that were developed in response to the EPA's GHG standards finalized
in 2012 for MYs 2017-2025 has largely continued despite the relaxation of GHG standards under
SAFE that were promulgated in April 2020 with implementation beginning in MY 2021. This
can also be seen within the increased penetration of GHG reducing technologies (Chapter 2.3). In
previous comments on EPA's light-duty GHG and other light-duty vehicle programs, automakers
have broadly stated that they require approximately five years to design, develop, and produce a
new vehicle model. Thus, in most cases, vehicles that automakers intend to sell during the first
years of these revised MY 2023 and later GHG standards were already designed under the
original, and more stringent, GHG standards finalized in 2012 for those model years. At the time
of the proposed rule, the relaxed GHG standards under the SAFE rule had been in place for little
more than one year. During this time, the ability of the industry to commit to a change of plans to
take advantage of the SAFE rule's relaxed standards, especially for MYs 2023 and later, was
highly uncertain in light of pending litigation, and the automobile industry regularly expressed
concern over the uncertain future of the SAFE standards. In fact, due in part to this uncertainty,
five automakers voluntarily agreed to more stringent national emission reduction targets under
the California Framework Agreements.3 Therefore, based on the automakers' own past
comments regarding product plan development and the regulatory and litigation history of the
GHG standards since 2012, we believe that automakers continue to be largely on track in terms
of technological readiness within their product plans to meet the approximate trajectory of
increasingly stringent light-duty vehicle GHG standards initially promulgated in 2012.
Although we do not believe that automakers have significantly changed their product plans in
response to the SAFE final rule issued in 2020, any that may have would have done so relatively
recently and we would anticipate that their earlier product plans could be reinstated or adapted
with minimal change. It is important to note that we have considered the need for manufacturers
to transition from the SAFE standards (or the California Framework Agreement) to standards
that are similar in stringency to the 2012 standards and have structured the revised standards to
be less stringent than the 2012 standards for model years 2023 and 2024, and of comparable
stringency for model year 2025. EPA considers this an important aspect of its analysis because it
mitigates concerns about lead-time for manufacturers to meet the revised standards beginning
with the 2023 model year. We see no reason to expect that the major GHG-reducing technologies
that automakers already developed, increasingly introduced (see Chapter 2.3), or already planned
for near-term implementation, will not be available for MY 2023 and later vehicles. Thus, in
contrast to the situation that existed prior to EPA's adoption of the initial light-duty GHG
standards in the 2012 rule, automakers now have had the benefit of at least 8-9 years of planning
and development for increasing levels of GHG-reducing technologies in preparation for meeting
these revised standards.
Further support that the technologies needed to meet the standards do not need to be
developed and are already widely available and in use on vehicles can be found in the fact that
five vehicle manufacturers, representing nearly 30 percent of U.S. auto sales, agreed in 2019
2-5
-------
with the State of California that their nationwide fleets would meet GHG emission reduction
targets more stringent than the applicable EPA standards for MYs 2021 and 2022, and similar to
the final EPA standards for MYs 2022 and 2023.3 These voluntary actions by automakers speak
directly to the feasibility of meeting standards at least as stringent as those under the California
Framework. Thus, the California Framework voluntary targets were another consideration in our
development and assessment of the Final EPA light-duty vehicle GHG standards.
It is important to note that our conclusion that the revised program is technologically feasible
is based in part on a projection that the standards will be met largely with the kinds of advanced
light-duty vehicle engine technologies, transmission technologies, electric drive systems,
aerodynamics, tires, and vehicle mass reduction already in place in vehicles within today's fleet.
Our updated analysis projects that the final standards can be met with a fleet that achieves a
gradually increasing market share of EVs and PHEVs, approximately 7 percent in MY 2023 up
to about 17 percent in MY 2026 (see Chapter 4 and also Section III.C within the Preamble to this
Final Rule). While this represents an increasing penetration of zero-emission and near-zero
emission vehicles into the fleet during the 2023-2026 model years, we believe that the growth in
the projected rate of penetration is consistent with current trends and market forces. We believe
that the continuation of trends already underway, as exemplified in part by manufacturers' public
announcements about their plans to transition to electrified vehicles, as well as continuing
advancements in EV technology, support the feasibility of this level of EV and PHEV
penetration during the time period of the rule. Moreover, EPA is committed to encouraging the
rapid deployment of zero-emission vehicles, and we are finalizing compliance flexibilities and
incentives to support this transition (see Sections I.B.2 and II.B within the Preamble to this Final
Rule).
2.2.2 Alternatives to the Revised Standards
In addition to the revised standards, we analyzed both a more stringent and a less stringent
alternative. The less stringent alternative in this FRM is equivalent to the Proposal Standards and
uses the corresponding footprint-based car and truck standards curves (i.e, the curve coefficients)
from the NPRM. The Proposed Standards were supported in public comments by traditional
internal combustion engine equipped vehicle manufacturers. The Proposal as analyzed also
included incentive multipliers consistent with those in EPA's NPRM.
The more stringent alternative (Alternative 2 minus 10) used the coefficients from the more
stringent alternative (Alternative 2) in the NPRM with the additional increase in stringency of 10
g/mile in MY 2026. Alternative 2 minus 10 differs from the final standards only in MY 2023 and
2024. Because we wanted maximum stringency for this alternative, we did not include the final
rule's advanced technology multipliers and cumulative credit cap associated with those
multipliers for MYs 2022 and later.
The fleet average targets for the two scenarios compared to the revised standards are provided in
Table 2-3 below. As described in 2.3.3 and elsewhere, the potential for increased penetration of
ZEVs was considered as a factor in the feasibility of the final standards and Alternative 2 minus
10 standards.
2-6
-------
Table 2-3: Projected Fleet Average Target Levels for Revised Standards and Alternatives (CO2 grams/mile)*
Model Year
Final Projected
Targets
Proposal Projected
Targets
Alternative 2 minus
10 Projected
Targets
2021**
229
229
229
2022**
224
224
224
2023
202
202
198
2024
192
192
186
2025
179
182
180
2026
161
173
161
* Targets shown are modeled results and, therefore, reflect fleet projections impacted by the
underlying standards. For that reason, slight differences in targets may occur despite equality
of standards in a given year.
** SAFE rule targets included here for reference.
260
240
S
220
200
180
160
SAFE FRM
2012 FRM
-Proposal
Alternative 2 minus 10
-Final Standards
140
2020	2021	2022	2023	2024	2025	2026	2027
Model Year
Figure 2-4: Final Rule Fleet Average Targets Compared to the Proposal and Alternative 2 minus 10
As shown in Figure 2-4, the range of analyzed scenarios that we considered was fairly narrow,
with the revised final standard targets differing from the Proposal and Alternative 2 minus 10
targets in any given model year in 2023-2026 by 2 to 12 g/mile. We believe that this approach is
reasonable and appropriate considering the relatively short lead time for the revised standards,
our assessment of feasibility, the existing automaker commitments to meet the California
Framework (representing about 28 percent of the auto market), the standards adopted in the 2012
2-7
-------
rule, and the need to reduce GHG emissions. The analysis of costs and benefits of the Final
Standards, the Proposal and Alternative 2 minus 10 are summarized within Chapters 4, 5, 6, and
10.
The revised final standards, the Proposal, and the Alternative 2 minus 10 all incorporate year-
over-year increases in GHG stringency, with varying starting stringencies in MY 2023, varying
ending stringencies in MY 2026, and fairly linear increases in stringency between MY 2023 and
2025 that would essentially follow the same slope as the 2012 program. All three potential
programs would also, by MY 2026, result in standards at least as stringent or more stringent
when compared to the last year (MY 2025) of the 2012 program.
For the Proposal, the standards would have slightly less stringency than the 2012 rule for
model years 2023-2025 and higher stringency in model year 2026, resulting in a less stringent
program compared to the 2012 rule until MY 2026. Chapter 5.1.1.2 shows the associated lower
amount of GHG reductions achieved under the Proposal when compared to the final standards.
For Alternative 2 minus 10, the standards for model years 2023 through 2025 would match
the stringency level of the standards in the 2012 rule and would continue to increase in
stringency for one additional year in MY 2026. Consistent with EPA's previous discussions
regarding feasibility, compliance costs, and lead time, we believe that Alternative 2 minus 10 are
also technologically feasible.
2.3 Vehicle Technologies
For a summary of the effectiveness and cost of technologies used by EPA for modeling
compliance with the final standards, see Chapter 4.1 of this RIA. A complete summary of vehicle
technologies and associated GHG effectiveness for internal combustion engine technologies,
transmission technologies, vehicle electrification, aerodynamics, tires, and vehicle mass
reduction can be found within Chapter 2.2 of the Technical Support Document (TSD) for the
November 2016 Proposed Determination.4 We still believe this document to be a sound and
thorough examination of the available technologies and their GHG effectiveness for the
timeframe of this rulemaking. In fact, some vehicle manufacturers have recently made public
statements regarding their plans to discontinue the development of conventional, internal
combustion engine-based technologies to focus on the electrified vehicle technologies.5 In their
press release announcing their goal to be carbon neutral in 2040, GM stated that "The company
will also continue to increase fuel efficiency of its traditional internal combustion vehicles in
accordance with regional fuel economy and greenhouse gas regulations. Some of these initiatives
include fuel economy improvement technologies, such as Stop/Start, aerodynamic efficiency
enhancements, downsized boosted engines, more efficient transmissions and other vehicle
improvements, including mass reduction and lower rolling resistance tires."6 Although some
manufacturers have indicated a reduced focus on internal combustion engine (ICE) technologies,
EPA has continued its independent evaluation of advanced engine and transmission technologies
and has updated and improved our assessment of light-duty vehicle GHG emissions over the
intervening years since publication of the TSD.7 The results of these analyses have been
published in over a dozen peer-reviewed technical and journal
8,9,10,11,12,13,14,15,16,17,18,19,20,21,22
pdpcl
The percentage share of specific MY 2015 to MY 2020 engine and transmission technologies
are summarized from EPA Automotive Trends Report data in Table 2-4 and Table 2-5
2-8
-------
respectively.23 In MY 2020, hybrid electric vehicles (HEV) accounted for approximately 6.5
percent of vehicle sales, while plug-in electric hybrids (PHEV) and battery electric vehicles
(BEV) together comprised 4 percent of sales. Thus, powertrain electrification of all types has
increased more than 3-fold from MY 2015 to MY 2020. The pace of introduction of new EV
models is rapidly increasing. Nearly 100 pure electric EV models are expected to be introduced
in the United States by the end of 2024.24 The sales of vehicles with 12V start/stop systems have
increased from approximately 7 percent to approximately 42 percent between MY 2015 and MY
2020.
As of MY 2020, more than half of light-duty gasoline spark ignition engines now use direct
injection (GDI) and more than a third are turbocharged.21,25 Nearly half of all light-duty vehicles
have planetary automatic transmissions with 8 or more gear ratios, and a fourth are using
continuously variable transmissions (CVT). We anticipate that these GHG reducing technologies
will continue to increasingly penetrate the light-duty vehicle fleet for MYs 2023-2026.
Table 2-4: Production Share by Engine Technologies for MY 2015-2020

Powertrain Technologies
Engine Technologies
Model Year
Gasoli
ne
Gasolin
e
HEV
Diesel
PHEV
BEV
GDI
Port
Avg.
Displ.
(L)
HP
VVT
CD
Turbo
Stop/
Start
2015
95.9%
2.4%
0.9%
0.2%
0.5%
41.9%
56.7%
2.90
229
97.2%
10.5%
15.7%
7.1%
2016
96.9%
1.8%
0.5%
0.3%
0.5%
48.0%
51.0%
2.85
230
98.0%
10.4%
19.9%
9.6%
2017
96.1%
2.3%
0.3%
0.8%
0.6%
49.7%
49.4%
2.85
234
98.1%
11.9%
23.4%
17.8%
2018
95.1%
2.3%
0.4%
0.8%
1.4%
50.2%
48.0%
2.82
241
96.4%
12.5%
30.0%
29.8%
2019
94.4%
3.8%
0.1%
0.5%
1.2%
52.9%
45.7%
2.85
245
97.2%
14.9%
30.0%
36.9%
2020
(prelim)
88.5%
6.5%
1.0%
0.7%
3.3%
55.3%
40.3%
2.75
247
94.0%
13.8%
35.3%
42.2%
23
Note: Adapted from the 2020 EPA Automotive Trends Report.
Table 2-5: Production Share by Transmission Technologies for MYs 2015-2020
Model Year
Manual
Automatic
Automatic
CVT
CVT
4 Gears
5
6
7
8+
Average


with
without
(Hybrid)
(Non-
Or
Gears
Gears
Gears
Gears
No. of


Lockup
Lockup

Hybrid)
Fewer




Gears
2015
2.6%
72.3%
1.4%
2.2%
21.5%
1.5%
4.5%
54.2%
3.1%
13.0%
5.9
2016
2.2%
72.3%
2.6%
1.7%
21.2%
1.1%
3.0%
54.9%
2.9%
15.3%
6.0
2017
2.1%
71.5%
2.6%
1.9%
21.8%
1.0%
2.4%
49.0%
3.4%
20.5%
6.1
2018
1.6%
72.8%
3.2%
1.7%
20.6%
1.9%
2.0%
37.6%
3.7%
32.5%
6.4
2019
1.4%
72.1%
2.4%
2.2%
21.9%
1.5%
1.6%
26.1%
2.6%
44.0%
6.6
2020
(prelim)
1.5%
66.1%
4.4%
3.1%
25.0%
3.4%
1.3%
15.8%
2.4%
49.0%
6.6
23
Note: Adapted from the 2020 EPA Automotive Trends Report.
a A technical assessment of the particulate matter (PM) emissions impacts of MY 2020-2021 light-duty vehicles
using engines equipped with gasoline direct injection (GDI) and port fuel injection is included within a memo to the
docket for this final rule.
2-9
-------
2.3.1 Recent Advances in Internal Combustion Engines
The Automotive Trends Report does not separately track the introduction of HEV and non-
HEV applications of Atkinson Cycle and Miller Cycle engines, however their application has
been increasing over the past five years. Atkinson Cycle and Miller Cycle engines represent
technologies that improve efficiency via use of increased expansion when compared to
convention (Otto cycle) spark ignition engines. Although Atkinson and Miller Cycles are
sometimes used interchangeably, EPA's use of the nomenclature refers specifically to either
naturally-aspirated (Atkinson) or turbocharged (Miller) implementations. Recent
implementations also include use of fast, wide-range of authority camshaft phasing to allow
variation of effective compression ratio for load control and additional reduction of pumping
losses. Most implementations over the last six years use gasoline direct injection (GDI) for
additional knock mitigation.b For additional information on these technologies, see Chapter
2.2.1.2 "Descriptions of Technologies and Key Developments since the FRM" within the
Technical Support Document for the November 2016 Proposed Determination (2016 TSD).26
Atkinson Cycle engines have been common in HEV applications for more than two decades.
More recently, Toyota, Mazda, and Hyundai/Kia have been expanding the use of these engines
in non-HEV applications to reduce fuel consumption and comply with GHG emissions
standards. Since the publication of the 2016 TSD, there has also been a broader range of product
introductions with Atkinson Cycle engines combined with gasoline direct injection (GDI) and
either cylinder deactivation or cooled EGR. Mazda introduced fixed cylinder deactivation0 on the
base 2.5L Atkinsons Cycle engine in the MY 2018 CX-5 CUV and Mazda 6 passenger car. It
was also introduced in the MY 2019 Mazda 3. Based on comparisons of certification data for
comparable chassis and trim levels, Mazda's implementation of fixed cylinder deactivation
provides an incremental effectiveness of approximately 2 percent beyond that of a 4-cylinder
Atkinson Cycle engine without fixed cylinder deactivation.
Atkinson Cycle with cooled EGR has been applied to a broad range of both HEV and non-
HEV passenger cars and crossover utility vehicles (CUV). Examples include the Toyota's
"Dynamic Force" range of engines added as part of the Toyota New Global Architecture
(TNGA).27'28'29'30'31 Cooled EGR is used to reduce pumping losses and to mitigate combustion
knock. These include the following Toyota engines: the M15A-FKS, M20A-FKS, and A25A-
FKS non-HEV engines; and the M15A-FXE, M20A-FXS, and A25A-FXS HEV-specific engines
used in the Toyota Corolla, Camry, Aval on, C-HR, RAV4, Highlander, Lexus ES and Lexus
UX. In 2018, EPA conducted engine dynamometer benchmark testing of the Toyota 2.5L A25A-
FKS engine with Atkinson Cycle and cooled EGR.17 During testing on Federal Tier 2
certification fuel, the Toyota A25A-FKS engine demonstrated a peak brake thermal efficiency
(BTE) of approximately 40 percent, the highest published BTE for a production, non-HEV
engine. This represents a significant improvement over the peak BTE (typically 35-37 percent)
of the naturally aspirated GDI engines that make up a majority of MY 2020 vehicle fleet.
Atkinson Cycle engines were estimated to have GHG effectiveness of approximately 3.2 to 3.8
percent relative to over otherwise comparable naturally-aspirated GDI engines in non-HEV
b Knock is an abnormal and potentially damaging form of combustion characterized by a very high rate of increase
in cylinder pressure and high peak cylinder pressure.
c Fixed cylinder deactivation disables a fixed number of engine cylinders to reduce pumping losses at light load.
2-10
-------
applications. EPA estimates that the addition of cooled EGR to an Atkinson Cycle engine further
reduces 2-cycle GHG emissions by an additional 4.4 percent over Atkinson Cycle alone.
Both engine-dynamometer developmental work and benchmarking of production engines by
EPA identified synergies between the use of fixed cylinder deactivation and cooled EGR on
Atkinson Cycle engines when used in non-HEV applications.32'33'34'17 Both EPA and other
researchers have also identified synergies between the use of dynamic cylinder deactivation and
cooled EGR on Atkinson Cycle engines.17'35 EPA estimates that the addition of either fixed
cylinder deactivation or dynamic cylinder deactivation11 to an Atkinson Cycle engine with cooled
EGR would provide an additional 2.3 percent or 7.9 percent reduction in 2-cycle GHG
emissions, respectively.17
VW now offers EA888-3B 2.0L Miller Cycle engine as the base engine in the Passat and
Arteon passenger cars and the Atlas and Tiguan CUVs. The MY 2022 Taos CUV will use the
EA211 1.5L evo Miller Cycle engine as the base engine, which has a peak brake thermal
efficiency of 38.1 percent.36 A hybrid-specific version of this engine is under development by
VW. When equipped with cooled-EGR and a variable-geometry turbo, it demonstrated a peak
BTE of 41.5 percent.36
2.3.2 Changes to Engine Technologies Represented in the Analysis for the Final Rule
Analytical revisions to the modeling of light-duty vehicle compliance with the final standards
and the resulting GHG emissions and vehicle technology costs are summarized within Chapter
4.1.
Within EPA's analysis, the different levels of HCR represent the following:
•	HCRO: Atkinson Cycle with GDI and a geometric compression ratio of 13:1.
o Examples: 2012 - 2016 Mazda vehicles with the SKYACTIV-G engine (which
we benchmarked)
•	HCR1: The addition of either cooled EGR or fixed cylinder deactivation (CDA) to
HCRO.
o Examples with cooled EGR: Many 2018 and nearly all 2019 and later Toyota
vehicles with 4-cylinder engines (we benchmarked the Toyota Camry);
o Examples with fixed CD A: 2017 and later Mazda vehicles with the
SKYACTIVE-G engine
•	HCR2: Atkinson Cycle with GDI, cooled EGR, and dynamic (individual cylinder)
cylinder deactivation, and a geometric compression ratio of 13:1.17
o For HEV applications, HCR2 represents the application of GDI, cooled EGR,
higher compression and expansion ratio, and the use of a dedicated hybrid
electric/engine powertrain strategy.6
d Dynamic cylinder deactivation is a newer, more capable system than fixed cylinder deactivation. Any number of
cylinders can be deactivated or activated on a cycle resolved basis. The first production examples became available
on GM full-frame trucks in MY 2019.
e Dedicated hybrid engines combine an engine and electric drive within a powertrain and calibrated in a synergistic
manner that increases engine efficiency and avoids areas of engine operation prone to knock and/or low-speed
preignition.
2-11
-------
The restriction within the analysis of all HCR technologies to naturally aspirated engines with
cylinder counts of 6 or less during compliance modeling was a means of restricting Atkinson
Cycle from application to trucks and other applications having a specific need for additional
torque reserve (e.g., trailer towing or high payload applications).
A change made in the analysis for the final rule relative to the proposal was to only include
HCR2 as part of a sensitivity analysis for MY2025 and later vehicles. The individual
technologies represented by HCR2 are all currently available within the current light-duty
vehicle fleet, however the specific combination of technologies represented by HCR2 are not yet
available in production light-duty vehicles. Thus, while we believe HCR2 to be technologically
feasible, we made the decision to limit application of HCR2 to a sensitivity analysis applied to
MY2025 and later as a conservative approach with respect to compliance with the light-duty
GHG standards. Some manufacturers may choose to pursue a relatively low cost internal
combustion engine technology like HCR2, however we believe that many manufactures will
choose instead to instead focus near-term and future investment on powertrain electrification.
Thus, we are showing 2 paths to compliance - a sensitivity analysis with an HCR2 compliance
path and modeled compliance without HCR2. For more information on use of HCR2 within the
analyses for the final rule, please refer to Chapters 4.1.1.3 and 4.1.5.1.
2.3.3 Vehicle Electrification
While we anticipate that the revised standards will be met primarily through the continued
penetration of conventional powertrain (e.g., internal combustion engine, transmission)
improvements and road-load reductions as outlined previously within the draft TAR,37 the PD
TSD,38 and in the previous sections of this chapter, we anticipate that the design of a future,
longer-term program beyond 2026 will further incorporate accelerating advances in zero-
emission technologies.
A proliferation of recent announcements from automakers signals a rapidly growing shift in
investment away from internal-combustion technologies and toward high levels of
electrification. These automaker announcements are supported by continued advances in
automotive electrification technologies, and further driven by the need to compete in a global
market as other countries implement aggressive zero-emission transportation policies.
For example, in January 2021, General Motors announced plans to become carbon neutral by
2040, including an effort to shift its light-duty vehicles entirely to zero-emissions by 2035.39 In
March 2021, Volvo announced plans to make only electric cars by 2030,40 and Volkswagen
announced that it expects half of its U.S. vehicle sales will be all-electric by 2030.41 In April
2021, Honda announced a full electrification plan to take effect by 2040, with 40 percent of its
North American vehicle sales expected to be fully electric or fuel cell vehicles by 2030, 80
percent by 2035 and 100 percent by 2040.42 In May 2021, Ford announced that they expect 40
percent of their global light-duty vehicle sales will be all-electric by 2030.43 In June 2021, Fiat
announced a move to all electric vehicles by 2030,44 and in July 2021 its parent corporation
Stellantis announced an intensified focus on electrification across all of its brands.45 Also in July
2021, Mercedes-Benz announced that all of its new architectures would be electric-only from
2025, with plans to become ready to go all-electric by 2030 where possible.46 In September
2021, Toyota announced large new investments in battery production and development to
support an increasing focus on electrification.47 On August 5, 2021, in conjunction with the
announcement of Executive Order 14037, many of these automakers, as well as the United Auto
2-12
-------
Workers and the Alliance for Automotive Innovation, expressed continued commitment to these
announcements and support for the goal of achieving 40 to 50 percent sales of zero emissions
vehicles by 2030.
These announcements, and others like them, continue a pattern over the past several years in
which many manufacturers have taken steps to aggressively pursue zero-emission technologies,
introduce a wide range of ZEV models, and reduce their reliance on the internal-combustion
engine in various markets around the globe, including the U.S.48'49 These goals and investments
have been coupled with a continuing increase in the market penetration of new zero-emission
vehicles (3.6 percent of new U.S. light-duty vehicle sales so far in calendar year 2021,50 and
projected to be 4.1 percent of production in MY 2021 up from 2.2 percent of production in MY
2020),51 as well as a rapidly increasing diversity of plug-in vehicle models in the U.S.52 For
example, the number of battery electric vehicle (BEV) and plug-in hybrid electric vehicle
(PHEV) models available for sale in the U.S. has more than doubled from about 24 in MY 2015
to about 60 in MY 2021, with offerings in a growing range of vehicle segments.53'54 Recent
model announcements indicate that this number will increase to more than 80 models by MY
2023, with many more expected to reach production before the end of the decade.55 Market
forecasts suggest a combined EV and PHEV (including range-extended EV) sales share of
approximately 16 percent to 24 percent in the U.S. by 2026,56>57 which compare favorably to the
ZEV projection for this final rule. Many of the ZEVs already on the market today cost less to
drive than conventional vehicles,58'59 offer improved performance and handling,60 and can be
charged at a growing network of public chargers as well as at home.61
Recent BEV product announcements also include a growing number of dedicated battery
electric vehicle platforms, such as the GM BEV2 light-duty vehicle (LDV) and BEV3 light-duty-
truck (LDT) platforms, the Tesla Model 3/Model Y LDV and crossover utility vehicle (CUV)
platform, the VW MEB LDV and CUV platform, and the Hyundai E-GMP LDV and CUV
platform.62 Dedicated BEV platforms eliminate provisions for internal combustion engine (ICE)
powertrain, exhaust emissions, evaporative emissions, and fuel systems that would otherwise
need to be accommodated on platforms that are shared between BEV, PHEV, HEV, and
conventional ICE vehicle models. This dedicated BEV platform approach typically allows
integration of the battery pack entirely within the vehicle floor structure, reduces vehicle weight,
reduces manufacturing costs, increases available passenger and cargo volume, and in some cases,
has the battery pack integrated as part of the vehicle's crash mitigation structure.
An increasing number of global jurisdictions and U.S. states are planning to take actions to
shift the light-duty fleet toward zero-emissions technology. In 2020, California announced an
intention to require increasing volumes of ZEVs to meet the goal that, by 2035, all new light-
duty vehicles sold in the state be ZEVs.63 New York has adopted similar targets and
requirements to take effect by 2035,64>65 with Massachusetts poised to follow.66 Several other
states may adopt similar provisions by 2050 as members of the International Zero-Emission
Vehicle Alliance.67 Globally, at least 12 countries, as well as numerous local jurisdictions, have
announced similar goals to shift all new passenger car sales to ZEVs in the coming years,
including Norway (2025), the Netherlands, Denmark, Iceland, Ireland, Sweden, and Slovenia
(2030), Canada and United Kingdom (2035), France and Spain (2040) and Costa Rica
(2050).68>69 Together, these countries represent approximately 13 percent of the global market for
passenger cars, in addition to that represented by the aforementioned U.S. states and other global
jurisdictions.70
2-13
-------
2.3.4 Automotive Li-ion Battery Costs
In response to numerous stakeholder comments (see Preamble III. A), EPA reviewed the
battery costs used in the SAFE rulemaking, which had been carried over to the analysis for the
proposal. We considered the inputs that had previously been used to derive the costs, and
compared those costs to estimates we had derived in previous and ongoing analyses and to the
current and expected future costs of batteries as widely reported in the trade and academic
literature. We concluded that the battery costs used in the proposal were significantly higher than
indicated by this evidence, and that the likely effect of using an updated set of assumptions
would be more in agreement with the emerging consensus on the level and direction of battery
costs within the industry.
Based on an assessment of the effect of using updated inputs in place of those used in the
SAFE rulemaking, we found technical justification for reducing battery costs by approximately
25 percent. Details on the technical basis for this change can be found in Section 4.1.1.2 of this
RIA.
We also considered the effect of this reduction on the projected battery costs for future years
beyond the time frame of the rule. Applying the existing learning curve to the downward
adjusted costs past the time frame of the rule would produce costs gradually declining to below
$80 per kWh (for an example 60 kWh battery) in the mid-2030s and to about $75/kWh by the
mid-2040s. EPA is currently uncertain about the potential for battery costs to reach this level due
in part to uncertainties about the effect of increased demand for critical minerals and other
factors, which we also received comment on, and also because our current battery modeling tools
such as BatPaC 4.0 are unable to generate costs at these levels using inputs that can reasonably
be validated. Due to the widely acknowledged uncertainty of quantitatively projecting declines in
battery costs far into the future, and particularly in the context of the downwardly adjusted
battery costs, we chose to flatten the rate of learning past 2029 so as to prevent future costs from
declining below $90 per kWh for a 60 kWh battery, a level that we can technically validate at
this time. More information on the technical basis for this change can be found in Section 4.1.1.2
of this RIA.
We believe that holding learning constant after 2029 is likely a conservative assumption, as
we continue to expect that some level of continued learning will occur beyond 2029 but there is
uncertainty at this point on what the appropriate level of learning would be. Thus, our battery
cost estimates beyond 2029 in this final rulemaking may be conservatively high. EPA continues
to study the potential for cost reductions in batteries during and after the time frame of the rule.
For example, we expect that pending updates to the ANL BatPaC model, as well as collection of
emerging data on forecasts for future mineral prices and production capacity, will make it
possible to characterize the rate of decline in battery costs that we continue to believe will occur
from 2030 to 2050, as well as trends in costs in the nearer term, and we will incorporate this
information in the subsequent rulemaking for MYs 2027 and beyond. For more discussion please
see Section 4.1.1.2 of this RIA.
2.4 Analysis of Manufacturers Generation and Use of GHG Credit
EPA believes that the multi-year nature of auto design and development means that the
industry's product plans originally developed in response to the EPA's 2012 GHG standards
rulemaking for MYs 2017-25 have largely continued notwithstanding the SAFE rule that was
2-14
-------
promulgated in April 2020, including relaxed standards beginning in MY 2021. Thus, in most
cases, the vehicles that automakers will be producing during the first years of the Final MY
2023-26 program were already designed under the original, more stringent GHG standards for
those model years finalized in 2012. Manufacturers are also already demonstrating the ability to
comply with the Final 2023 model year standards with many vehicles currently for sale.
For the Final Rule, EPA performed an analysis of 2021 model year vehicles to assess how
changes in sales mix could help facilitate vehicle manufacturer compliance to more stringent
standards. This analysis examined certification and projected sales data for 2021 model year
vehicles. EPA assumed that manufacturers continue to utilize credits for off-cycle technologies,
as well as A/C credits for reduced refrigerant leakage and improved efficiency. The level of off-
cycle credits was based on average manufacturer's MY 2019 off-cycle credits for cars and trucks,
respectively (so it does not reflect the Final Rule's increased cap to 15 g/mi of menu off-cycle
credits). EPA applied the industry average of 19 g/mi and 24 g/mi of total A/C credits for car and
truck models, respectively, to each manufacturer. Table 2-6 and Table 2-7 show the availability
of "credit generators" (the number of unique vehicle models that outperform their revised
individual footprint-based standard for 2023 model year), grouped by market segment. The
smallest market segments, by total sales volume, are shaded in gray and collectively represent
only about 5 percent of all sales. Projected performance is based on actual 2021 tailpipe CO2
emissions and adjusting for assumed A/C and off-cycle credits.
The analysis accounted for the various trim levels by manufacturers, as there are 1370 unique
vehicle model types in the 2021 model year. Of those 1370 unique vehicles, 216 models (over 16
percent of all models sold) already outperform the revised 2023 standards. 125 of these models
are advanced gasoline or hybrid vehicles while an additional 91 models are plug-in hybrids or
battery electric vehicles.
Table 2-6: Distribution of 2021 MY Vehicle Models and Number of Vehicles Which Generate Credits vs. 2023
MY Standards (All Vehicles)
Vehicle Category
Total
Models
Credit
Generators
Mkt Segment
% of 2021 Sales
Minicompact Cars
35
1
<)"„
Subcompact Cars
1 l<>
'J
2"..
Compact Cars
1 l(.
15

Two Seaters
(4
0
<)"„
Midsize Cars
158
28
13%
Large Cars
87
21
5%

0
i".,
Small SUVs
140
32
28%
Standard SUVs
288
34
25%
Small Pick-up Trucks
40
0

Standard Pick-up Trucks
256
61
13%
Totals
1370
216

Gray shading denotes niche vehicle segments at or below 3 percent of total sales
2-15
-------
Table 2-7: Distribution of 2021 MY Vehicle Models and Number of Vehicles Which Generate Credits vs. 2023
MY Standards (Gasoline ICE and Hybrid Vehicles)
Vehicle Category
Total
Models
Credit
Generators
Mkt Segment
% of 2021 Sales
Minicompact Cars
"4
(i
()"„
Subcompact Cars
1 in
(i

Compact Cars
In"
(.

Two Seaters

(i
()"„
Midsize Cars
143
13
13%
Large Cars
70
7
5%
Small Station Wagons
22
3
3%
Midsize Station Wagons
i:
(i
<)"„
Minivans
-
:

Vans
it.
(i
I-.,
Small SUVs
i ^i
23
28%
Standard SUVs
:<4
10
25%
Small Pick-up Trucks
4ii
(i

Standard Pick-up Trucks
25b
61
13%
Totals
1275
125

Gray shading denotes niche vehicle segments at or below 3 percent of total sales
Some niche market segments (shaded in gray within Table 2-6 and Table 2-7) including the
smallest vehicles (minicompact and subcompact cars), two-seaters, and small pickup trucks -
show few or no credit-generating models. However, credit-generators are currently available to
manufacturers in market segments that represent nearly 95 percent of the total sales volume.
Using the same analytical approach, these 2021 vehicle models offer additional credits and
more opportunities against the 2022 model year standards (Table 2-8). It is evident that
manufacturers are already well-positioned to earn significant credits against the 2022 model year
standards (and thus, also against the 2021 standards) with their 2021 vehicles. These credits can
be banked to provide margin for later years as a potential compliance strategy.
Table 2-8: Distribution of 2021 MY Vehicle Models and Number of Vehicles Which Generate Credits vs. 2022
MY Standards (All Vehicles)
Vehicle Category
Total Models
Credit Generators
Market Segment
% of 2021 Sales
Minicompact Cars
35
1
<)"„
Subcompact Cars
1 l«>


Compact Cars
1 l(.
25

Two Seaters
(4
u
<)"„
Midsize Cars
158
40
13%
Large Cars
87
29
5%
Small Station Wagons
31
14
3%
Midsize Station Wagons
i:
u
<)"„
Minivans
s
5

Vans
i(.
S
i".,
Small SUVs
140
52
28%
Standard SUVs
288
61
25%
Small Pick-up Trucks
40
0
v.,
Standard Pick-up Trucks
15b
92
13%
Totals
1370
336

Gray shading denotes niche vehicle segments at or below 3 percent of total sales
2-16
-------
References for Chapter 2
1	49 CFR Part 523.
2	77 FR 62781.
3	California Air Resources Board. Framework Agreements on Clean Cars. August 17, 2020. Last accessed on the
Internet on 5/25/2021 at the following URL: https://ww2.arb.ca.gov/sites/default/files/2020-08/clean-car-
framework-documents-all-bmw-ford-honda-volvo-vw.pdf
4	U.S. EPA. Proposed Determination on the Appropriateness of the Model Year 2022-2025 Light-Duty Vehicle
Greenhouse Gas Emissions Standards under the Midterm Evaluation: Technical Support Document. Chapter 2.2 -
State of Technology and Advancements since the 2012 Final Rule. EPA-420-R-16-021, November 2016.
5	Johannsen, F. VW brand will phase out combustion engines, CEO says. Automotive News Europe, March 21,
2021. Accessed on the Internet on 7/28/2021 at the following URL: https://europe.autonews.com/automakers/vw-
brand-will-phase-out-combustion-engines-ceo-says
6	General Motors, the Largest U.S. Automaker, Plans to be Carbon Neutral by 2040;
https://media.gm.com/media/us/en/gm/home.detail.html/content/Pages/news/us/en/2021/jan/0128-carbon.html
7	U.S. EPA. Benchmarking Advanced Low Emission Light-Duty Vehicle Technology. Last accessed on the Internet
on 5/26/2021 at the following URL: https://www.epa.gov/vehicle-and-fuel-emissions-testing/benchmarking-
advanced-low-emission-light-duty-vehicle-technology
8	Conway, G., Robertson, D., Chadwell, C., McDonald, J., Kargul, J., Barba, D., & Stuhldreher, M. (2018).
Evaluation of Emerging Technologies on a 1.6 L Turbocharged GDI Engine (No. 2018-01-1423). SAE Technical
Paper.
9	Dekraker, P., Barba, D., Moskalik, A., & Butters, K. (2018). Constructing engine maps for full vehicle simulation
modeling (No. 2018-01-1412). SAE Technical Paper.
10	Moskalik, A., Bolon, K., Newman, K., & Cherry, J. (2018). Representing GHG Reduction Technologies in the
Future Fleet with Full Vehicle Simulation. SAE International journal of fuels and lubricants, 11(4), 469.
11	Bolon, K., Moskalik, A., Newman, K., Hula, A., Neam, A., & Mikkelsen, B. (2018). Characterization of GHG
Reduction Technologies in the Existing Fleet (No. 2018-01-1268). SAE Technical Paper.
12	Lee, S., Cherry, J., Safoutin, M., McDonald, J., & Olechiw, M. (2018). Modeling and Validation of 48V Mild
Hybrid Lithium-Ion Battery Pack. SAE International Journal of Alternative Powertrains, 7(3), 273-288.
13	Lee, S., Cherry, J., Safoutin, M., Neam, A., McDonald, J., & Newman, K. (2018). Modeling and Controls
Development of 48 V Mild Hybrid Electric Vehicles (No. 2018-01-0413). SAE Technical Paper.
14	Stuhldreher, M., Kargul, J., Barba, D., McDonald, J., Bohac, S., Dekraker, P., & Moskalik, A. (2018).
Benchmarking a 2016 Honda Civic 1.5-liter L15B7 turbocharged engine and evaluating the future efficiency
potential of turbocharged engines. SAE International journal of engines, 11(6), 1273.
15	Robertson, D., Conway, G., Chadwell, C., McDonald, J., Barba, D., Stuhldreher, M., & Birckett, A.
(2018). Predictive GT-Power Simulation for VNT Matching on a 1.6 L Turbocharged GDI Engine (No. 2018-01-
0161). SAE Technical Paper.
16	Smith, L., Smith, I., Hotz, S., & Stuhldreher, M. (2018). Selective Interrupt and Control: An Open ECU
Alternative (No. 2018-01-0127). SAE Technical Paper.
17	Kargul, J., Stuhldreher, M., Barba, D., Schenk, C., Bohac, S., McDonald, J., & Dekraker, P. (2019).
Benchmarking a 2018 Toyota camry 2.5-liter atkinson cycle engine with cooled-EGR. SAE International journal of
advances and current practices in mobility, 1(2), 601.
18	Wang, Y., Conway, G., McDonald, J., & Birckett, A. (2019). Predictive GT-power simulation for VNT matching
to EIVC strategy on a 1.6 L turbocharged GDI engine (No. 2019-01-0192). SAE Technical Paper.
19	Moskalik, A., & Newman, K. (2020). Assessment of Changing Relationships between Vehicle Fuel Consumption
and Acceleration Performance (No. 2020-01-5067). SAE Technical Paper.
2-17
-------
20	Moskalik, A., Stuhldreher, M., & Kargul, J. (2020). Benchmarking a 2018 Toyota Camry UB80E Eight-Speed
Automatic Transmission (No. 2020-01-1286). SAE Technical Paper.
21	Moskalik, A. (2020). Using Transmission Data to Isolate Individual Losses in Coastdown Road Load
Coefficients. SAE International Journal of Advances and Current Practices in Mobility, 2(2020-01-1064), 2156-
2171.
22	Lee, S., Fulper, C. R., Cullen, D., McDonald, J., Fernandez, A., Doorlag, M. H.,... & Olechiw, M. (2020). On-
Road Portable Emission Measurement Systems Test Data Analysis and Light-Duty Vehicle In-Use Emissions
Development. SAE International Journal of Electrified Vehicles, 9(2), 111.
23	U.S. EPA. The 2020 EPA Automotive Trends Report - Greenhouse Gas Emissions, Fuel Economy, and
Technology since 1975. EPA-420-R-21-003, January 2021.
24	Consumer Reports. "Hot New Electric Cars That Are Coming Soon - Automakers are offering new models in all
shapes and sizes for 2021 and beyond." May 20, 2021. Last accessed on the Internet on 5/26/2021 at the following
URL: https ://www.consumerreports.org/hybrids-evs/hot-new-electric-cars-are-coming-soon/
25	McDonald, J. (2021) Particulate Matter Emissions from Light-duty Vehicles Equipped with Port Fuel Injection
and Gasoline Direct Injection. Memo to the Docket.
26	U.S. EPA. Proposed Determination on the Appropriateness of the Model Year 2022-2025 Light-Duty Vehicle
Greenhouse Gas Emissions Standards under the Midterm Evaluation: Technical Support Document. EPA-420-R-16-
021. November 2016.
27	Toyota Motor Corporation. Toyota Develops TNGA-based Powertrain Units for Smooth, Responsive, 'As
Desired' Driving. December 06, 2016. Last accessed on the Internet on 5/26/2021 at the following URL:
https://pressroom.toyota.com/toyota-tnga-powertrain-responsive-driving/
oo
Itabashi, S., Murase, E., Tanaka, H., Yamaguchi, M., & Muraguchi, T. (2017). New combustion and powertrain
control technologies for fun-to-drive dynamic performance and better fuel economy (No. 2017-01-0589). SAE
Technical Paper.
29 Hakariya, M., Toda, T., & Sakai, M. (2017). The new Toyota inline 4-cylinder 2.5 L gasoline engine (No. 2017-
01-1021). SAE Technical Paper.
Toyota Motor Corporation. Toyota Announces New Powertrain Units Based on TNGA. February 26, 2018. Last
accessed on the Internet on 5/26/2021 at the following URL:
https://global.toyota/en/newsroom/corporate/21179861 .html
31 Yamaji, K., Tomimatsu, M., Takagi, I., Higuchi, A., Yoshida, T., & Murase, E. (2018). New 2.0 L 14 gasoline
direct injection engine with Toyota new global architecture concept (No. 2018-01-0370). SAE Technical Paper.
U.S. EPA. Proposed Determination on the Appropriateness of the Model Year 2022-2025 Light-Duty Vehicle
Greenhouse Gas Emissions Standards under the Midterm Evaluation: Technical Support Document. Chapter
2.3.4.1.8.1 - Effectiveness Data Used and Basis for Assumptions. EPA-420-R-16-021, November 2016.
33	Lee, S., Schenk, C., & McDonald, J. (2016). Air Flow Optimization and Calibration in High-Compression-Ratio
Naturally Aspirated SI Engines with Cooled-EGR (No. 2016-01-0565). SAE Technical Paper.
34	Schenk, C., & Dekraker, P. (2017). Potential fuel economy improvements from the implementation of cEGR and
CDA on an Atkinson Cycle Engine (No. 2017-01-1016). SAE Technical Paper.
35	Bowyer, S., Ortiz-Soto, E., Younkins, M., & VENKADASAMY, V. (2021). Evaluation of New High Efficiency
Engine Concept with Atkinson Cycle, Cooled EGR and Dynamic Skip Fire (No. 2021-01-0459). SAE Technical
Paper.
36	Brannys, S., Gehrke, S., Hoffmeyer, H., Hentschel, L., Blumenroder, K., Helbing, C., & Dinkelacker, F. (2019,
October). Maximum efficiency concept of a 1.5 1TSI evo for future hybrid powertrains. In 28th Aachen Colloquium
Automobile and Engine Technology.
37	U.S. EPA, CA-EPA/ARB, U.S. DOT/NHTSA. Draft Technical Assessment Report: Midterm Evaluaton of Light-
Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average Fuel Economy Standards for Model
Years 2022-2025. EPA-420-D-16-900, July 2016.
2-18
-------
TO
U.S. EPA Proposed Determination on the Appropriateness of the Model Year 2022-2025 Light-Duty Vehicle
Greenhouse Gas Emissions Standards under the Midterm Evaluation. EPA-420-R-16-020, November 2016.
39	See RIA Chapter 2 Endnote No.8.
40	Volvo Car Group, "Volvo Cars to be fully electric by 2030," Press Release, March 2, 2021.
41	Volkswagen Newsroom, "Strategy update at Volkswagen: The transformation to electromobility was only the
beginning," March 5, 2021. Accessed June 15, 2021 at https://www.volkswagen-newsroom.com/en/stories/strategy-
update-at-volkswagen-the-transformation-to-electromobility-was-only-the-beginning-6875
42	Honda News Room, "Summary of Honda Global CEO Inaugural Press Conference," April 23, 2021. Accessed
June 15, 2021 at https://global.honda/newsroom/news/2021/c210423eng.html
43	Ford Motor Company, "Superior Value From EVs, Commercial Business, Connected Services is Strategic Focus
of Today's 'Delivering Ford+' Capital Markets Day," Press Release, May 26, 2021.
44	Stellantis, "World Environment Day 2021 - Comparing Visions: Olivier Francois and Stefano Boeri, in
Conversation to Rewrite the Future of Cities," Press Release, June 4, 2021.
45	Stellantis, "Stellantis Intensifies Electrification While Targeting Sustainable Double-Digit Adjusted Operating
Income Margins in the Mid-Term," Press Release, July 8, 2021.
46	Mercedes-Benz, "Mercedes-Benz prepares to go all-electric," Press Release, July 22, 2021.
47	Toyota, "Video: Media briefing & Investors briefing on batteries and carbon neutrality" (transcript), September 7,
2021. Accessed on September 16, 2021 at https://global.toyota/en/newsroom/corporate/35971839.html#presentation
Environmental Defense Fund and M.J. Bradley & Associates, "Electric Vehicle Market Status - Update,
Manufacturer Commitments to Future Electric Mobility in the U.S. and Worldwide," April 2021.
49	International Council on Clean Transportation, "The end of the road? An overview of combustion-engine car
phase-out announcements across Europe," May 10, 2020.
50	Argonne National Laboratory, "Light Duty Electric Drive Vehicles Monthly Sales Updates," September 2021,
accessed on October 20, 2021 at: https://www.anl.gov/es/light-duty-electric-drive-vehicles-monthly-sales-updates
51	U.S. EPA. The 2020 EPA Automotive Trends Report, Greenhouse Gas Emissions, Fuel Economy, and
Technology since 1975. EPA-420-R-21-003, January 2021
52	Muratori et al., "The rise of electric vehicles - 2020 status and future expectations," Progress in Energy v3n2
(2021), March 25, 2021. Accessed July 15, 2021 at https://iopscience.iop.org/article/10.1088/2516-1083/abe0ad
53	Fueleconomy.gov, 2015 Fuel Economy Guide.
54	Fueleconomy.gov, 2021 Fuel Economy Guide.
55	See RIA Chapter 2 Endnote No. 48.
56	IHS Markit Automotive Research and Analysis. US EPA Proposed Greenhouse Gas Emissions Standards for
Model Years 2023-2026; What to Expect. August 9, 2021. Accessed on the Internet on 10/29/2021 at
https://ihsmarkit.com/research-analysis/us-epa-proposed-greenhouse-gas-emissions-standards-MY2023-26.html
57	Loren McDonald, EVAdoption, LLC. EV Sales Forecasts - US EVs (BEV & PHEV) Sales and Sales Share
Forecast: 2021-2030. Accessed on the Internet on 10/29/2021 at https://evadoption.com/ev-sales/ev-sales-forecasts/
CO
Department of Energy Vehicle Technologies Office, Transportation Analysis Fact of the Week #1186, "The
National Average Cost of Fuel for an Electric Vehicle is about 60% Less than for a Gasoline Vehicle," May 17,
2021.
59	Department of Energy Vehicle Technologies Office, Transportation Analysis Fact of the Week #1190, "Battery-
Electric Vehicles Have Lower Scheduled Maintenance Costs than Other Light-Duty Vehicles," June 14, 2021.
60	Consumer Reports, "Electric Cars 101: The Answers to All YourEV Questions," November 5, 2020. Accessed
June 8, 2021 at https://www.consumerreports.org/hybrids-evs/electric-cars-101-the-answers-to-all-your-ev-
questions/
2-19
-------
61	Department of Energy Alternative Fuels Data Center, Electric Vehicle Charging Station Locations. Accessed on
May 19, 2021 at https://afdc.energy.gov/fuels/electricity _locations.html#/find/nearest?fuel=ELEC
62	Oreizi, D. "Overview of Electric Vehicle Platroms in 2021." Charged Future, February 2, 2021. Las accessed on
the Internet on 7/21/2021 at the following URL: https://www.chargedfuture.com/electric-vehicle-platforms-in-2021/
63	State of California Office of the Governor, "Governor Newsom Announces California Will Phase Out Gasoline-
Powered Cars & Drastically Reduce Demand for Fossil Fuel in California's Fight Against Climate Change," Press
Release, September 23, 2020.
64	New York State Senate, Senate Bill S2758, 2021-2022 Legislative Session. January 25, 2021.
65	Governor of New York Press Office, "In Advance of Climate Week 2021, Governor Hochul Announces New
Actions to Make New York's Transportation Sector Greener, Reduce Climate-Altering Emissions," September 8,
2021. Accessed on September 16, 2021 at https://www.governor.ny.gov/news/advance-climate-week-2021-
governor-hochul-announces-new-actions-make-new-yorks-transportation
66	Commonwealth of Massachusetts, "Request for Comment on Clean Energy and Climate Plan for 2030,"
December 30, 2020.
67	ZEV Alliance, "International ZEV Alliance Announcement," Dec. 3, 2015. Accessed on July 16, 2021 at
http://www.zevalliance.org/international-zev-alliance-announcement/.
68	International Council on Clean Transportation, "Update on the global transition to electric vehicles through
2019," July 2020.
69	Reuters, "Canada to ban sale of new fuel-powered cars and light trucks from 2035," June 29, 2021. Accessed July
1, 2021 from https://www.reuters.com/world/americas/canada-ban-sale-new-fuel-powered-cars-light-trucks-2035-
2021-06-29/
70
International Council on Clean Transportation, "Growing momentum: Global overview of government targets for
phasing out new internal combustion engine vehicles," posted 11 November 2020, accessed April 28, 2021 at
https://theicct.org/blog/staff/global-ice-phaseout-nov2020.
2-20
-------
Chapter 3: Economic and Other Key Inputs
3.1 Rebound
3.1.1 Accounting for the Fuel Economy Rebound Effect
In the context of light-duty vehicles (LDVs), rebound effects might occur when an increase in
vehicle fuel efficiency results in individuals driving more as a result of the lower cost per mile of
driving. Because this additional driving consumes fuel and generates emissions, the magnitude of
the rebound effect is one determinant of the actual fuel savings and emission reductions that will
result from adopting GHG emissions standards. The rebound effect generally refers to the
additional energy consumption that may arise from the introduction of a more efficient, lower
cost energy service. This effect offsets, to some degree, the energy savings benefits of that
efficiency improvement.1,2,3
The rebound effect for personal vehicles can, in theory, be estimated directly from the change
in vehicle use, in terms of vehicle miles traveled (VMT), which results from a change in vehicle
fuel efficiency.21 In practice, any attempt to quantify this "VMT rebound effect" (sometimes also
labeled the "direct rebound effect," or "direct VMT rebound effect") is complicated by the
difficulty in identifying an applicable data source from which the response to a significant
improvement in fuel efficiency can be estimated.b'4 Analysts, instead, often estimate the VMT
rebound indirectly, as the change in vehicle use that results from a change in fuel cost per mile
driven or a change in fuel price. When a fuel cost per mile approach is used, it does not
distinguish the relative contributions of changes in fuel efficiency and changes in fuel price to
the rebound effect, since both factors are determinants of fuel cost per mile.c When expressed as
positive percentages, the elasticities give the percentage increase in vehicle use that is presumed
to result from an increase in fuel efficiency or a decrease in fuel price.
The VMT rebound effect can also be divided into: (1) the short- to medium-run and (2) the
long-run rebound effect. Typically, studies estimating the short- to medium-run VMT rebound
effect are based upon a time period of roughly one to two years when the vehicle stock and land
use patterns are not changing significantly. The long-run rebound effect is estimated over a
longer time period when households can adjust where they work and live and the vehicle stock
can change more significantly than in the short/medium-run time frame. It is oftentimes difficult
to directly identify a long-run rebound effect, as many factors influencing travel behavior are
also changing over time. Thus, many VMT rebound estimates in the transportation policy and
economics literature are based on short- and medium-run responses because these responses are
easier to identify. Ideally, the evolution of VMT rebound effects over time from short- and
a Vehicle fuel efficiency is sometimes measured in terms of fuel consumption (gallons per mile) rather than fuel
economy (miles per gallon) in rebound estimates.
b Many of time series studies of the LDV rebound effect examine time periods before 2010. U.S. LDV fleet-wide
fuel economy has only been increasing since 2005. From 2005 to 2010, U.S. LDV fleet-wide fuel economy
improvements were fairly modest. Thus, there may be insufficient variability in LDV fuel economy to estimate a
relationship between fuel economy and VMT. See reference citation [4],
c Fuel cost per mile is equal to the price of fuel in dollars per gallon divided by fuel economy in miles per gallon (or
multiplied by fuel consumption in gallons per mile), so this figure declines when a vehicle's fuel efficiency
increases.
3-1
-------
medium-run effects to long-run effects would be utilized in an analysis of LDV GHG standards.
However, there is not a sufficient understanding of the evolution of VMT rebound effects over
time to incorporate this pattern in this final rule.
While we focus on the VMT rebound effect in our analysis of this LDV final rule, there are at
least two other types of rebound effects discussed in the transportation policy and economics
literature: the "indirect rebound effect," which typically refers to the purchase of other energy-
consuming goods or services using the cost savings from energy efficiency improvements, and
the "economy-wide rebound effect." The economy-wide rebound effect refers to the increased
demand for energy throughout the whole economy, in response to the reduced market price of
energy that results from energy efficiency improvements.
Because research on indirect and economy-wide rebound effects is scant, the rebound effect
discussed in this section refers solely to the effect of increased fuel efficiency on vehicle use.
The terms "VMT rebound effect," "direct VMT rebound effect," and "rebound effect" are used
interchangeably, and are distinguished from other rebound effects that could potentially impact
the fuel savings and emissions reductions from EPA's final LDV standards, including the indirect
and the economy-wide rebound effects.d
3.1.2 Summary of Historical Literature on the LDV Rebound Effect
This section provides a brief summary of historical literature on the LDV rebound effect. It is
important to note that a majority of the studies previously conducted rely on data from the 1950
-1990s. While these older studies provide useful information on the potential magnitude of the
rebound effect, studies based on more recent information (e.g., within the last decade) provide
more applicable estimates of how the final LDV standards will affect future driving behavior. A
number of more recent studies on LDV rebound effects (i.e., after 2010) are summarized in
Section 3.1.3 below.
Estimates from published studies covering the period from roughly 2010 and earlier using
data from 1950-2004 have found long-run rebound effects on the order of 10-30 percent. Some
of these studies are summarized in Table 3-1 and Table 3-2. In addition, Table 3-3 provides
estimates of the rebound effect using U.S. household survey data.
Table 3-1: Estimates of the Rebound Effect Using U.S. Aggregate Time-Series Data on Vehicle Travel
Author (year)
Short-Run
Long-Run
Time Period
Mayo & Mathis (1988)
22%
26%
1958-1984
Gately (1992)
9%
9%
1966-1988
Greene (1992)
Linear 5-19%
Log-linear 13%
Linear 5-19%
Log-linear 13%
1957-1989
Jones (1992)
13%
30%
1957-1989
Schimek (1996)
5-7%
21-29%
1950-1994
Source: Sorrell and Dimitropolous (2007) Table 4.6.5
d The indirect and economy-wide rebound effects do not justify applying a rebound rate higher than 10 percent in
this analysis. These additional rebound effects, to the extent they exist, may be small and their contribution to the
overall rebound rate would be offset by other considerations discussed below, such as how future GDP could reduce
the VMT rebound rate and how consumers' total VMT may be more responsive to salient changes in fuel prices than
to gradual reductions in fuel costs per mile from these LDV GHG standards.
3-2
-------
Table 3-2: Estimates of the Rebound Effect Using U.S./State and Canadian/Province Level Data
Author (year)
Short-Run
Long-Run
Time Period
Haughton & Sarkar
(1996)
9-16%
22%
1973-1992
Small and Van Dender
5%
22%
1966-2001
(2007)
2%
11%
1997-2001
Hymel, Small and Van
3%
14%
1966-2004
Dender (2010)
5%
16%
1984-2004
Barla et al. (2009)
8%
18%
1990-2004
Source: Sorrell and Dimitropolous (2007) Table 4.7, with the addition of Small and Van Dender (2007),
Hymel, Small and Van Dender (2010) and Barla et al. (2009). The Barla et al. study is based upon Canadian
Province data.



Table 3-3: Estimates of the Rebound Effect Using U.S. Household Survey Data
Author (year)
Estimate of Rebound Effect
Time Period
Goldberg (1996)
0%
1984-1990
Greene, Kahn, and
Gibson (1999)
23%
1979-1994
Pickrell & Schimek
(1999)
4-34%
1995
Puller & Greening
(1999)
49%
1980-1990
West (2004)
87%
1997
Source: Sorrell and Dimitropolous (2007).
While studies using national (Table 3-1) and state-level (Table 3-2) data have found a
relatively consistent range of long-run estimates of the rebound effect, household surveys display
more variability (Table 3-3). One explanation for this variability is that these studies consistently
find that the magnitude of the rebound effect differs according to the number of household
vehicles, and the average number of household vehicles differs among the surveys used to derive
these estimates. Still another possibility is that it is difficult to distinguish the impact of fuel cost
per mile on vehicle use from other, unobserved factors. For example, commuting distance might
influence both the choice of the vehicle and VMT. Residential density may also influence both
fuel cost per mile and VMT since households in urban areas are likely to simultaneously face
both higher fuel prices and shorter travel distances. Also, given that household data tends to be
collected on an annual basis, there may not be enough variability in the fuel price data to
estimate the magnitude of the rebound effect.6
Since there has been little variation in fuel economy over the time frame of most studies,
isolating the impact of fuel economy on VMT can be difficult using econometric analysis of
historical data. Therefore, studies that estimate the rebound effect using time series data often
examine the impact of gasoline prices or fuel cost per mile (i.e., the combined impact of both
gasoline prices and fuel economy) on VMT. However, if drivers are more responsive to changes
in fuel price or the cost of driving than to the variable directly of interest, fuel economy, these
studies may overstate the potential impact of the rebound effect resulting from this final rule. For
example, drivers may respond more to changes in fuel prices that are highly visible (i.e., salient)
3-3
-------
than to changes in fuel economy from vehicle standards that are gradually implemented over
time.
Another important distinction among studies is whether they assume that the rebound effect is
constant or varies over time in response to the absolute levels of fuel costs, personal income, or
household vehicle ownership. Most studies using aggregate annual data for the U.S. assume a
constant rebound effect, although some test whether the effect can vary as changes in retail fuel
prices or average fuel efficiency alter fuel cost per mile driven. Many studies using household
survey data estimate significantly different rebound effects for households owning varying
numbers of vehicles, with most finding that the rebound effect is larger among households that
own more vehicles.
Some of the studies, such as Small and Van Dender (2007) and Hymel, Small and Van
Dender (2010), using a combination of state-level and national data, conclude that the rebound
effect varies directly in response to changes in personal income, as well as fuel costs. These
studies indicate that the rebound effect has decreased over time as incomes have risen. One
reason that the rebound effect could vary over time is that the responsiveness to the fuel cost of
driving will be larger when it is a larger proportion of the total cost of driving. For example, as
incomes rise, the responsiveness to the fuel cost per mile of driving will decrease if households
view the time cost of driving - which is likely to be related to their income levels - as a larger
component of the total cost.
Small and Van Dender (2007) combine time series data for each of the 50 states and the
District of Columbia to estimate the rebound effect, allowing the magnitude of the rebound to
vary over time.7 For the time period 1966-2001, their study finds a long-run rebound effect of 22
percent, which is generally consistent with previously published studies.6 But for the five-year
period (1997-2001) estimated in their study, the long-run rebound effect decreases to 11 percent.
Hymel, Small and Van Dender (2010) extend the Small and Van Dender model by adding
congestion's impact on driving behavior.8 Controlling for congestion modestly increases their
estimates of the rebound effect in the study. For the time period 1966-2004, they estimate a
long-run rebound effect of 14 percent. For the time period, 1984-2004, they find a long-run
rebound effect of 16 percent, while for the most recent year in their data set, 2004, they estimate
a long-run rebound effect of 9 percent.
Barla et al. (2009) uses Canadian, province-level, panel data from 1990-2004 of light-duty
vehicles to estimate a VMT rebound effect.9 The model uses a similar methodological approach
as Small and Van Dender (2007) use, with a simultaneous three-equation model of aggregate
demand for vehicle kilometers traveled, vehicle stock and fuel efficiency. Barla et al. find short-
and long-run VMT rebound effects of 8 percent and 18 percent, respectively/
e The Small and Van Dender (2007) methodology uses a lagged dependent variable to calculate a long-run rebound
estimate. The idea is that by using the coefficient on the lagged dependent variable, the speed of adjustment can be
utilized to develop an estimate of the long-run equilibrium rebound value. In follow-on studies, Hymel, Small and
Van Dender (2010) and Hymel and Small (2015) use the same methodology to estimate a long-run VMT rebound
effect.
-F
Barla et al. (2009), using a methodology similar to Small and Van Dender (2007), break out short-run from long-
run VMT rebound effects.
3-4
-------
There is some evidence in the literature that consumers are more responsive to an increase in
fuel prices than to a decrease in fuel prices. At the aggregate level, Dargay and Gately (1997)
and Sentenac-Chemin (2012) provide some evidence that demand for transportation fuel is
asymmetric.10 In other words, given the same magnitude of change, the response to a decrease in
gasoline price is smaller than the response to an increase. Gately (1993) shows that the response
to an increase in oil prices can be on the order of five times larger than the response to a price
decrease.11 Furthermore, Dargay and Gately and Sentenac-Chemin also find evidence that
consumers respond more to a large shock than to a small, gradual change in fuel prices. Since
these final standards would decrease the cost of driving gradually over time, it is possible that the
rebound effect would be much smaller than some of the historical estimates included in the
literature.
3.1.3 Review of Recent Literature on LDV Rebound
More recent studies on LDV rebound effects have become available in the last decade (i.e.,
since 2010) and are summarized in Section 3.1.3 below. Most of the VMT rebound estimates
reported in this section are short- and medium-run estimates.
A national, U.S. study by Greene (2012) concludes that the magnitude of the rebound effect
"is by now on the order of 10 percent."12 In this study, Greene looks at how VMT is influenced
by the gasoline price fluctuations, light-duty fuel consumption patterns, U.S. real personal
income, and the number of registered vehicles in the U.S., among other factors. Over the entire
time period analyzed, 1966-2007, Greene finds that fuel prices have a statistically significant
impact on VMT, while fuel efficiency did not. From this perspective, if the impact of fuel
efficiency on VMT is not statistically significant, the VMT rebound effect could be zero. Like
Small and Van Dender, Greene finds that the VMT rebound effect is declining modestly over
time as household incomes rise and travel costs increase. When using Greene's preferred
functional form, the projected rebound effect is approximately 12 percent in 2008, and drops to
10 percent in 2020 and to 9 percent in 2030.
Using data on household characteristics and vehicle use from the 2009 NHTS, Su (2012)
analyzes the effects of locational and demographic factors on household vehicle use and
investigates how the magnitude of the rebound effect varies with vehicles' annual use.13 Using
variation in the fuel economy and per-mile cost of driving and detailed controls for the
demographic, economic, and locational characteristics of the households that owned them (e.g.,
road and population density) and each vehicle's main driver (as identified by survey
respondents), Su employs specialized regression methods to capture the variation in the rebound
effect across ten different categories of vehicle use.
Su estimates that the overall rebound effect for all vehicles in the sample averages 13 percent,
and that its magnitude varies from 11-19 percent among the ten different categories of annual
vehicle use. The smallest rebound effects were estimated for vehicles at the two extremes of the
distribution of annual use - those vehicles driven comparatively little, and those vehicles used
most intensively - while the largest estimated effects applied to vehicles that were driven slightly
more than average. Controlling for the possibility that high-mileage drivers respond to the
increased importance of fuel costs by choosing vehicles that offer higher fuel economy narrowed
the range of Su's estimates of rebound effects slightly (to 11-17 percent) but did not alter the
finding that they are smallest for lightly- and heavily-driven vehicles and largest for those with
slightly above average use. The 2009 NHTS is based upon data collected from April 2008 to
3-5
-------
April 2009. This time period may have been an unusual time period, since it was during the time
period of the Great Recession. It is not clear how the impacts of employment and output losses
from the Great Recession influenced, and resulted in, unusual travel patterns in the U.S.
Frondel and Vance (2013) use panel estimation methods and household diary travel data
collected in Germany between 1997-2009 to identify an estimate of a private transport rebound
value.14 The study focuses on single-car households that did not change their car ownership over
the timeframe each household was surveyed, up to a maximum of three years. Frondel and
Vance find a rebound effect for single-vehicle households of 46-70 percent.
Liu et al. (2014) employ the 2009 NHTS to develop an elaborate model of an individual
household's choices about how many vehicles to own, what types and ages of vehicles to
purchase, and how much combined driving to do using all of the household's vehicles for the
D.C. Metro area.15 Their analysis uses a complex mathematical formulation and statistical
methods to represent and measure the interdependence among households' choices of the
number, types, and ages of vehicles to purchase, as well as how intensively to use them. The
complexity of the relationships among a number of factors incorporated in their model - the
number of vehicles owned, their specific types and ages, fuel economy levels, and use - requires
them to measure these effects by introducing variation in income, neighborhood attributes, and
fuel costs, and observing the response of households' annual driving. Their results imply a
rebound effect of approximately 40 percent in response to a significant (25-50 percent) variation
in fuel costs, with almost exactly symmetrical responses to increases and declines in fuel costs.
Like Su and Liu et al., Linn (2016) also uses the 2009 NHTS to develop an approach to
estimate the relationship between the VMT of vehicles belonging to each household and a
variety of different factors: fuel costs, vehicle characteristics other than fuel economy (e.g.,
horsepower, the overall "quality" of the vehicle), and household characteristics (e.g., age,
income).16 Linn reports a fuel economy rebound effect with respect to VMT of between 20-40
percent.
One interesting result of the study is that when the fuel efficiency of all vehicles on the road
increases - which would be the long-run effect of rising fuel efficiency standards - two factors
have opposing effects on the VMT of a particular vehicle in a multi-vehicle household. First,
VMT increases when a vehicle's own fuel economy increases. But the increase in fuel economy
of the household's other vehicles cause the vehicle's own VMT to decrease. Since the vehicle's
own VMT response to a fuel economy increase is larger in magnitude than the VMT response to
changes in other vehicles' fuel economy, VMT increases if the fuel economy of all vehicles
increases proportionately. Linn also finds that VMT responds much more strongly to vehicle fuel
economy than to gasoline prices, which is at variance with the Hymel et al. and Greene results
discussed above.
Gillingham (2014) examines a period of significant swings in retail gasoline prices, along
with media reports of changing household driving habits, to examine how households respond to
changes in gasoline prices.17 This study uses a vehicle-level dataset of all new vehicles registered
in California in 2001-2003, and subsequently given a smog check (i.e., odometer readings) over
the 2005-2009 time frame, a period of steady economic growth but rapidly increasing gasoline
prices. Gillingham estimates the effect of differences in average monthly fuel price on monthly
vehicle use - at a county level. The primary empirical result of the responsiveness of new vehicle
VMT to gasoline prices is a medium-run estimate of 22 percent. There is evidence of
3-6
-------
considerable heterogeneity in this responsiveness across buyer types, demographics, and
geographic conditions.
In a follow-up paper, Gillingham (2020) states that this 2014 study examines the response to
the 2008 gasoline price shock in California, an unusual period when gasoline prices were
particularly salient to consumers.18 Thus, according to Gillingham, the results of his 2014 study
should not be used for developing an estimate of the VMT rebound effect for fuel
economy/GHG standards. Gillingham points to his own PhD dissertation (2011) which examines
travel patterns for California drivers from 2001 to 2009 using odometer readings as more
suggestive of the VMT rebound effect of LDV fuel economy/GHG standards.19 His PhD
estimates a VMT rebound effect of one percent.
Gillingham's results in his 2014 paper find that vehicle-level responsiveness to fuel price
increases with income, which is the opposite of the conclusions that Hymel, Small and Van
Dender and Greene find in previous studies. Gillingham hypothesizes that the increase in the per-
vehicle rebound effect with higher incomes may relate to wealthier households having more
discretionary driving, or to switching between flying and driving. Alternatively, wealthier
households tend to own more vehicles, and it is possible that within-household switching of
vehicles may account for the greater responsiveness at higher income levels.
Wang and Chen (2014) examine the responsiveness of VMT to fuel prices across income
groups, using a system of structural equations with VMT and fuel efficiency (i.e., miles per
gallon) from the 2009 NHTS.20 They find that the rebound effect is only significant for the
lowest income households (up to $25,000). Wang and Chen hypothesize that low-income
households have numerous unfilled travel needs. Thus, fuel efficient vehicles spur more driving
by low-income households.
Hymel and Small (2015) revisit the simultaneous equations methodology of Small and Van
Dender (2007) and Hymel, Small and Van Dender (2010), to see whether their previous
estimates of the VMT rebound effect have changed by adding in more recent data (2005-
2009).21 Their estimates of the long-run light-duty vehicle rebound effect over 2000-2009 are 4-
18 percent, when evaluated at average values of income, fuel cost, and urbanization in the U.S.
during this time period. However, these results also show strong evidence of asymmetry in
responsiveness to fuel price increases and decreases. Results suggest that a rebound adjustment
to fuel price rises takes place quickly; the rebound response is large in the year of, and the first
year following, a price rise, then diminishes to a smaller value. The rebound response to price
decreases occurs more slowly.
One commenter (Center for Biological Diversity, et al.) suggests that while EPA's proposed
rule reports a range of VMT rebound estimates from the Hymel and Small (2015) study of 4 to
18 percent, that only the lower value of the range, 4 percent, should be used in developing an
estimate of the VMT rebound effect for use in this rule. The basis for this commenters'
suggestion is a statement by Small in the context of the SAFE rule that: "A better
characterization of the most recent study would be that it finds a long-run rebound effect of 4.0
percent or 4.2 percent under two more realistic models that are supported by the data".g The 18
percent VMT rebound estimate in this study is based upon a model that does not consider
g See Small, Kenneth, Comment Letter on Proposed MY 2021-2026 Standards, 2018.
3-7
-------
whether drivers respond asymmetrically to increases or decreases in fuel prices/cost of driving,
which Small labels his "base model". According to Small, the base model is the starting point for
the development of asymmetric models which are the main objective of the paper. In the new
models that Hymel and Small develop in the paper, they find asymmetric responses to fuel
price/fuel cost changes. In other words, drivers respond more to fuel price/cost of driving
increases as compared to fuel price/cost of driving decreases. Since this rule will result in
increases in vehicle fuel economy which lowers the cost of driving, the applicable VMT rebound
estimate is 4 percent. To summarize, we agree with the commenter that the 4 percent VMT
rebound value is more applicable than other estimates from this study for estimating a VMT
rebound effect for this rule.
Consistent with previous results using the same modeling framework used previously in their
other published studies, Small et al. find that the VMT rebound effect declines with increasing
income and urbanization and increases with increasing fuel cost. By far the most important of
these sources of variation is income, the effect of which is large enough to reduce the projected
rebound effect for time periods of interest for this final rule.
The study by Hymel and Small also finds a strengthening of the VMT rebound effect for the
years 2003-2009 when compared to their earlier results, suggesting that some additional,
unaccounted for factors have increased the rebound effect. Three potential factors are
hypothesized to have caused the upward shift in the VMT rebound effect in the 2003-2009 time
period: (1) media coverage, (2) price volatility, and (3) asymmetric responses to fuel price
changes.11 While media coverage and volatility are important for understanding the rebound
effect based upon fuel prices, they may not be as relevant to influencing the rebound effect due
to fuel efficiency from LDV standards.
Hymel and Small find that there is an upward shift in the rebound effect of 2.5-2.8 percent
starting in 2003. Results suggest that the media coverage and volatility variables may explain
about half of the upward shift in LDV rebound in the 2003-2009 time period. Nevertheless,
these influences are small enough that they do not fully offset the downward trend in VMT
response due to higher incomes and other factors. Hence, even assuming the variables retain their
2003-2009 values into the indefinite future, they will not prevent a further diminishing of the
magnitude of the rebound effect if incomes continue to grow through time.
West et al. (2017) attempt to estimate the VMT rebound effect with household level data from
Texas, using a discontinuity in the eligibility requirements for the 2009 U.S. Car Allowance
Rebate System (CARS). This program, known as "Cash for Clunkers," incentivized eligible
households to purchase more fuel-efficient vehicles.22 Households that owned "clunkers" -
defined as vehicles with a fuel economy of 18 miles per gallon (MPG) or less - were eligible for
the subsidy, as long as their replacement vehicle was at least 22 MPG. The empirical strategy of
the paper is to compare the fuel economy of vehicle purchases and subsequent VMT of "barely
eligible" households to those households who were "barely ineligible."
h The media coverage variable is measured by constructing measures of media coverage based upon gasoline-price
related articles appearing in the New York Times newspaper. Using the ProQuest historical database, they tally the
annual number of article titles containing the words gasoline (or gas) and price (or cost). They then form a variable
equal to the annual fraction of all New York Times articles that are gasoline-price-related. This fraction ranged from
roughly 1/4000 during the 1960s to a high of 1/500 in 1974.
3-8
-------
Based upon odometer data reporting VMT, the paper finds a meaningful discontinuity in the
fuel economy of new vehicles purchased by CARS-eligible relative to ineligible households.
West et al. report that the increases in fuel economy realized by households who utilized the
program were not accompanied by increased use of the higher-MPG replacement vehicles. They
suggest this is because of the replacement vehicles' other attributes. Because households chose to
buy cheaper, smaller, and lower-performing vehicles, they did not drive any additional miles
after the purchase of the fuel-efficient vehicle. They conclude there is no evidence of a rebound
effect in response to improved fuel economy from the CARS program.
It is difficult to generalize the VMT response from the CARS program to a program for LDV
GHG standards. This was a one-time program for a fixed fleet of existing vehicles with specific
characteristics. The change in vehicle attributes from the program may not be representative of
any vehicle attribute changes from LDV GHG standards. Thus, this study does not provide
useful implications about the likely response of vehicle use to increases in LDV GHG standards.
Gillingham et al. (2015) use detailed annual vehicle-level emissions inspection test data from
Pennsylvania for 2000-2010 - including odometer readings, inspection zip codes, and extensive
vehicle characteristics - to examine both the responsiveness of driving to changing gasoline
prices, and heterogeneity in this responsiveness by geography, the fuel economy of the vehicle,
and the age of the vehicle.23 The study finds a short-run driving response (i.e., VMT) to gasoline
prices of 10 percent.
Leung (2015) examines how VMT is allocated across a typical household's vehicles in
response to a gasoline price increase.24 Leung uses 2009 NHTS data to decompose household
decreased demand for gasoline in response to a gasoline price increase into: (1) changes to VMT
and (2) changes to fuel economy or MPG (via a household reallocating its VMT to a vehicle with
a different MPG). Leung finds a VMT responsiveness to gasoline prices of 10 percent.
Langer et al. (2017) develop a model of motorists' demand for automobile travel that
explicitly accounts for heterogeneity across drivers and their vehicles for the state of Ohio. The
study estimates drivers' responses to changes in the marginal cost of driving. The study is based
upon data from State Farm Mutual Automobile Insurance Company on individual drivers who, in
return for a discount on their insurance, allowed a private firm to remotely record their vehicles'
VMT from odometer readings from 2009-2013. The model allows for a comparison of the
effects of gasoline and VMT taxes on fuel consumption, among other factors. They find a
responsiveness of VMT with respect to the price of automobile travel is 12 percent.25
Knittel and Sandler (2018) estimate the VMT responsiveness to gasoline price, in the context
of the gasoline tax as an emission reduction policy tool. The study looks at California LDVs over
the period of 1998-2008, using odometer readings (i.e., Smog Check data).26 They find an
average VMT responsiveness of 13 percent. They also observe significant heterogeneity across
different types of vehicles, suggesting that VMT responsiveness to gasoline prices can vary
significantly based upon the specific sub-classes of vehicles considered.
One interesting study of VMT rebound is by De Borger et al. (2016). They analyze the
response of vehicle use to changes in fuel economy among a sample of nearly 350,000 Danish
households owning a single vehicle, of which almost one-third replaced it with a different model
during the 2001-2011 time period.27 By comparing the change in households' driving between
those who replaced their vehicles during the intervening period to those who did not, the authors
3-9
-------
attempt to isolate the effect of changes in fuel economy on vehicle use from those of other
factors. Their data allow them to control for the effects of household characteristics and vehicle
features other than fuel economy on vehicle use. The authors use complex statistical methods to
account for the fact that some households replacing their vehicles may have done so in
anticipation of changes in their driving demands (rather than the reverse), as well as for the
possibility that some households who replaced their vehicles may be doing so because their
driving behavior is more sensitive to fuel prices than other households.
De Borger et al. measure the rebound effect from the change in households' vehicle use in
response to changes in fuel economy that are a consequence of their decisions to replace their
vehicles. Thus, the authors directly estimate the fuel economy rebound effect itself, in contrast to
studies that rely on indirect measures, such as fuel prices or the costs per mile of driving. Their
preferred estimates of the fuel economy rebound effect range from 8-10 percent. De Borger et al.
also find no evidence that the rebound effect is smaller among lower-income households than
among their higher-income counterparts.
Gillingham et al. (2016) undertake a summary and review of the general rebound literature,
including rebound effects from LDV studies considered for this final rule, as well as electricity
used in stationary applications.28 According to Gillingham et al., the literature suggests that
differences in estimates of the rebound effect stem from its varying definitions, as well as
variation in the quality of data and empirical methodologies used to estimate it. Gillingham et al.
seek to clarify the definition of each of the channels of the rebound effect, and to critically assess
the state of the literature that estimates its magnitude.
Gillingham et al. note that most analyses assume a "zero cost breakthrough" (ZCB) - their
term for an improvement in efficiency that results in energy savings and related energy or fuel
cost savings but does not have associated increased costs of technology or implementation. Thus,
the authors argue, most analyses do not reflect the true costs of a "policy-induced improvement."
Gillingham et al. also caution that failing to account for the increased costs of equipment and/or
implementation of a policy-induced improvement may result in different estimates of the
rebound effect, compared to a ZCB improvement in efficiency.
Wenzel and Fujita (2018) examine the responsiveness of driving to changes in the price of
gasoline and driving costs.29 Using detailed odometer readings from over 30 million vehicles in
four urban areas of Texas from 2005-2010, they estimate that the responsiveness of the demand
for VMT with respect to the price of gasoline in Texas is 9 percent, after accounting for
differences in vehicle models. They also use the rated combined city /highway fuel economy of
each vehicle to calculate the cost of driving, in cents per mile, since a vehicle's previous
inspection. They find a VMT responsiveness with respect to the cost of driving of 16 percent.
A study by Gillingham and Munk-Nielsen (2019) provides an estimate for the fuel price
elasticity of driving for Denmark in the period from 1998-2011.30 They find a one-year elasticity
of 30 percent. An interesting aspect of this study is that it finds two tails of more responsive
drivers. The first tail is drivers living in the outskirts of cities with long commutes, but with
adequate access to public transport. The second tail is composed of drivers who commute very
little and tend to live in cities. Households with long commutes can readily switch to public
transport, while households who commute very little largely use their vehicles for a diverse set of
non-work trips, many of which can be readily switched to other modes of transport.
3-10
-------
The finding of the two tails may explain differences in the results in fuel price elasticities
between the U.S. and Europe, according to Gillingham and Munk-Nielsen. The Gillingham and
Munk-Nielsen study finds a price responsiveness of driving of 30 percent for Denmark drivers
but, if ample access to public transport is eliminated, this responsiveness changes to 13 percent.
This is more in line with recent estimates from the U.S. for the fuel price responsiveness of
driving.
In an additional study, Gillingham (2020) develops a rationale for the use of a 10 percent
VMT rebound effect, and argues that the 20 percent used by the agencies in the most recent joint
LDV rulemaking for the 2020-2026 GHG/fuel economy standards is too high.31'32 Gillingham
points out that the agencies argue that odometer reading data is the most reliable data when they
are discussing the relationship between vehicle miles traveled and vehicle age, but do not make
this distinction in the discussion of the VMT rebound effect. Gillingham argues that, when
reviewing VMT rebound studies and attempting to develop a single value of a VMT rebound
effect, studies based upon odometer readings should be given greater weight. This is because
odometer reading data is more reliable, since it is measured rather than self-reported, and may be
more representative of travel behavior by covering nearly the entire LDV fleet in a region.
Based upon a list of recent VMT studies that the agencies reviewed in the proposed 2022-
2026 LDV standards, Gillingham presents a summary of literature relevant for his central
estimate of the rebound effect of fuel economy standards in the U.S. He restricts his review to
publicly available U.S.-based literature from the past decade. His review excludes estimates from
outside of the U.S., in particular Europe, as travel behavior has been shown to be different due to
a variety of factors including different urban forms and public transportation access. Second,
Gillingham excludes some estimates from unpublished work that are inaccessible, or that
estimate something other than the VMT rebound effect (i.e., response of gasoline demand to fuel
price). Third, Gillingham excludes estimates that are inappropriate for using as an estimate of the
rebound effect, based upon individual author's judgements. For example, as mentioned above,
Gillingham excludes his own study published in 2014, which examines the driving response to
the 2008 gasoline price shock, an unusual period when gasoline prices were particularly salient
to consumers.
According to Gillingham, a few clear findings are apparent. First, there is a relatively wide
range of estimates. In general, studies using household survey data tend to have much higher
rebound effect estimates than those using odometer reading data. Second, the average rebound
effect over all studies that are considered by Gillingham is 14 percent, and the average over all
studies using just odometer readings is 8 percent. According to Gillingham, based upon his
review of relevant studies, he casts doubt on the argument for a central case estimate of 20
percent for the VMT rebound effect of U.S. LDV GHG/fuel economy standards.33
A study by the Dimitropoulos et al. (2018) presents a meta-analysis of 76 empirical studies
and 1,138 estimates of elasticities of travel from 18 countries (i.e., the U.S., European countries,
China and India) over the last fifty years, which can serve as possible measures of the VMT
rebound effect.34 Some of the most recent U.S. state-level studies using odometer readings data
such as Knittel and Sandler (2018), Langer et al. (2017) and Wenzel and Fujita (2018) are not
included in the meta-analysis. The meta-analysis uses an econometric approach to assess the
sources of heterogeneity in rebound effect estimates across the studies. The overall world VMT
rebound effect is estimated to be, on average, around 12 percent in the short-run, and roughly 32
3-11
-------
percent in the long-run, across all of the studies considered. Other findings by Dimitropoulos et
al. suggest that studies using household survey data typically produce long-run rebound
estimates twice or more as large as studies based on aggregate data. The meta-analysis also finds
that the VMT rebound effect is declining worldwide, at a rate of roughly 0.7 percentage points
per year.
Dimitropoulos et al. provide VMT rebound estimates that vary by the price of gasoline,
population density, and gross domestic product (GDP) per capita, based upon the meta-analysis
results. They conclude that the VMT rebound effect increases with the price of gasoline and
population density, and decreases with per capita GDP, making rebound estimates from different
countries not directly comparable. Using 2018 U.S. values for gasoline price ($0.63/liter),
population density (33.75/km2), and GDP per capita ($51,552) for the U.S., Dimitropoulos et al.
results predict a long-run VMT rebound effect of roughly 20 percent for the U.S.35
In a Report entitled, "Science Advisory Board (SAB) Consideration of the Scientific and
Technical Basis of the EPA's Proposed Rule titled The Safer Affordable Fuel-Efficient (SAFE)
Vehicles Rule for Model Years 2021-2026 Passenger Cars and Light Trucks",36 EPA's
Scientific Advisory Board (SAB) provides comments on the VMT rebound estimate used in the
proposed SAFE rule. On the magnitude of the 20 percent rebound value used, the SAB provides
several recommendations. First, the SAB suggests that the Agency's consider several recent
odometer-based VMT rebound studies (e.g., Langer et al. (2017); Knittel and Sandler (2018);
and Wenzel and Fujita (2018)) 26'27>30 which were not considered for the proposed SAFE rule.
The SAB also recommends that the Agency not over-generalize on the importance of the
rebound effect, assuming the implications of increased efficiency will be seen uniformly across
sectors of the economy. Finally, the SAB recommends that the Agency consider the relative
saturation of demand for VMT, the increasing role that the travel behaviors of Millennials, Baby
Boomers and ride sharing services have in reducing the magnitude of the U.S. VMT rebound
effect. In a concluding statement, the SAB comments, "Due to these concerns, the SAB
recommends that the rebound estimate be reconsidered to account for the broader literature, and
that it be determined through a full assessment of the quality and relevance of the individual
studies rather than a simple average of results. A more in-depth analysis will allow the Agency to
weight papers based on their quality and applicability: recent papers using strong methodology
and U.S. data should be weighted more heavily than older papers, or those from outside the U.S.,
or those with weaker methodology."
3-12
-------
3.1.4 Basis for Rebound Effect Used in this Final LDV Rule
EPA uses a single point estimate for the direct VMT rebound effect as an input to the agency's
analyses for this final LDV GHG rule (2023-2026). Based on a review of estimates of the VMT
rebound effect from recent analyses (i.e., since 2010) and some of the insights from VMT
rebound studies completed before 2010, EPA is using a value of 10 percent for the direct
rebound effect for this final rule. In Chapter 4.10, as sensitivities, EPA presents estimates of the
impacts of using a five percent and 15 percent VMT rebound effect.
There is a wide variety of estimates of the VMT rebound effect from the recent analyses, in
part, due to the many different methodologies and data sources used to try to quantify this
impact. Given the broad range of values, EPA believes it is important to critically evaluate which
studies are most likely to be reflective of the rebound effect that is relevant to the final standards
(2023-2026). In other words, one cannot just take the "average" rebound estimates from
literature to use for the VMT rebound effect for this final rule.
EPA applies the following critical factors when choosing a VMT rebound estimate for this
final rulemaking:
1.	Geographic/Timespan relevance: Priority is given to U.S., as opposed to international
rebound studies, since U.S. studies are based upon U.S. LDV travel, land use patterns,
and socio-economic conditions. U.S. national-level studies are most useful since they are
based upon the geographic scale of this final rulemaking. Priority is given to studies that
are based on U.S. demographic/land use patterns over timespans most relevant to this
rulemaking's analytical timeframe (e.g., 2023-2050). Thus, we focus on studies relying
upon time series data rather than single-year studies. Even well-executed single year
studies have difficulty in controlling for confounding factors influencing the VMT
rebound effect, so these studies are not given significant weight;
2.	Time period of study: Priority is given to more recent rebound studies in the last decade,
since their driving patterns are more likely to resemble driving patterns over the time
frame of this final LDV rule;
3.	Reliability/Replicability of study: Priority is given to studies that use measured
odometer reading data for VMT. Many household survey studies rely on self-reported
VMT data, which may not produce as reliable estimates of the VMT rebound effect as
studies based on measured data. Also, odometer reading data is likely to more
representative of travel behavior by covering nearly the entire LDV fleet in a region.
Finally, the 2009 NHTS data was collected during the Great Recession time period. It is
not clear how representative travel patterns in the U.S. were during this time period for
developing estimates over timespans most relevant to this rulemaking's analytical
timeframe (e.g., 2023-2050); and
4.	Strong statistical/methodological basis: Priority is given to studies using strong
statistical methods that effectively attempt to control and isolate the impacts of the VMT
rebound effect.
3-13
-------
The critical factors listed above are consistent with the SAB's recommendations on how to
determine a preferred estimate of the magnitude of the VMT rebound effect for use in this final
LDV GHG rule. EPA undertakes a comprehensive, overall, in-depth assessment of the full range
of VMT rebound studies relevant for developing a preferred VMT rebound estimate for this final
rule. EPA weighs the applicability and quality of each individual VMT rebound study in this
overall assessment. EPA does not simply average the results of the relevant VMT studies in
developing a VMT rebound estimate for this final rule. EPA gives more weight to U.S. rebound
studies as opposed to international VMT rebound studies. In addition, EPA gives more weight to
recent rebound studies (i.e., in the last decade). The application of the critical factors listed above
to the relevant VMT rebound literature is presented below.
Studies that provide a U.S. estimate of the LDV VMT rebound effect are most applicable to
estimating the overall VMT effects of the final LDV standards. The most recent national, U.S.
studies are by Hymel and Small (2015), which estimates a rebound effect ranging from 4-18
percent, and Greene (2012), which concludes that the rebound effect "is by now on the order of
10 percent." Based upon the comments received on the proposal for this rule referencing Small's
comments on the SAFE rule, EPA is using the 4 percent value from the Hymel and Small study
(2015), and not the range. See Table 3-4 below. Since GHG standards result in improved vehicle
efficiency, which lowers the cost of driving, and Hymel and Small found an asymmetric
response to the costs of driving, the lower end of the range in the Hymel and Small estimates is
more applicable for evaluating the final LDV GHG standards.
Both studies, Greene (2012) and Hymel and Small (2015), are based upon U.S. vehicle travel
patterns, as opposed to relying on international (i.e., outside the U.S.) travel patterns. Both
studies have been published in the last decade and are based upon the geographic scale of this
final rulemaking - the national, U.S. level. Both studies estimate the VMT rebound effect
looking at travel behavior over many years, as opposed to studies that rely on only a single year.
As noted above, even well executed, single year studies may have difficulty in controlling for
confounding factors influencing the VMT rebound effect. Both studies use solid statistical
methods that are generally effective at isolating the impacts of the VMT rebound effect. See
Table 3-4 below for the list of national, U.S. studies given significant weight in developing an
estimate of the VMT rebound effect for this final rule.
The set of studies at the U.S. state-level using odometer readings further support the 10
percent VMT rebound estimate for the U.S. as a whole. These studies, for Pennsylvania:
Gillingham et al. (2015); for California: Gillingham (2011)/Knittel and Sandler (2018); for Ohio:
Langer et al. (2017); and for Texas: Wenzel and Fujita (2018), find VMT rebound effects of 10,
1, 13, 12, and 9-16 percent, respectively. See Table 3-4 below for the list of U.S. state-level,
odometer studies given significant weight in developing an estimate of the VMT rebound effect
for this final rule.
3-14
-------
Table 3-4: Studies Given Significant Weight in Developing an Estimate of the VMT Rebound Effect for this
Final Rule
Author
Year
Estimate of Rebound Effect
Description/
Time Period
U.S. National
Greene
2012
10%
Aggregate
1966-2007
Hymel and
Small
2015
4%
State-level
2000-2009
State-Level Odometer
Gillingham
2011
1%
California
2001-2009
Gillingham et al.
2015
10%
Pennsylvania
2000-2010
Langer et al.
2017
12%
Ohio
2009-2013
Wenzel and
Fujita
2018
9-16%
Texas
2005-2010
Knittel and
Sandler
2018
13%
California
1998-2010
All of the state-level studies are based upon U.S. vehicle travel patterns, as opposed to relying
on international (i.e., outside the U.S.) travel patterns. All five of the studies have been published
in the past decade. These state-level studies use odometer readings to measure VMT, as opposed
to self-reported data, which provides more confidence in the reliability of their results. In
addition, odometer reading data is likely to be more representative of travel behavior by covering
nearly the entire LDV fleet in a region. Also, these studies all use time series, rather than single
year, data to estimate the VMT rebound effect, avoiding possible confounding effects of using a
single year's data. All of the U.S. state-level studies use solid statistical methods that are
generally effective at isolating the impacts of the VMT rebound effect. The Gillingham (2014)
study, which found a 22 percent VMT rebound effect in California, is excluded from
consideration in the set of state-level rebound studies using odometer data. As Gillingham points
out, this study assesses the response to driving from a salient 2008 gasoline price shock, which is
quite different than gradual changes in fuel economy from the final LDV standards.
The four states considered in the studies - Pennsylvania, Ohio, Texas and California - are
geographically diverse, with different population sizes, incomes, demographic characteristics,
and vehicle fleet characteristics. Nevertheless, these studies provide estimates of VMT rebound
effects that are roughly clustered in the 10 percent range. Thus, these U.S. state-level studies,
based on odometer readings, provide support for the use of a 10 percent rebound effect in
developing a single VMT rebound estimate for the U.S. nation as a whole.
The West et al. study (2017) on the CARS (Cash for Clunkers) program did not find a VMT
rebound effect (i.e., a VMT rebound effect of zero). This study uses odometer data from the state
of Texas. But the VMT response to a vehicle scrappage program could be very different than for
a program that results in a gradual increase in fuel economy over time, such as the LDV final
rule considered here. For example, West et al. find that vehicle attribute changes (i.e., lower curb
weight/horsepower) offset the lower costs of driving, resulting in a zero-rebound effect. It is not
3-15
-------
clear how vehicle attributes will change with this final LDV rule. Therefore, little to no weight is
given to the West et al. study in determining a VMT rebound effect for this final rule.
Su (2012), Liu et al. (2014), Leung (2015) and Linn (2016), each using NHTS 2009 data, find
rebound effects that vary from 10-40 percent. Wang and Chen (2014), using the 2009 NHTS
data as well, find a rebound effect only for low-income households. These widely different
results based upon the same dataset suggest that these studies may not provide reliable estimates
of the VMT rebound effect. The concern is that different methodological approaches with the
same set of data yield different results. All of the household survey studies are based on self-
reported VMT data, suggesting that the results may not be as reliable as studies based on
odometer readings. Further, the NHTS data are for a single year. Even well executed studies
based upon a single year of data may have difficulty controlling for confounding factors
influencing estimates of the VMT rebound effect. Also, travel and household data from the 2009
NHTS was collected while the U.S. was in the midst of the Great Recession. The Great
Recession led to significant employment and output losses in the U.S., which may have possibly
led to unusual travel patterns.
This final rule uses AEO 2021 as the basis for projecting economic and fuel market trends
during time frame of analysis of this final rule.37 The AEO 2021 projects that U.S. Gross
Domestic Product will increase over time. Some of the national, aggregate studies of the U.S.,
Hymel and Small (2015) and Greene (2012), find that the VMT rebound effect decreases as
household incomes rise. As incomes rise, the value of time spent driving is typically assumed to
rise as well. Thus, the time cost of travel becomes a larger fraction of total travel costs, so
vehicle use may become less responsive to variations in fuel costs. Wang and Chen find that only
low-income households have a rebound effect, which is consistent with the VMT rebound effect
diminishing with increases in income. On the other hand, Gillingham, (2014) finds that the VMT
rebound effect increases with household income. But the Gillingham (2014) study examines the
travel response to a salient gasoline price increase, which is somewhat different than a gradual
improvement in fuel economy from this final LDV rule. Also, De Borger et al. (2016) did not
find a significant impact of income on the VMT rebound effect. Thus, the evidence of how the
rebound effect varies with income is somewhat mixed. While the relationship between the VMT
rebound effect and income is supported by some of the national, aggregate studies; some, but
less, weight is given to this factor in determining a VMT rebound value for this final rule.
In summary, the 10 percent VMT rebound value chosen for use in these final LDV GHG
standards (2023-2026) is based upon applying a set of critical factors - geographic/timespan
relevance, time period, repeatability/reliability, and statistical/methodological basis - and the
weight of evidence from multiple recent studies (i.e., studies since 2010), based upon an updated
and rigorous review of the large body of literature on this topic. A combination of the recent
U.S., national VMT rebound studies and recent, odometer-based, VMT rebound studies for
different states - Pennsylvania, Ohio, Texas and California - that are geographically diverse,
with different population sizes, incomes, demographic characteristics, and vehicle fleet
characteristics, support a single point value of 10 percent for the direct VMT rebound effect. All
of the studies estimate the VMT rebound effect over many years, as opposed to a single year, and
use strong statistical methods. A simple average of all seven studies listed in Table 3-4 of this
final rule results in a VMT rebound value of 8.9 percent. A simple average of all of the state-
level, odometer-based VMT rebound studies listed in Table 3-4 results in a VMT rebound
estimate of 9.7 percent. The four most recent state-level, odometer-based VMT rebound studies
3-16
-------
since 2015 (i.e., Gillingham et al., Langer et al., Wenzel and Fujita and Knittel and Sandler) have
a simple average VMT rebound estimate of 11.9 percent. A variety of factors discussed in this
rulemaking could potentially lower the VMT rebound effect estimate, including: using salient
fuel prices to represent the gradual cost of driving from GHG standards; asymmetric driver
responses to fuel prices/costs of driving; the impacts of increases in income on the VMT rebound
effect; and, the VMT rebound effect with high fuel efficient vehicles. More research is needed on
these factors, particularly using odometer-based data, before the impacts of these factors can
more significantly influence the overall VMT rebound estimate being used in this rule. To
conclude, we believe that EPA's evaluation of the recent, U.S. VMT rebound literature provides
a very reliable estimate of the VMT rebound effect, 10 percent, and we have used this value
within this LDV GHG final rule.
3.2 Energy Security Impacts
This final rule is designed to require improvements in the fuel economy of light-duty vehicles
(LDV) and thereby reduce fuel consumption and GHG emissions. In turn, this final rule helps to
reduce U.S. petroleum imports. A reduction of U.S. petroleum imports reduces both financial
and strategic risks caused by potential sudden disruptions in the supply of imported petroleum to
the U.S., thus increasing U.S. energy security. In other words, reduced U.S. oil imports act as a
"shock absorber" when there is a supply disruption in world oil markets. This section
summarizes the agency's estimates of U.S. oil import reductions and energy security benefits of
the final light-duty GHG standards for model years 2023-2026.
Energy independence and energy security are distinct but related concepts, and an analysis of
energy independence informs our analysis of energy security.38 The goal of U.S. energy
independence is the elimination of all U.S. imports of petroleum and other foreign sources of
energy.39 U.S. energy security is broadly defined as the continued availability of energy sources
at an acceptable price.40 Most discussions of U.S. energy security revolve around the topic of the
economic costs of U.S. dependence on oil imports.1'41 We note that energy independence
associated with oil is theoretically achievable (i.e. the U.S. produces all of its own oil). On the
other hand, energy security risks can never be completely eliminated because the price of energy
consumed (i.e., oil) in the U.S. is affected by global commodity markets.
The U.S.'s oil consumption has been gradually increasing in recent years (2015-2019) before
dropping dramatically as a result of the COVID pandemic in 2020.42 The U.S. has increased its
production of oil, particularly "tight" (i.e., shale) oil, over the last decade.43 As a result of the
recent increase in U.S. oil production, the U.S. became a net exporter of crude oil and product in
2020 and is now projected to be a net exporter of crude oil and product through 2023 to 2050, the
time frame of this analysis.44 This is a significant reversal of the U.S.'s net export position since
the U.S. has been a substantial net importer of crude oil and product starting in the early 1950s.45
Given that the U.S. is projected to be a net exporter of crude oil and product for the
foreseeable future, one could reason that the U.S. does not have a significant energy security
problem anymore. However, U.S. refineries still rely on significant imports of heavy crude oil
1 The issue of cyberattacks is another energy security issue that could grow in significance over time. For example,
one of the U.S.'s largest pipeline operators, Colonial Pipeline, was forced to shut down after being hit by a
ransomware attack. The pipeline carries refined gasoline and jet fuel from Texas to New York.
3-17
-------
from potentially unstable regions of the world. Also, oil exporters with a large share of global
production have the ability to raise or lower the price of oil by exerting the market power
associated with the Organization of Petroleum Exporting Countries (OPEC) to alter oil supply
relative to demand. These factors contribute to the vulnerability of the U.S. economy to episodic
oil supply shocks and price spikes, even when the U.S. is projected to be an overall net exporter
of crude oil and product.
3.2.1 Review of Historical Energy Security Literature
Energy security discussions are typically based around the concept of the oil import premium.
The oil import premium is the extra cost of importing oil beyond the price of the oil itself as a
result of: (1) potential macro-economic disruption and increased oil import costs to the economy
from oil price spikes or "shocks" and (2) monopsony impacts. Monopsony impacts stem from
changes in the demand for imported oil, which changes the price of all imported oil.
The so called oil import premium gained attention as a guiding concept for energy policy in
the aftermath around of the oil shocks of the 1970's (Bohi and Montgomery 1982, EMF 1982).46
Plummer (1982) provided valuable discussion of many of the key issues related to the oil import
premium as well as the analogous oil stockpiling premium.47 Bohi and Montgomery (1982)
detailed the theoretical foundations of the oil import premium and established many of the
critical analytic relationships.48 Hogan (1981) and Broadman and Hogan (1986, 1988) revised
and extended the established analytical framework to estimate optimal oil import premia with a
more detailed accounting of macroeconomic effects.49'50 Since the original work on energy
security was undertaken in the 1980's, there have been several reviews on this topic by Leiby,
Jones, Curlee and Lee (1997) and Parry and Darmstadter (2004).51'52
The economics literature on whether oil shocks are the same level of threat to economic
stability as they once were, is mixed. Some of the literature asserts that the macroeconomic
component of the energy security externality is small. For example, the National Research
Council (2009) argued that the non-environmental externalities associated with dependence on
foreign oil are small, and potentially trivial.53 Analyses by Nordhaus (2007) and Blanchard and
Gali (2010) questioned the impact of oil price shocks on the economy in the early 2000 time
frame.54 They were motivated by attempts to explain why the economy actually expanded during
the oil shock in the early 2000 time frame, and why there was no evidence of higher energy
prices being passed on through higher wage inflation. One reason, according to Nordhaus and
Blanchard and Gali, is that monetary policy has become more accommodating to the price
impacts of oil shocks. Another reason is that consumers have simply decided that such
movements are temporary and have noted that price impacts are not passed on as inflation in
other parts of the economy.
Hamilton (2012) reviewed the empirical literature on oil shocks and suggested that the results
are mixed, noting that some work (e.g. Rasmussen and Roitman (2011)) finds less evidence for
economic effects of oil shocks or declining effects of shocks (Blanchard and Gali (2010)), while
other work continues to find evidence regarding the economic importance of oil shocks.55 For
example, Baumeister and Peersman (2011) find that an "oil price increase of a given size seems
to have a decreasing effect over time, but noted that the declining price-elasticity of demand
means that a given physical disruption had a bigger effect on price and turned out to have a
similar effect on output as in the earlier data".56 Hamilton observed that "a negative effect of oil
prices on real output has also been reported for a number of other countries, particularly when
3-18
-------
nonlinear functional forms have been employed" (citing as examples Kim (2012), Engemann,
Kliesen, and Owyang (2011)).57'58 Alternatively, rather than a declining effect, Ramey and Vine
(2010) find "remarkable stability in the response of aggregate real variables to oil shocks once
we account for the extra costs imposed on the economy in the 1970s by price controls and a
complex system of entitlements that led to some rationing and shortages."59
Some of the literature on oil price shocks emphasizes that economic impacts depend on the
nature of the oil shock, with differences between price increases caused by a sudden supply loss
and those caused by rapidly growing demand. Recent analyses of oil price shocks have
confirmed that "demand-driven" oil price shocks have greater effects on oil prices and tend to
have positive effects on the economy while "supply-driven" oil shocks still have negative
economic impacts (Baumeister, Peersman and Robays (2010)).60 A paper by Kilian and
Vigfusson (2014), for example, assigned a more prominent role to the effects of price increases
that are unusual, in the sense of being beyond the range of recent experience.61 Kilian and
Vigfussen also concluded that the difference in response to oil shocks may well stem from the
different effects of demand- and supply-based price increases: "One explanation is that oil price
shocks are associated with a range of oil demand and oil supply shocks, some of which stimulate
the U.S. economy in the short-run and some of which slow down U.S. growth (see Kilian
2009a)".62
The general conclusion that oil supply-driven shocks reduce economic output is also reached
in a paper by Cashin et al. (2014) which focused on 38 countries from 1979-2011.63 They stated:
"The results indicate that the economic consequences of a supply-driven oil-price shock are very
different from those of an oil-demand shock driven by global economic activity, and vary for oil-
importing countries compared to energy exporters". Cashin et al. continues "oil importers
(including the U.S.) typically face a long-lived fall in economic activity in response to a supply-
driven surge in oil prices". But almost all countries see an increase in real output for an oil-
demand disturbance.
EPA's assessment of the energy security literature finds that there are benefits to the U.S.
from reductions in oil imports. But there is some debate as to the magnitude, and even the
existence, of energy security benefits from U.S. oil import reductions. Over the last decade,
differences in economic impacts from oil demand and oil supply shocks have been distinguished.
The oil security premium calculations in this analysis are based on price shocks from potential
future supply events only. Oil supply shocks, which reduce economic activity, have been the
predominant focus of oil security issues since the oil price shocks/oil embargoes of the 1970's.
3.2.2 Review of Recent Energy Security Literature
There have also been a handful of more recent studies undertaken in the last few years that are
relevant for the issue of energy security: one by Resources for the Future (RFF), a study by
Brown, two studies by Oak Ridge National Laboratory (ORNL), and a couple of studies, Newell
and Prest and Bjornland et al., on the responsiveness of U.S. tight oil (i.e., shale oil) to world oil
price changes. We provide a brief review and high-level summary of each of these studies below.
The RFF study (2017) attempts to develop updated estimates of the relationship among gross
domestic product (GDP), oil supply and oil price shocks, and world oil demand and supply
elasticities.64 In a follow-on study, Brown summarized the RFF study results well.65 The RFF
work argues that there have been major changes that have occurred in recent years which have
3-19
-------
reduced the impacts of oil shocks on the U.S. economy. First, the U.S. is less dependent on
imported oil than in the early 2000s due in part to the 'Tracking revolution" (i.e., tight/shale oil),
and to a lesser extent, increased production of renewable fuels. In addition, RFF argues that the
U.S. economy is more resilient to oil shocks than in the earlier 2000 time frame. Some of the
factors that make the U.S. more resilient to oil shocks include increased global financial
integration and greater flexibility of the U.S. economy (especially labor and financial markets),
many of the same factors that Nordhaus and Blanchard and Gali pointed to as discussed above.
In the RFF effort, a number of comparative modeling scenarios are conducted by several
economic modeling teams using three different types of energy-economic models to examine the
impacts of oil shocks on U.S. GDP. The first is a dynamic stochastic general equilibrium model
developed by Balke and Brown.66 The second set of modeling frameworks use alternative
structural vector autoregressive models of the global crude oil market.67'68'69 The last of the
models utilized is the National Energy Modeling System (NEMS).70
Two key parameters are focused upon to estimate the impacts of oil shock simulations on U.S.
GDP: oil price responsiveness (i.e., the short-run price elasticity of demand for oil) and GDP
sensitivity (i.e., the elasticity of GDP to an oil price shock). The more inelastic (i.e., the less
responsive) short-run oil demand is to changes in the price of oil, the higher will be the price
impacts of a future oil shock. Higher price impacts from an oil shock result in higher GDP
losses. The more inelastic (i.e., less sensitive) GDP is to an oil price change, the less the loss of
U.S. GDP with future oil price shocks.
For oil price responsiveness, RFF reports three different values: a short-run price elasticity of
oil demand from their assessment of the "new literature", -0.17; a "blended" elasticity estimate; -
0.05, and short-run oil price elasticities from the "new models" RFF uses, ranging from -0.20 to -
0.35. The "blended" elasticity is characterized by RFF in the following way: "Recognizing that
these two sets of literature [old and new] represent an evolution in thinking and modeling, but
that the older literature has not been wholly overtaken by the new, Benchmark-E [the blended
elasticity] allows for a range of estimates to better capture the uncertainty involved in calculating
the oil security premiums."
The second parameter that RFF examines is the GDP sensitivity. For this parameter, RFF's
assessment of the "new literature" finds a value of -0.018, a "blended elasticity" estimate of -
0.028, and a range of GDP elasticities from the "new models" that RFF uses that range from -
0.007 to -0.027. One of the limitations of the RFF study is that the large variations in oil price
over the last fifteen years are believed to be predominantly "demand shocks": for example, a
rapid growth in global oil demand followed by the Great Recession and then the post-recession
recovery.
The only supply-side oil shock in the last several years was the attack on the Saudi Aramco
Abqaiq oil processing facility and the Khurais oil field (which took place after the publication of
RFF's study). On September 14th, 2019, a drone and cruise missile attack damaged the Saudi
Aramco Abqaiq oil processing facility and the Khurais oil field in eastern Saudi Arabia. The
Abqaiq oil processing facility is the largest crude oil processing and stabilization plant in the
world, with a capacity of roughly 7 MMBD or roughly seven percent of global crude oil
production capacity.71 On September 16th, the first full day of commodity trading after the
attack, both Brent and West Texas Intermediate (WTI) crude oil prices surged by $7.17/barrel
3-20
-------
and $8.34/barrel, respectively, in response to the attack, the largest price increase in roughly a
decade.
However, by September 17th, Saudi Aramco reported that the Abqaiq plant was producing 2
MMBD, and they expected its entire output capacity to be fully restored by the end of
September.72 Tanker loading estimates from third-party data sources indicated that loadings at
two Saudi Arabian export facilities were restored to the pre-attack levels.73 As a result, both
Brent and WTI crude oil prices fell on September 17th, but not back to their original levels. The
oil price spike from the attack on the Abqaiq plant and Khurais oil field was prominent and
unusual, as Kilian and Vigfusson (2014) describe. While pointing to possible risks to world oil
supply, the oil price shock was short-lived, and generally viewed by market participants as being
transitory, so it did not influence oil markets over a sustained time period. Thus, there is little
recent empirical evidence to estimate the response of the U.S. economy to an oil supply shock of
a significant magnitude^
A second set of recent studies related to energy security are from ORNL. In the first study,
ORNL (2018) undertakes a quantitative meta-analysis of world oil demand elasticities based
upon the recent economics literature.74 The ORNL study estimates oil demand elasticities for
two sectors (transportation and non-transportation) and by world regions (OECD and Non-
OECD) by meta-regression. To establish the dataset for the meta-analysis, ORNL undertakes a
literature search of peer reviewed journal articles and working papers between 2000 and 2015
that contain estimates of oil demand elasticities. The dataset consisted of 1,983 observations
from 75 published studies. The study finds a weighted short-run price elasticity of world oil
demand of -0.07 and a long-run price elasticity of world oil demand of -0.26.
The second relevant ORNL (2018) study from the standpoint of energy security is a meta-
analysis that examines the impacts of oil price shocks on the U.S. economy as well as many
other net oil-importing economies.75 Nineteen studies after the year 2000 were identified that
contain quantitative/accessible estimates of the economic impacts of oil price shocks. Almost all
studies included in the review were published since 2008. The key result that the study finds is a
short-run oil price elasticity of U.S. GDP, roughly one year after an oil shock, of -0.021, with a
68 percent confidence interval of -0.006 to -0.036.
Only in recent years have the implications of the "tight oil revolution" been felt in the
international oil market where U.S. production of oil is rising to be roughly on par with Saudi
Arabia and Russia. Recent economics literature of the tight (i.e., shale/unconventional) oil
expansion in the U.S. has a bearing on the issue of energy security as well. It could be that the
large expansion in shale oil has eroded the ability of OPEC to set world oil prices to some
degree, since OPEC cannot directly influence shale oil production decisions. Also, the growth in
U.S. oil supply is reducing the share of global oil supply controlled by OPEC, also possibly
limiting OPEC's degree of market power. But given that the shale oil expansion is a relatively
3 The Hurricanes Katrina/Rita in 2005 primarily caused a disruption in U.S. oil refinery production, with a more
limited disruption of some crude supply in the U.S. Gulf Coast area. Thus, the loss of petroleum product exceeded
the loss of crude oil, and the regional impact varied even within the U.S. The Katrina/Rita Hurricanes were a
different type of oil disruption event than is quantified in the Stanford EMF risk analysis framework, which provides
the oil disruption probabilities than ORNL is using.
3-21
-------
recent trend, it is difficult to know how much of an impact the increase in shale oil is having, or
will have, on OPEC behavior.
Two recent studies have examined the characteristics of tight oil supply that have relevance
for the topic of energy security. In the context of energy security, the question that arises is: can
tight oil respond to an oil price shock more quickly and substantially than conventional oil?76 If
so, then tight oil could potentially lessen the impacts of future oil shocks on the U.S. economy by
moderating the price increases from a future oil supply shock.
Newell and Prest (2019) look at differences in the price responsiveness for oil wells, using a
detailed dataset of 164,000 oil wells, during the time frame of 2000-2015 in five major oil-
producing states: Texas, North Dakota, California, Oklahoma, and Colorado.77 They find that
unconventional oil wells are more price responsive than conventional oil wells, mostly due to
their much higher productivity, but the estimated price elasticity is still small. Newell and Prest
also estimate a medium-run price elasticity of oil supply of 0.12. Newell and Prest note that the
shale oil supply response still takes more time to arise than is typically considered for a "swing
producer", referring to a supplier able to increase production quickly, within 30 to 90 days. In the
past, only Saudi Arabia and possibly one or two other oil producers in the Middle East, have
been able to ramp up oil production in a short period of time. From the standpoint of energy
security, the most relevant time frame of analysis is roughly a year, considered the short-run
responsiveness of oil demand to price.
Another study, by Bjornland et al. (2021), uses a well-level monthly production data set
covering more than 15,000 crude oil wells in North Dakota to examine differences in supply
responses between conventional and tight oil/shale oil.78 They find a short-run (i.e., one-month)
supply elasticity with respect to oil price for tight oil wells of 0.076, whereas the one-month
response of conventional oil supply was not statistically different from zero. Both the results
from the Newell and Prest and Bjornland et al. suggest that tight oil may have a larger supply
response to oil prices in the short-run than conventional oil, although the estimated short-run
elasticity is still small.
Finally, despite continuing uncertainty about oil market behavior and outcomes and the
sensitivity of the U.S. economy to oil shocks, it is generally agreed that it is beneficial to reduce
petroleum fuel consumption from an energy security standpoint. U.S. oil markets are expected to
remain tightly linked to trends in the world crude oil market. It is not just U.S. crude oil imports
alone, but both imports and consumption of petroleum from all sources and their role in
economic activity, that exposes the U.S. to risk from price shocks in the world oil price. The
relative significance of petroleum consumption and import levels for the macroeconomic
disturbances that follow from oil price shocks is not fully understood. Recognizing that changing
petroleum consumption will change U.S. imports, this assessment of oil costs focuses on those
incremental social costs that follow from the resulting changes in net imports, employing the
usual oil import premium measure.
3.2.3 Cost of Existing U.S. Energy Security Policies
An additional often-identified component of the full economic costs of U.S. oil imports is the
costs to the U.S. taxpayers of existing U.S. energy security policies. The two primary examples
are maintaining the Strategic Petroleum Reserve (SPR) and maintaining a military presence to
help secure a stable oil supply from potentially vulnerable regions of the world.
3-22
-------
The SPR is the largest stockpile of government-owned emergency crude oil in the world.
Established in the aftermath of the 1973/1974 oil embargo, the SPR provides the U.S. with a
response option should a disruption in commercial oil supplies threaten the U.S. economy. While
the costs for building and maintaining the SPR are more clearly related to U.S. oil use and
imports, historically these costs have not varied in response to changes in U.S. oil import levels.
Thus, while the effect of the SPR in moderating price shocks is factored into the analysis that
EPA is using to estimate the macroeconomic oil security premiums, the cost of maintaining the
SPR is excluded.
EPA also has considered the possibility of quantifying the military benefits components of
energy security but has not done so here for several reasons. The literature on the military
components of energy security has described four broad categories of oil-related military and
national security costs, all of which are hard to quantify. These include possible costs of U.S.
military programs to secure oil supplies from unstable regions of the world, the energy security
costs associated with the U.S. military's reliance on petroleum to fuel its operations, possible
national security costs associated with expanded oil revenues to "rogue states" and relatedly the
foreign policy costs of oil insecurity.
Of these categories listed above, the one that is most clearly connected to petroleum use and
is, in principle, quantifiable is the first: the cost of military programs to secure oil supplies and
stabilize oil supplying regions. There is an ongoing literature on the measurement of this
component of energy security, but methodological and measurement issues - attribution and
incremental analysis - pose two significant challenges to providing a robust estimate of this
component of energy security. The attribution challenge is to determine which military programs
and expenditures can properly be attributed to oil supply protection, rather than some other
objective. The incremental analysis challenge is to estimate how much the petroleum supply
protection costs might vary if U.S. oil use were to be reduced or eliminated. Methods to address
both of these challenges are necessary for estimating the effect on military costs arising from a
modest reduction (not elimination) in oil use attributable to this final rule.
Since "military forces are, to a great extent, multipurpose and fungible" across theaters and
missions (Crane et al. 2009), and because the military budget is presented along regional
accounts rather than by mission, the allocation to particular missions is not always clear.79
Approaches taken usually either allocate "partial" military costs directly associated with
operations in a particular region, or allocate a share of total military costs (including some that
are indirect in the sense of supporting military activities overall) (Koplow and Martin 1998).80
The challenges of attribution and incremental analysis have led some to conclude that the
mission of oil supply protection cannot be clearly separated from others, and the military cost
component of oil security should be taken as near zero (Moore et al. 1997).81
Stern (2010), on the other hand, argues that many of the other policy concerns in the Persian
Gulf follow from oil, and the reaction to U.S. policies taken to protect oil.82 Stern presents an
estimate of military cost for Persian Gulf force projection, addressing the challenge of cost
allocation with an activity-based cost method. He uses information on actual naval force
deployments rather than budgets, focusing on the costs of carrier deployment. As a result of this
different data set and assumptions regarding allocation, the estimated costs are much higher,
roughly 4 to 10 times, than other estimates. Stern also provides some insight on the analysis of
incremental effects, by estimating that Persian Gulf force projection costs are relatively strongly
3-23
-------
correlated to Persian Gulf petroleum export values and volumes. Still, the issue remains of the
marginality of these costs with respect to Persian Gulf oil supply levels, the level of U.S. oil
imports, or U.S. oil consumption levels.
Delucchi and Murphy (2008) seek to deduct from the cost of Persian Gulf military programs
the costs associated with defending U.S. interests other than the objective of providing more
stable oil supply and price to the U.S. economy.83 Excluding an estimate of cost for missions
unrelated to oil, and for the protection of oil in the interest of other countries, Delucchi and
Murphy estimated military costs for all U.S. domestic oil interests of between $24 and $74
billion annually. Delucchi and Murphy assume that military costs from oil import reductions can
be scaled proportionally, attempting to address the incremental issue.
Crane et al. considers force reductions and cost savings that could be achieved if oil security
were no longer a consideration. Taking two approaches and guided by post-Cold War force draw
downs and by a top-down look at the current U.S. allocation of defense resources, they
concluded that $75—$91 billion, or 12-15 percent of the current U.S. defense budget, could be
reduced.
Finally, an Issue Brief by Securing America's Future Energy (SAFE) (2018) found a
conservative estimate of approximately $81 billion per year spent by the U.S. military protecting
global oil supplies.84 This is approximately 16 percent of the recent U.S. Department of
Defense's budget. Spread out over the 19.8 million barrels of oil consumed daily in the U.S. in
2017, SAFE concludes that the implicit subsidy for all petroleum consumers is approximately
$11.25 per barrel of crude oil, or $0.28 per gallon. According to SAFE, a more comprehensive
estimate suggests the costs could be greater than $30 per barrel, or over $0.70 per gallon.85
As in the examples above, an incremental analysis can estimate how military costs would vary
if the oil security mission is no longer needed, and many studies stop at this point. It is
substantially more difficult to estimate how military costs would vary if U.S. oil use or imports
are partially reduced, as is projected to be a consequence of this final rule. Partial reduction of
U.S. oil use surely diminishes the magnitude of the security problem, but there is uncertainty that
supply protection forces and their costs could be scaled down in proportion, and there remains
the associated goal of protecting supply and transit for U.S. allies and other importing countries,
if they do not decrease their petroleum use as well.86 While military costs are an important
consideration, EPA continues to be unaware of a robust methodology for assessing the effect on
military costs of a partial reduction in U.S. oil use and imports. Therefore, we do not include
military cost impacts in EPA's benefit and cost analysis for this final rule.
3.2.4 U.S. Oil Import Reductions from this Final Rule
Over the time frame of analysis for this final rule, 2023-2050, the U.S. Department of
Energy's (DOE) Energy Information Administration's (EIA) Annual Energy Outlook (AEO)
2021 (Reference Case) projects that the U.S. will be both an exporter and an importer of crude
oil.87 The U.S. produces more light crude oil than its refineries can refine. Thus, the U.S. exports
lighter crude oil and imports heavier crude oils to satisfy the needs of U.S. refineries, which are
configured to efficiently refine heavy crude oil. U.S. crude oil exports are projected to be
gradually increasing from 3 million barrels per day (MMBD) in 2023 to 3.5 MMBD in 2026 and
remain above 3 MMBD through 2050. U.S. crude oil imports, meanwhile, are projected to
3-24
-------
decline modestly from 7.8 MMBD in 2023 to 7.5 MMBD in 2026. U.S. crude oil imports
continue to decrease modestly to 6.9 MMBD by 2030, before rising to the 7.6 MMBD in 2050.
The AEO 2021 projects that U.S. net oil product exports will be 5.3 MMBD in 2023 and rise
modestly to 5.6 MMBD in 2026. After 2026, U.S. net oil product exports are projected to be
somewhat greater than five MMBD until 2045, before decreasing modestly to 4.6 MMBD in
2050. Given the pattern of U.S. crude oil exports/imports, and U.S. net oil product exports, the
U.S. is projected to be a net petroleum (crude oil and product) exporter from 2023 through 2050.
For example, from 2023 to 2026, the U.S. net crude oil and product exports increase steadily
from 0.5 to 1.6 MMBD. U.S. net crude oil and product exports increase to roughly 2 MMBD in
the 2030 to 2035 time frame, before tapering off to 0.1 MMBD by 2050.
Since the U.S. is projected to continue importing significant quantities of crude oil through
2050, EPA's assessment is that the U.S. it is not expected to achieve the overall goal of U.S.
energy independence during the analytical time frame of this rule. However, the U.S. is projected
to be a net exporter of crude oil and products through 2050.
U.S. oil consumption is projected to be fairly steady for the time period from 2023 to 2050.
From 2023 to 2040, projected U.S. oil consumption is fairly constant at roughly 20 MMBD
before increasing modestly to roughly 21 MMBD in the 2045-2050 time period. During the
2023-2050 time frame, the AEO projects that the U.S. will continue to consume significant
quantities of oil and will likewise continue to rely on significant quantities of crude oil imports.
Estimated fuel consumption changes from this final GHG rule are presented in Chapter 5.2.
Based on a detailed analysis of differences in U.S. fuel consumption, crude oil imports/exports
and exports of petroleum products for the time frame 2023-2050, and using the AEO 2021
(Reference Case) and two alternative sensitivity cases, i.e., (Low Economic Growth) and (High
Economic Growth), EPA estimates that approximately 91 percent of the change in fuel
consumption resulting from the final LDV GHG standards is likely to be reflected in reduced
U.S. imports of crude oil.k The 91 percent oil import factor is calculated by taking the ratio of the
changes in U.S. net crude oil and product imports divided by the change in U.S. oil consumption
in the different AEO cases. Thus, on balance, each gallon of petroleum reduced as a result of the
final LDV GHG Rule is anticipated to reduce total U.S. imports of petroleum by 0.91 gallons.
Based upon the changes in fuel consumption estimated in Chapter 5.2 and the 91 percent oil
import reduction factor, the reduction in U.S. oil imports as a result of the final LDV GHG
standards are estimated in Table 3-5 below for the 2023-2050 time frame. For comparison
purposes, based upon the AEO 2021 (Reference Case), Table 3-5 also shows the U.S.'s projected
crude oil exports and imports, net oil product exports, net crude oil/product exports and U.S. oil
consumption for the years 2023-2050.88
k We looked at changes in U.S. crude oil imports/exports and net petroleum products in the AEO 2021 Reference
Case, Table 11. Petroleum and Other Liquids Supply and Disposition, in comparison to two alternative cases from
the AEO 2021. See the spreadsheet, "AEO2021 Change in oil product demand on imports".
3-25
-------
Table 3-5: Projected Trends in U.S. Oil Exports/Imports, Net Oil Product Exports, Net Crude Oil/Product
Exports, Oil Consumption and U.S. Oil Import Reductions Resulting from the Final LDV GHG rule from
2023 to 2050 (Millions of barrels per day (MMBD))*




U.S. Net Crude

U.S. Oil Import

U.S. Crude
U.S. Crude
U.S.Net Oil
Oil and Product
U.S. Oil
Reductions from
Year
Oil Exports
Oil Imports
Product Exports
Exports
Consumption
Final Rule*
2023
3.0
7.8
5.3
0.5
20.0
0.0
2024
3.4
7.8
5.4
0.9
20.1
0.1
2025
3.3
7.5
5.6
1.4
20.2
0.1
2026
3.5
7.5
5.6
1.6
20.2
0.2
2030
3.1
6.9
5.9
2.0
20.2
0.4
2035
3.3
7.0
5.6
1.9
20.4
0.7
2040
3.2
7.5
5.5
1.2
20.6
0.8
2045
3.1
7.3
5.1
1.0
21.0
0.9
2050
3.1
7.6
4.6
0.1
21.6
0.9
* Chapter 5.2 presents the total barrels of oil reduced due to the final standards. Here we present the barrels of
imported oil reduced (per day). The values shown here account for the estimated oil import reduction as percent of
total oil demand reduction (91 percent) and divides by 365 days in a year.


3.2.5 Oil Security Premiums Used for this Final Rule
In order to understand the energy security implications of reducing U.S. oil imports, EPA has
worked with Oak Ridge National Laboratory (ORNL), which has developed approaches for
evaluating the social costs and energy security implications of oil use. The energy security
estimates provided below are based upon a methodology developed in a peer-reviewed study
entitled, "The Energy Security Benefits of Reduced Oil Use, 2006-2015," completed in 2008.89
This ORNL study is an updated version of the approach used for estimating the energy security
benefits of U.S. oil import reductions developed in a 1997 ORNL Report.90 This approach has
been used to estimate energy security benefits for the LDV GHG and fuel economy standards
(2012-2016)/(2017-2025) and the HDV GHG/fuel economy standards Phase I (2014-2018) and
Phase II (2018 and later).91'92'93
When conducting this analysis, ORNL considers the full cost of importing petroleum into the
U.S. The full economic cost (i.e., labeled oil security premiums below) is defined to include two
components in addition to the purchase price of petroleum itself. These are: (1) the higher
costs/benefits for oil imports resulting from the effect of U.S. demand on the world oil price (i.e.,
the "demand" or "monopsony" costs/benefits); and (2) the risk of reductions in U.S. economic
output and disruption to the U.S. economy caused by sudden disruptions in the supply of
imported oil to the U.S. (i.e., the avoided macroeconomic disruption/adjustment costs).
For this final LDV GHG rule, EPA is using updated oil security premium values estimated
using ORNL's methodology, which incorporates the oil price projections and energy market and
economic trends, particularly regional oil supplies and demands at a global level (i.e.,
U.S./OPEC/rest of the world), from the AEO 2021 into its model. For the proposed LDV GHG
rule, we used the AEO 2018 for estimating the oil security premiums. The macroeconomic oil
security premium values in this final LDV rule, based on AEO 2021 data, are 10-15 percent
lower than the ones in the proposed LDV rule using AEO 2018. The smaller values result from
lower projections of U.S. crude oil net import levels and U.S. crude oil import prices in AEO
2021 relative to AEO 2018. The projected 2023-2026 average oil import cost for the U.S. is 28
3-26
-------
percent lower in the AEO 2021 than the AEO 2018 (19 percent lower for 2023-2050) and the
projected AEO 2021 U.S. net crude oil imports are also 28 percent lower, on average, for 2023-
2026, than in AEO 2018 (28 percent lower as well for 2023-2050).
EPA only considers the avoided macroeconomic disruption/adjustment costs oil security
premiums (i.e., labeled macroeconomic oil security premiums below), since the monopsony
impacts of this final rule are considered transfer payments. In previous LDV GHG rules when
the U.S. was forecasted by the U.S. Energy Information Administration (EIA) to be a net
importer of crude oil and product, monopsony impacts represented reduced payments by U.S.
consumers to oil producers outside of the U.S. There was some debate among economists as to
whether the U.S. exercise of its monopsony power in oil markets, for example from the
implementation of LDV GHG rules, was a "transfer payment" or a "benefit". Given the
redistributive nature of this monopsony impact from a global perspective, and since there are no
changes in resource production costs when the U.S. exercises it monopsony power, some
economists argued that it is a transfer payment. Other economists argued that monopsony
impacts were a benefit since they partially address, and partially offset, the market power of
OPEC. In previous EPA LDV GHG rules, after weighing both countervailing arguments, EPA
concluded that the U.S.'s exercise of its monopsony power was a transfer payment, and not a
benefit.94
In the context of this LDV GHG rule, the U.S.'s oil trade balance is quite a bit different than
in previous LDV GHG rules. The U.S. is projected to be a net exporter of oil in the time frame of
this analysis of this rule, 2023-2050. As a result, reductions in U.S. oil consumption and, in turn,
U.S. oil imports, still lower the world oil price modestly. But the net effect of the lower world oil
price is now a decrease in revenue for U.S. exporters of crude oil and products, instead of a
decrease in payments to foreign oil producers. The argument that monopsony impacts address
the market power of OPEC is no longer appropriate. Thus, EPA continues to consider the U.S.
exercise of monopsony power to be transfer payments. As a result, EPA does not believe that
excluding monopsony effects stemming from this rule results in an underestimate of the energy
security benefits of this rule.
For this rule, EPA and ORNL worked together to develop oil security premiums based upon
the recent energy security literature on this topic. EPA is continuing to use the same oil security
premium methodology for this final rule as it used in the proposal. The recent economics
literature (discussed in Section 3.2.2 above) focuses on three factors that can influence the
macroeconomic oil security premiums. We discuss each factor below and provide a rationale for
how we are updating two out of three of the factors to develop new estimates of the
macroeconomic oil security premiums. We are not accounting for how shale oil is influencing
the macroeconomic oil security premiums in this final rule.
First, we assess the price elasticity of demand for oil. In previous EPA vehicle rulemakings,
EPA has used a short-run elasticity of demand for oil of -0.045.95 From the recent RFF study, the
"blended" price elasticity of demand for oil is -0.05. The ORNL meta-analysis estimate of this
parameter is -0.07. We find the elasticity estimates from what RFF characterizes as the "new
literature," -0.175, and from the "new models" that RFF uses, -0.20 to -0.33, somewhat high.
Most of the world's oil demand is concentrated in the transportation sector and there are
currently limited alternatives to oil use in this sector. According to the IEA, the share of global
oil consumption attributed to the transportation sector grew from 60 percent in 2000 to 66
3-27
-------
percent in 2018.96 The next largest sector by oil consumption, and an area of recent growth, is
petrochemicals. Thus, we believe it would be surprising if short-run oil demand responsiveness
has changed in a dramatic fashion. Increases in future electric vehicle use could influence the
price elasticity of demand for oil, but there is little empirical evidence available to assess this
issue. We may attempt to address this issue in the future if new information and data becomes
available.
The ORNL meta-analysis estimate encompasses the full range of the economics literature on
this topic and develops a meta-analysis estimate from the results of many different studies in a
structured way, while the RFF study's "new models" results represent only a small subset of the
economics literature's estimates. Thus, for the analysis of this final rule, we are increasing the
short-run price elasticity of demand for oil from -0.045 to -0.07, a 56 percent increase.1 This
increase has the effect of lowering the macroeconomic oil security premiums estimated for this
rulemaking.
Second, we consider the elasticity of GDP to an oil price shock. For previous EPA vehicle
rulemakings, a GDP elasticity to an oil shock of -0.032 was used.97 The RFF "blended" GDP
elasticity is -0.028, the RFF's "new literature" GDP elasticity is -0.018, while the RFF "new
models" GDP elasticities range from -0.007 to -0.027. The ORNL meta-analysis GDP elasticity
is -0.021, a 35 percent reduction from the GDP elasticity used in previous EPA rulemakings. We
believe that the ORNL meta-analysis value is representative of the recent literature on this topic
since it considers a wide range of recent studies and does so in a structured way. Also, the ORNL
meta-analysis estimate is within the range of GDP elasticities of RFF's "blended" and "new
literature" elasticities.
For the proposed rule, energy security premiums were developed using the ORNL
methodology, the AEO 2018 and a GDP elasticity of -0.023. For this final rule, in addition to
updating the energy security premium estimates to use the most recently available AEO (i.e.,
AEO 2021), we have also updated the GDP elasticity to -0.021, the value from the ORNL meta-
analysis discussed above. These updates resulted in a lower per-barrel energy security premium
than was used in the analysis for the proposed rule. Note, however, that the overall energy
security benefits are greater in this final rule than in the proposal because the final standards are
estimated to result in a greater reduction of gasoline consumption, which more than offsets the
decrease to the per-barrel premium. Finally, we have not factored in how increases in U.S. tight
oil might influence U.S. oil security values, other than how they significantly reduce net oil
imports, given the complexity of this issue.
Table 3-6 below provides estimates of ORNL's macroeconomic oil security premiums for
selected years from 2023-2050 based upon the AEO 2021. In terms of cents per gallon, the
macroeconomic oil security premiums range from 7.5 cents/gallon in 2023 to 7.7 cents/gallon in
2026. In the later years of the time frame of this analysis, the macroeconomic oil security
premiums range from 8.1 cents/gallon in 2030 to 11.8 cents/gallon in 2050.
1 EPA and ORNL worked together to develop an updated estimate of the short-run elasticity of demand for oil for
use in the ORNL model.
3-28
-------
Table 3-6: Macroeconomic Oil Security Premiums for Selected Years from 2023-2050 (2018$/Barrel)*
Year (range)
Macroeconomic Oil Security Premiums
(Range)
2023
$3.15
($0.92-$5.71)
2024
$3.17
($0.84 - $5.80)
2025
$3.18
($0.77 - $5.89)
2026
$3.23
($0.74 - $6.00)
2030
$3.41
($0.62-$6.41)
2035
$3.76
($0.70 - $7.05)
2040
$4.21
($1.04-$7.77)
2045
$4.54
($1.18-$8.29)
2050
$4.94
($1.46-$8.91)
* Top values in each cell are the midpoints, the values in parentheses are the 90 percent confidence
intervals. The macroeconomic oil security premium estimates for the years 2023, 2024 and 2026 are
linearly interpolated values from ORNL estimates, which are reported in five-year time intervals.
3.2.6 Energy Security Benefits of the Final Rule
Using the ORNL oil security premium methodology with: (1) estimated oil savings calculated
by EPA, (2) an oil import reduction factor of 91 percent, and (3) updated oil security premium
estimates based upon the recent energy security literature and the AEO 2021, EPA presents the
annual energy security benefits of the final LDV GHG standards for selected years from 2023-
2050 in Table 3-7 below. We do not consider the monopsony effect or military cost impacts of
oil import changes in the energy security benefits provided below.
Table 3-7: Annual Energy Security Benefits of the Final LDV GHG/Fuel Economy Rule for Selected Years
2023-2050 (in Billions of 2018$)
Year
Benefits (2018$)
2023
$0.03
2026
$0.18
2030
$0.51
2035
$0.92
2040
$1.3
2050
$1.6
PV, 3%
$14
PV, 7%
$7
Annualized, 3%
$0.73
Annualized, 7%
$0.56
3-29
-------
3.3 Social Cost of Greenhouse Gases
We estimate the climate benefits for this rulemaking using measures of the social cost of three
greenhouse gases: carbon, methane, and nitrous oxide. The social cost of each gas (i.e., the social
cost of carbon (SC-CO2), methane (SC-CH4), and nitrous oxide (SC-N2O)) is the monetary value
of the net harm to society associated with a marginal increase in emissions in a given year, or the
benefit of avoiding that increase. Collectively, these values are referenced as the "social cost of
greenhouse gases" (SC-GHG). In principle, SC-GHG 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-
GHG therefore, reflects the societal value of reducing emissions of the gas in question by one
metric ton. The SC-GHG is the theoretically appropriate values to use in conducting benefit-cost
analyses of policies that affect CO2, CH4, and N2O emissions.
We estimate the global social benefits of CO2, CH4, and N2O emission reductions expected
from this final rule using the SC-GHG estimates presented in the Technical Support Document:
Social Cost of Carbon, Methane, and Nitrous Oxide Interim Estimates under Executive Order
13990. We have evaluated the SC-GHG estimates in the TSD and have determined that these
estimates are appropriate for use in estimating the global social benefits of CO2, CH4, and N2O
emission reductions expected from this final rule.98 These SC-GHG estimates are interim values
developed for use in benefit-cost analyses until updated estimates of the impacts of climate
change can be developed based on the best available science and economics. After considering
the TSD, and the issues and studies discussed therein, EPA finds that these estimates, while
likely an underestimate, are the best currently available SC-GHG estimates. The SC-GHG
estimates used in this RIA are the same as those used in the July 2016 Draft Technical
Assessment Report: Midterm Evaluation of Light-Duty Vehicle Greenhouse Gas Emission
Standards and Corporate Average Fuel Economy Standards for Model Years 2022-2025,
adjusted for inflation to 2018 dollars.
The SC-GHG 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. 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.99'100'101 In August 2016 the IWG published estimates of the social cost of methane (SC-
CH4) and nitrous oxide (SC-N2O) using methodologies that are consistent with the methodology
underlying the SC-CO2 estimates. 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
3-30
-------
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.102 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 SAFE rule RIAm, 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
and other greenhouse gases reflect the best available science and the recommendations of the
National Academies.103 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.
As noted above, EPA participated in the IWG but has also independently evaluated the interim
SC-GHG estimates published in February 2021 and determined they are appropriate to use here
to estimate the climate benefits for this final rule. EPA and other agencies intend to undertake a
fuller update of the SC-GHG estimates that takes into consideration the advice of the National
Academies and other recent scientific literature.103
The EPA has also evaluated the content of the TSD, including the studies and methodological
issues discussed therein and concludes that it agrees with the rationale for these estimates
presented in the TSD and summarized below.
In particular, the IWG found that the SC-GHG estimates used under E.O. 13783 fail to reflect
the full impact of GHG emissions in multiple ways. First, the IWG 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, U.S. military assets and interests abroad, and tourism, and spillover pathways such as
economic and political destabilization and global migration that can lead to adverse impacts on
U.S. national security, public health, and humanitarian concerns. In addition, assessing the
benefits of U.S. GHG mitigation activities requires consideration of how those actions may
m The values used in the SAFE rule RIA were interim values developed under E.O. 13783 for use in regulatory
analyses. EPA followed E.O. 13783 in the SAFE rule by using SC-CO2 estimates reflecting impacts occurring
within U.S. borders and 3% and 7% discount rates in our central analysis for the SAFE rule RIA.
3-31
-------
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.
In addition, a wide range of scientific and economic experts have emphasized the issue of
reciprocity as support for considering global damages of GHG emissions. Using a global
estimate of damages in U.S. analyses of regulatory actions allows the U.S. to continue to actively
encourage other nations, including emerging major economies, to take significant steps to reduce
emissions. The only way to achieve an efficient allocation of resources for emissions reduction
on a global basis—and so benefit the U.S. and its citizens—is for all countries to base their
policies on global estimates of damages.
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.
Furthermore, as an empirical matter, the development of a domestic SC-GHG is greatly
complicated by the relatively few region- or country-specific estimates of the SC-C02 in the
literature. At present, the only quantitative characterization of domestic damages from GHG
emissions is based on the share of damages arising from climate impacts occurring within U.S.
borders as represented in current IAMs. This is both incomplete and an underestimate of the
share of total damages that accrue to the citizens and residents of the U.S. because these models
do not capture the regional interactions and spillovers discussed above. EPA, as a member of 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.
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 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.n'103'103'104'105'106 Furthermore,
the damage estimates developed for use in the SC-GHG are estimated in consumption-equivalent
terms, and so an application of OMB Circular A-4's guidance for regulatory analysis would then
use the consumption discount rate to calculate the SC-GHG. EPA agrees with this assessment
and will continue to follow developments in the literature pertaining to this issue. EPA also notes
that while OMB Circular A-4, as published in 2003, recommends using 3% and 7% discount
rates as "default" values, Circular A-4 also reminds agencies that "different regulations may call
11 GHG emissions are stock pollutants, where damages are associated with what has accumulated in the atmosphere
over time, and they are long lived such that subsequent damages resulting from emissions today occur over many
decades or centuries depending on the specific greenhouse gas under consideration. In calculating the SC-GHG, the
stream of future damages to agriculture, human health, and other market and non-market sectors from an additional
unit of emissions are estimated in terms of reduced consumption (or consumption equivalents). Then that stream of
future damages is discounted to its present value in the year when the additional unit of emissions was released.
Given the long time horizon over which the damages are expected to occur, the discount rate has a large influence
on the present value of future damages.
3-32
-------
for different emphases in the analysis, depending on the nature and complexity of the regulatory
issues and the sensitivity of the benefit and cost estimates to the key assumptions." On
discounting, Circular A-4 recognizes that "special ethical considerations arise when comparing
benefits and costs across generations," and Circular A-4 acknowledges that analyses may
appropriately "discount future costs and consumption benefits.. .at a lower rate than for
intragenerational analysis." In the 2015 Response to Comments on the Social Cost of Carbon for
Regulatory Impact Analysis, OMB, EPA, and the other IWG members recognized that "Circular
A-4 is a living document" and "the use of 7 percent is not considered appropriate for
intergenerational discounting. There is wide support for this view in the academic literature, and
it is recognized in Circular A-4 itself." Thus, EPA concludes that a 7% discount rate is not
appropriate to apply to value the social cost of greenhouse gases in this regulatory analysis. In
this analysis, to calculate the present and annualized values of climate benefits, EPA uses the
same discount rate as the rate used to discount the value of damages from future GHG emissions,
for internal consistency. That approach to discounting follows the same approach that the
February 2021 TSD recommends "to ensure internal consistency—i.e., future damages from
climate change using the SC-GHG at 2.5 percent should be discounted to the base year of the
analysis using the same 2.5 percent rate." EPA has also consulted the National Academies' 2017
recommendations on how SC-GHG estimates can "be combined in RIAs with other cost and
benefits estimates that may use different discount rates." The National Academies reviewed
"several options," including "presenting all discount rate combinations of other costs and benefits
with [SC-GHG] estimates." In Table 10-6, EPA presents all combinations of the SC-GHG values
at the different discount rates appropriate to climate effects (2.5%, 3%, and 5%) together with
other costs and benefits discounted at the 3% and 7% rates, consistent with the options outlined
by the National Academies.
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 adopted as interim estimates 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
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 percent, 3 percent, and 5 percent), 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 percent 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.
3-33
-------
Table 3-8, Table 3-9, and Table 3-10 summarize the interim global SC-CO2, SC-CH4, and
SC-N2O estimates for the years 2015 to 2070.° These estimates are reported in 2018 dollars but
are otherwise identical to those presented in the IWG's 2016 TSD. For purposes of capturing
uncertainty around the SC-GHG estimates in analyses, the IWG's February 2021 TSD
emphasizes the importance of considering all four of the SC-GHG values. The SC-GHG
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 3-8: Interim Global Social Cost of Carbon Values, 2020-2070 (2018$/Metric Tonne CO2)99
Emissions

Discount Rate and Statistic

Year





5%
3%
2.5%
3%

Average
Average
Average
95th Percentile
2020
$14
$50
$74
$147
2025
$16
$55
$81
$164
2030
$19
$60
$87
$181
2035
$22
$66
$93
$200
2040
$24
$71
$100
$218
2045
$28
$77
$107
$235
2050
$31
$82
$113
$252
2055
$34
$86
$119
$258
2060
$37
$91
$124
$268
2065
$42
$98
$132
$292
2070
$48
$105
$139
$318
Note: The 2020-2050 SC-CO2 values are identical to those reported in the 2016 TSD (IWG 2016a) adjusted for
inflation to 2018 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 estimates were extended for the period 2051 to 2070
using methods, assumptions, and parameters identical to the 2020-2050 estimates. The values are stated in
$/metric tonne CO2 and vary depending on the year of C02 emissions. This table displays the values rounded to
the nearest dollar; the annual unrounded values through 2050 are available on OMB's website:
. The annual unrounded 2051-2070 values used in the calculations in
this RIA are available in the rule docket.


0 The February 2021 TSD provides SC-GHG estimates through emissions year 2050. Estimates were extended for
the period 2051 to 2070 using the IWG methods, assumptions, and parameters identical to the 2020-2050 estimates.
Specifically, 2051-2070 SC-GHG estimates were calculated in Mimi.jl, an open-source modular computing platform
used for creating, running, and performing analyses on IAMs (www.mimiframework.org). For CO2, the 2051-2054
SC-GHG values were calculated by linearly interpolating between the 2050 TSD values and the 2055 Mimi-based
values. The annual unrounded 2051-2070 values used in the calculations in this RIA are available in the rule docket,
and the replication code is available upon request.
3-34
-------
Table 3-9: Interim Global Social Cost of Methane Values, 2020-2070 (2018$/Metric Tonne CH4)99
Emissions

Discount Rate and Statistic

Year





5%
3%
2.5%
3%

Average
Average
Average
95th Percentile
2020
$650
$1,400
$1,900
$3,800
2025
$780
$1,700
$2,200
$4,400
2030
$910
$1,900
$2,400
$5,000
2035
$1,100
$2,200
$2,700
$5,800
2040
$1,200
$2,400
$3,100
$6,500
2045
$1,400
$2,700
$3,400
$7,200
2050
$1,600
$3,000
$3,700
$7,900
2055
$1,700
$3,100
$3,800
$8,100
2060
$1,800
$3,300
$4,000
$8,300
2065
$2,400
$4,100
$4,800
$11,000
2070
$3,000
$4,800
$5,700
$14,000
Note: The 2020-2050 SC-CH4 values are identical to those reported in the 2016 TSD (IWG 2016a) adjusted for
inflation to 2018 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 estimates were extended for the period 2051 to 2070
using methods, assumptions, and parameters identical to the 2020-2050 estimates. The values are stated in
$/metric tonne CH4 and vary depending on the year of CH4 emissions. This table displays the values rounded to
the nearest dollar; the annual unrounded values through 2050 are available on OMB's website:
. The annual unrounded 2051-2070 values used in the calculations in
this RIA are available in the rule docket.


Table 3-10: Interim Global Social Cost of Nitrous Oxide Values, 2020-2070 (2018$/Metric Tonne N2O)99
Emissions

Discount Rate and Statistic

Year





5%
3%
2.5%
3%

Average
Average
Average
95th Percentile
2020
$5,600
$18,000
$26,000
$47,000
2025
$6,600
$20,000
$29,000
$53,000
2030
$7,600
$22,000
$32,000
$59,000
2035
$8,800
$24,000
$35,000
$65,000
2040
$10,000
$27,000
$38,000
$72,000
2045
$11,000
$29,000
$41,000
$79,000
2050
$13,000
$32,000
$44,000
$86,000
2055
$14,000
$35,000
$47,000
$92,000
2060
$16,000
$37,000
$50,000
$98,000
2065
$19,000
$42,000
$55,000
$110,000
2070
$22,000
$46,000
$60,000
$130,000
Note: The 2020-2050 SC-N20 values are identical to those reported in the 2016 TSD (IWG 2016a) adjusted for
inflation to 2018 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 estimates were extended for the period 2051 to 2070
using methods, assumptions, and parameters identical to the 2020-2050 estimates. The values are stated in
$/metric tonne N20 and vary depending on the year of N20 emissions. This table displays the values rounded to
the nearest dollar; the annual unrounded values through 2050 are available on OMB's website:
. The annual unrounded 2051-2070 values used in the calculations in
this RIA are available in the rule docket.


3-35
-------
There are a number of limitations and uncertainties associated with the SC-GHG estimates
presented in Table 3-8 through Table 3-10. 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 3-1, Figure 3-2, and Figure 3-3 present the quantified sources of uncertainty in the form
of frequency distributions for the SC-CO2, SC-CFU, and SC-N2O estimates for emissions in
2030. The distributions of SC-GHG 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-GHG estimates for each discount
rate. As illustrated by the figures, the assumed discount rate plays a critical role in the ultimate
estimate of the SC-GHG. This is because GHG 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.
CN
O
CO
£=
O
E
CO
M—
o
c
o
u
03
5% Average = $19
OsJ
o
o
o
3% Average = $60
i 2.5% Average = $87
3%
95th Pet.
$181
I
IBBBBbbbu I
yQDQyQHBB
~
~
Discount Rate
~
5.0%
~
3.0%
~
2.5%
5th - 95ln Percentile
of Simulations
i ii ii i i i i i
=th
I I
I II I I I I I I I I I I I I I I I I I II II I I II
20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320
Social Cost of Carbon in 2030 [2018$ / metric ton COz]
Figure 3-1: Frequency Distribution of SC-CO2 Estimates for 2030 p
p Although the distributions and numbers are based on the full set of model results (150,000 estimates for each
discount rate and gas), for display purposes the horizontal axis is truncated with 0.02 to 0.68 percent of the estimates
falling below the lowest bin displayed and 0.12 to 3.11 percent of the estimates falling above the highest bin
displayed, depending on the discount rate and GHG.
3-36
-------
10
c
o
E
CO
4—
o
CN1
O
5% Average = $940
3% Average = $2000
2.5% Average = $2500
3%
95th Pet. = $5200
l~: :i :ca-t	¦ —
L
L
II
I I I I II I I
0 400
1 I I III
1200
i i i i :i i i
2000
l I: ITI I
2800
I I II I
3600
TTTTTr
4400
I I 1 I 11
5200
[M il I
6000
Discount Rate
m
~
5.0%
0 1
~
3.0%
1
1
~
2.5%
I 5th - 95'" Percentile
] J of Simulations
111111 ri i rrn ii 1111111
6800 7600 8400
i
Social Cost of Methane in 2030 [2018$ / metric ton CH4]
Figure 3-2: Frequency Distribution of SC-CH4 Estimates for 20301
o
-------
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 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-GHG estimates. However, as discussed in the February 2021 TSD, the IWG has
recommended that, taken together, the limitations suggest that the SC-GHG estimates used in
this final rule likely underestimate the damages from GHG emissions. EPA concurs that the
values used in this rulemaking conservatively underestimate the rule's climate benefits. In
particular, the Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report,
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.107 Since then, the peer-reviewed literature has continued
to support this conclusion, as noted in the IPCC's Fifth Assessment report and other recent
scientific assessments.108'109'110'111'112'113'114'115 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.108
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.114 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-GHG estimates. The IWG is currently working on a comprehensive update of the
SC-GHG estimates taking into consideration recommendations from the National Academies of
Sciences, Engineering and Medicine, recent scientific literature, public comments received on
the February 2021 TSD and other input from experts and diverse stakeholder groups.116
3-38
-------
Table 3-11 through Table 3-13 shows the estimated global climate benefits from changes in
CO2, CH4, N2O, respectively and Table 3-14 through Table 3-16 shows the combined total
climate benefits expected to occur over 2023-2070 under the final revised GHG standards and
the two alternatives analyzed (see Chapter 2.2.2 and also Preamble Section II.C for more detail
on the alternatives considered by EPA). EPA estimated the dollar value of the GHG-related
effects for each analysis year between 2023 through 2050 by applying the SC-GHG estimates,
shown in Tables 3-8 through 3-10, to the estimated changes in GHG emissions inventories
resulting from including tailpipe emissions from light-duty cars and trucks, and the upstream
emissions associated with the fuels used to power those vehicles.q EPA then calculated the
present value and annualized benefits from the perspective of 2021 by discounting each year-
specific value to the year 2021 using the same discount rate used to calculate the SC-GHG.
q According 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. Circular A-4 also
reminds analysts that "[d]ifferent regulations may call for different emphases in the analysis, depending on the
nature and complexity of the regulatory issues." 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, how U.S. GHG mitigation
activities affect mitigation activities by other countries, and spillover effects from climate action elsewhere. The SC-
GHG estimates used in regulatory analysis under revoked E.O. 13783, including in the RIA for the SAFE rule, were
an attempt to approximate the climate damages occurring within U.S. borders only (e.g., $7/mtC02 and $1 l/mtC02
(2018 dollars) using a 3% discount rate for emissions occurring in 2023 and 2050, respectively; $207/mtCH4 and
$376/mtCH4 (2018 dollars) using a 3% discount rate for emissions occurring in 2023 and 2050, respectively; and
$2437/mtN20 and $3986/mtN20 (2018 dollars) using a 3% discount rate for emissions occurring in 2023 and 2050,
respectively). Applying the same estimates that were used in the SAFE rule (based on a 3% discount rate) to the
GHG emission reduction expected from this final rule would yield benefits from climate impacts within U.S borders
of $37 million in 2023, increasing to $1.9 billion in 2050 for C02; $1 million in 2023, increasing to $67 million in
2050 for CH4; $0.3 million in 2023, increasing to $15 million in 2050 for N2O; and combined GHG benefits of $38
million in 2023, increasing to $1.9 billion in 2050. However, as discussed at length in the IWG's February 2021
TSD, estimates focusing on the climate impacts occurring solely within U.S. borders are an underestimate of the
benefits of GHG mitigation accruing to U.S. citizens and residents, as well as being subject to a considerable degree
of uncertainty due to the manner in which they are derived. In particular, the estimates developed under revoked
E.O. 13783 did not capture significant regional interactions, spillovers, and other effects and so are incomplete
underestimates. The U.S. District Court for the Northern District of California found that by omitting such impacts,
those "interim domestic" estimates "fail[ed] to consider.. .important aspect[s] of the problem" and departed from the
"best science available" as reflected in the global estimates. California v. Bernhardt, 472 F. Supp. 3d 573, 613-14
(N.D.Cal. 2020). EPA continues to center attention in this regulatory analysis on the global measures of the SC-
GHG as the appropriate estimates and as necessary for all countries to use to achieve an efficient allocation of
resources for emissions reduction on a global basis, and so benefit the U.S. and its citizens.
3-39
-------
Table 3-11: Estimated Global Climate Benefits from Changes in CO2 Emissions 2023 - 2050 for the Final
Rule (Billions of 2018$)
Discount Rate and Statistic
Calendar Year
5% Average
3% Average
2.5% Average
% 95th percentile
2023
$0,076
$0.26
$0.38
$0.77
2024
$0.16
$0.53
$0.78
$1.6
2025
$0.28
$0.94
$1.4
$2.8
2026
$0.45
$1.5
$2.2
$4.5
2027
$0.67
$2.2
$3.2
$6.6
2028
$0.92
$3
$4.3
$9
2029
$1.2
$3.7
$5.4
$11
2030
$1.4
$4.4
$6.4
$13
2031
$1.6
$5.2
$7.5
$16
2032
$1.9
$5.9
$8.5
$18
2033
$2.2
$6.6
$9.5
$20
2034
$2.4
$7.3
$11
$22
2035
$2.6
$8
$11
$24
2036
$2.9
$8.6
$12
$26
2037
$3.1
$9.2
$13
$28
2038
$3.3
$9.7
$14
$30
2039
$3.5
$10
$14
$31
2040
$3.7
$11
$15
$33
2041
$3.9
$11
$16
$34
2042
$4
$11
$16
$35
2043
$4.2
$12
$17
$36
2044
$4.3
$12
$17
$37
2045
$4.5
$12
$17
$38
2046
$4.6
$13
$18
$39
2047
$4.7
$13
$18
$40
2048
$4.9
$13
$18
$40
2049
$5
$13
$18
$41
2050
$5.1
$14
$19
$42
PV
$29
$120
$190
$370
Annualized
$1.9
$6.3
$9.1
$19
Notes:




Climate benefits are based on
changes (reductions) in C02 emissions and are calculated using four different estimates of the social cost of
carbon (SC-C02) (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-C02 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.



3-40
-------
Table 3-12: Estimated Global Climate Benefits from Changes in CH4 Emissions 2023 - 2050 for the Final
Rule (Billions of 2018$)
Discount Rate and Statistic
Calendar Year
5% Average
3% Average
2.5% Average
3% 95th percentile
2023
$0.0037
$0.0081
$0,011
$0,021
2024
$0.0076
$0,016
$0,021
$0,043
2025
$0,014
$0,029
$0,038
$0,077
2026
$0,022
$0,047
$0,061
$0.12
2027
$0,033
$0.07
$0.09
$0.19
2028
$0,045
$0,096
$0.12
$0.25
2029
$0,058
$0.12
$0.15
$0.32
2030
$0.07
$0.15
$0.19
$0.39
2031
$0,083
$0.17
$0.22
$0.46
2032
$0,097
$0.2
$0.25
$0.53
2033
$0.11
$0.22
$0.29
$0.6
2034
$0.12
$0.25
$0.32
$0.67
2035
$0.14
$0.28
$0.35
$0.74
2036
$0.15
$0.3
$0.38
$0.8
2037
$0.16
$0.32
$0.41
$0.86
2038
$0.17
$0.34
$0.43
$0.92
2039
$0.18
$0.36
$0.46
$0.97
2040
$0.2
$0.38
$0.48
$1
2041
$0.21
$0.4
$0.5
$1.1
2042
$0.22
$0.42
$0.52
$1.1
2043
$0.22
$0.43
$0.54
$1.2
2044
$0.23
$0.45
$0.56
$1.2
2045
$0.24
$0.46
$0.57
$1.2
2046
$0.25
$0.47
$0.59
$1.3
2047
$0.26
$0.49
$0.6
$1.3
2048
$0.27
$0.5
$0.62
$1.3
2049
$0.28
$0.52
$0.64
$1.4
2050
$0.29
$0.53
$0.65
$1.4
PV
$1.5
$4.4
$6
$12
Annualized
$0,099
$0.22
$0.29
$0.6
Climate benefits are based on changes (reductions) in CH4 emissions and are calculated using four different
estimates of the social cost of methane (SC-CH4) (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-CH4 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.


3-41
-------
Table 3-13: Estimated Global Climate Benefits from Changes in N2O Emissions 2023 - 2050 (Billions of
2018$)
Discount Rate and Statistic
Calendar Year
5% Average
3% Average
2.5% Average
3% 95th percentile
2023
$0.0009
$0.0028
$0.0041
$0.0073
2024
$0.0019
$0.0057
$0.0084
$0,015
2025
$0.0034
$0.01
$0,015
$0,027
2026
$0.0056
$0,017
$0,024
$0,044
2027
$0.0082
$0,024
$0,035
$0,065
2028
$0,011
$0,033
$0,048
$0,088
2029
$0,014
$0,042
$0.06
$0.11
2030
$0,017
$0.05
$0,072
$0.13
2031
$0.02
$0,059
$0,084
$0.16
2032
$0,023
$0,067
$0,096
$0.18
2033
$0,027
$0,076
$0.11
$0.2
2034
$0.03
$0,084
$0.12
$0.22
2035
$0,033
$0,092
$0.13
$0.24
2036
$0,036
$0.1
$0.14
$0.27
2037
$0,039
$0.11
$0.15
$0.28
2038
$0,042
$0.11
$0.16
$0.3
2039
$0,044
$0.12
$0.17
$0.32
2040
$0,047
$0.13
$0.18
$0.33
2041
$0,049
$0.13
$0.18
$0.35
2042
$0,051
$0.14
$0.19
$0.36
2043
$0,053
$0.14
$0.2
$0.37
2044
$0,055
$0.14
$0.2
$0.39
2045
$0,057
$0.15
$0.21
$0.4
2046
$0,059
$0.15
$0.21
$0.41
2047
$0,061
$0.16
$0.22
$0.42
2048
$0,063
$0.16
$0.22
$0.43
2049
$0,065
$0.16
$0.22
$0.44
2050
$0,066
$0.17
$0.23
$0.44
PV
$0.36
$1.4
$2.2
$3.8
Annualized
$0,024
$0,073
$0.11
$0.2
Climate benefits are based on changes (reductions) in N20 emissions and are calculated using four different
estimates of the social cost of nitrous oxide (SC-N20) (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-N20 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.


3-42
-------
Table 3-14: Estimated Global Climate Benefits from Changes in GHG Emissions 2023 - 2050 (Billions of
2018$)
Discount Rate and Statistic
Calendar Year
5% Average
3% Average
2.5% Average
3% 95th percentile
2023
$0,081
$0.27
$0.4
$0.8
2024
$0.17
$0.55
$0.81
$1.6
2025
$0.3
$0.97
$1.4
$2.9
2026
$0.48
$1.6
$2.3
$4.7
2027
$0.71
$2.3
$3.3
$6.9
2028
$0.97
$3.1
$4.5
$9.3
2029
$1.2
$3.9
$5.6
$12
2030
$1.5
$4.6
$6.7
$14
2031
$1.7
$5.4
$7.8
$16
2032
$2
$6.2
$8.9
$19
2033
$2.3
$6.9
$9.9
$21
2034
$2.6
$7.7
$11
$23
2035
$2.8
$8.4
$12
$25
2036
$3.1
$9
$13
$27
2037
$3.3
$9.6
$14
$29
2038
$3.5
$10
$14
$31
2039
$3.7
$11
$15
$33
2040
$3.9
$11
$16
$34
2041
$4.1
$12
$16
$36
2042
$4.3
$12
$17
$37
2043
$4.5
$12
$17
$38
2044
$4.6
$13
$18
$39
2045
$4.8
$13
$18
$40
2046
$4.9
$13
$18
$41
2047
$5.1
$14
$19
$41
2048
$5.2
$14
$19
$42
2049
$5.3
$14
$19
$43
2050
$5.5
$14
$20
$44
PV
$31
$130
$200
$390
Annualized
$2
$6.6
$9.5
$20
Climate benefits are based on changes (reductions) in CO2, CH4, and N20 emissions and are calculated using four
different estimates of the social cost of carbon (SC-C02), the social cost of methane (SC-CH4), and the social cost
of nitrous oxide (SC-N20) (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, SC-CH4, and SC-N20 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.


3-43
-------
Table 3-15: Estimated Global Climate Benefits from Changes in GHG Emissions 2023 - 2050 for the Proposal
Standards (Billions of 2018$)
Discount Rate and Statistic
Calendar Year
5% Average
3% Average
2.5% Average
3% 95th percentile
2023
$0,055
$0.18
$0.27
$0.55
2024
$0.12
$0.39
$0.57
$1.2
2025
$0.22
$0.72
$1.1
$2.1
2026
$0.34
$1.1
$1.6
$3.3
2027
$0.49
$1.6
$2.3
$4.8
2028
$0.66
$2.1
$3.1
$6.3
2029
$0.83
$2.6
$3.8
$7.8
2030
$0.99
$3.1
$4.5
$9.3
2031
$1.2
$3.6
$5.2
$11
2032
$1.3
$4.1
$5.9
$12
2033
$1.5
$4.6
$6.5
$14
2034
$1.7
$5
$7.2
$15
2035
$1.8
$5.5
$7.8
$17
2036
$2
$5.9
$8.4
$18
2037
$2.1
$6.3
OO
OO
&
$19
2038
$2.3
$6.6
$9.3
$20
2039
$2.4
$6.9
$9.7
$21
2040
$2.5
$7.2
$10
$22
2041
$2.6
$7.4
$10
$23
2042
$2.7
$7.7
$11
$23
2043
$2.9
$7.9
$11
$24
2044
$3
$8.1
$11
$25
2045
$3.1
$8.3
$12
$25
2046
$3.1
$8.5
$12
$26
2047
$3.2
$8.7
$12
$27
2048
$3.3
$8.9
$12
$27
2049
$3.4
$9
$12
$28
2050
$3.5
$9.2
$13
$28
PV
$20
$83
$130
$250
Annualized
$1.3
$4.3
$6.2
$13
Climate benefits are based on changes (reductions) in CO2, CH4, and N20 emissions and are calculated using four
different estimates of the social cost of carbon (SC-C02), the social cost of methane (SC-CH4), and the social cost
of nitrous oxide (SC-N20) (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, SC-CH4, and SC-N20 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.


3-44
-------
Table 3-16: Estimated Global Climate Benefits from Changes in GHG Emissions 2023 - 2050 for Alternative
2 minus 10 (Billions of 2018$)
Discount Rate and Statistic
Calendar Year
5% Average
3% Average
2.5% Average
3% 95th percentile
2023
$0.12
$0.41
$0.6
$1.2
2024
$0.23
$0.77
$1.1
$2.3
2025
$0.37
$1.2
$1.8
$3.6
2026
$0.56
$1.8
$2.7
$5.5
2027
$0.79
$2.6
$3.7
$7.6
2028
$1.1
$3.4
$4.9
$10
2029
$1.3
$4.1
$6
$12
2030
$1.6
$4.9
$7.1
$15
2031
$1.8
$5.7
$8.2
$17
2032
$2.1
$6.4
$9.2
$19
2033
$2.4
$7.2
$10
$22
2034
$2.6
$7.9
$11
$24
2035
$2.9
$8.6
$12
$26
2036
$3.1
$9.2
$13
$28
2037
$3.3
$9.8
$14
$30
2038
$3.5
$10
$15
$31
2039
$3.7
$11
$15
$33
2040
$3.9
$11
$16
$34
2041
$4.1
$12
$16
$36
2042
$4.3
$12
$17
$37
2043
$4.5
$12
$17
$38
2044
$4.6
$13
$18
$39
2045
$4.8
$13
$18
$40
2046
$4.9
$13
$18
$41
2047
$5.1
$14
$19
$41
2048
$5.2
$14
$19
$42
2049
$5.3
$14
$19
$43
2050
$5.5
$14
$20
$44
PV
$32
$130
$200
$400
Annualized
$2.1
$6.7
$9.7
$20
Climate benefits are based on changes (reductions) in CO2, CH4, and N20 emissions and are calculated using four
different estimates of the social cost of carbon (SC-C02), the social cost of methane (SC-CH4), and the social cost
of nitrous oxide (SC-N20) (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, SC-CH4, and SC-N20 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.


3-45
-------
3.4 Drive Surplus, Congestion and Noise
As discussed in Chapter 3.1, the assumed rebound effect might occur when an increase in
vehicle fuel efficiency encourages people to drive more as a result of the lower cost per mile of
driving. Along with the safety considerations associated with increased vehicle miles traveled
(described in Chapter 5.3), additional driving can lead to other costs and benefits that can be
monetized.
The increase in travel associated with the rebound effect produces additional benefits to
vehicle drivers, which reflect the value of the added (or more desirable) social and economic
opportunities that become accessible with additional travel. Consistent with assumptions used in
the NPRM, this analysis estimates the economic benefits from increased rebound-effect driving
as the owner/operator surplus from the additional accessibility it provides.
The equation for the calculation of the Drive Value:
Drive Value = (1/2) (VMTrebound) [(S/ot/Z^noAction -	Action]
The economic value of the increased owner/operator surplus provided by added driving is
estimated as one half of the product of the decline in vehicle operating costs per vehicle-mile and
the resulting increase in the annual number of miles driven. Because it depends on the extent of
improvement in fuel consumption, the value of benefits from increased vehicle use changes by
model year and varies among alternative standards.
In contrast to the benefits of additional driving are the costs associated with that driving. If net
operating costs of the vehicle decline, then we expect a positive rebound effect. Increased vehicle
use associated with a positive rebound effect also contributes to increased traffic congestion and
highway noise. Depending on how the additional travel is distributed throughout the day and
where it takes place, additional vehicle use can contribute to traffic congestion and delays by
increasing traffic volumes on facilities that are already heavily traveled during peak periods.
These added delays impose higher costs on other road users in the form of increased travel time
and operating expenses. Because drivers do not take these external costs into account in deciding
when and where to travel, we account for them separately as a cost of the added driving
associated with the rebound effect.
EPA relies on estimates of congestion and noise costs developed by the Federal Highway
Administration to estimate the increased external costs caused by added driving due to the
rebound effect.117 EPA employed estimates from this source previously in the analysis
accompanying the light-duty 2010 and 2012 vehicle rulemakings and the 2016 Draft TAR and
Proposed Determination. We continue to find them appropriate for this analysis after reviewing
the procedures used by FHWA to develop them and considering other available estimates of
these values.
FHWA's congestion cost estimates focus on freeways because non-freeway effects are less
serious due to lower traffic volumes and opportunities to re-route around the congestion. The
agencies, however, applied the congestion cost to the overall VMT. The results of this analysis
potentially overestimate the congestions costs associated with increased vehicle use, and thus
lead to a conservative estimate of net benefits.
3-46
-------
EPA has used FHWA's "Middle" estimates for marginal congestion and noise costs caused by
increased travel from vehicles. This approach is consistent with the methodology used in our
prior analyses. The values used are shown in Table 3-17.
These congestion costs are consistent with those used in the NPRM, the 2016 Draft TAR and
the Proposed Determination. For this final rule, EPA has chosen not to adopt the approach from
the SAFE FRM where scaling factors were used to adjust the underlying FHWA congestion cost
estimates. In particular, EPA concluded that scaling the marginal per-mile congestion costs by
the change in VMT per lane-mile on US highways from 1997 to 2017 does not account for
changes in average speeds and improved road design, and may have the potential to over-
estimate costs. We are using the FHWA congestion estimates without scaling, consistent with the
SAFE NPRM and prior EPA rulemakings, and adjusting to represent 2018 dollars.
Table 3-17: Costs Associated with Congestion and Noise (2018 dollars per vehicle mile)

Passenger cars
Van/SUVs
Pickups
Congestion
0.0634
0.0634
0.0566
Noise
0.0009
0.0009
0.0009
3-47
-------
References for Chapter 3
I	Winebrake, J., Green, E., Comer, B., Corbett, J., Froman, S. 2012. "Estimating the direct rebound effect for on-
road freight transportation," Energy Policy 48, 252-259.
Greene, D., Kahn, J., Gibson, R. 1999. "Fuel economy rebound effect for U.S. household vehicles," The Energy
Journal, 20.
3	Greening, L., Greene, D., Difiglio, C. 2000. "Energy efficiency and consumption - the rebound effect - a survey,"
Energy Policy, 28, 389-401.
4	Figure ES-1. Estimated Real World Fuel Economy and CO2. pp. ES-3. "The 2020 EPA Automotive Trends
Report. Greenhouse gas emissions, fuel economy and technology since 1975. Executive Summary." U.S. EPA 420-
5-21-001. January 2021.
5	Sorrell, S. and Dimitropoulos, J. 2007. "UKERC Review of Evidence of the Rebound Effect, Technical Report 2:
Econometric Studies," Sussex Energy Group, Working Paper.
6	Pickrell, D. and Schimek, P. 1999. "Growth in Motor Vehicle Ownership and Use: Evidence from the Nationwide
Personal Transportation Survey," Journal of Transportation and Statistics, vol. 2, no. 1, pp. 1-17.
H
Pickrell, D. and Schimek, P., 1999. "Growth in Motor Vehicle Ownership and Use: Evidence from the Nationwide
Personal Transportation Survey," Journal of Transportation and Statistics, vol. 2, no. 1, pp. 1-17.
8	Hymel, K., Small, K., and Van Dender, K. 2010. "Induced demand and rebound effects in road transport,"
Transportation Research Part B: Methodological, Volume 44, Issue 10, December, Pages 1220-1241, ISSN 0191-
2615, DOI: 10.1016/j.trb.2010.02.007. [EPA-HQ-OAR-2010-2010-0799-0758],
9	Barla, P., Lamonde, B., Miranda-Morano, L. and Boucher, N. 2009. Travel distance, stock and fuel efficiency of
private vehicles in Canada; price elasticities and rebound effect. Transportation, 36: 389-402.
10	Dargay, J and Gately, D. 1997. "The demand for transportation fuels: imperfect price-reversibility?"
Transportation Research Part B 31(1). Sentenac-Chemin, E. 2012, "Is the price effect on fuel consumption
symmetric: Some evidence from an empirical study," Energy Policy, Volume 41, pp. 59-65.
II	Gately, D. 1993. "The Imperfect Price-Reversibility of World Oil Demand," The Energy Journal, International
Association for Energy Economics, vol. 14(4), pp. 163-182.
12	Greene, D. 2012. "Rebound 2007: Analysis of U.S. light-duty vehicle travel statistics," Energy Policy, vol. 41, pp.
14-28.
13	Su, Q. 2012. A quantile regression analysis of the rebound effect: Evidence from the 2009 National Household
Transportation Survey in the United States, Energy Policy, 45, pp. 368-377.
14	Frondel, M. and Vance, C. 2013. Re-Identifying the Rebound: What about Asymmetry? The Energy Journal, 34
(4):43-54.
15	Lui, Y., Tremblay, J. and Cirillo, C. 2014. "An integrated model for discrete and continuous decisions with
application to vehicle ownership, type and usage," Transportation Research Part A, pp. 315-328.
16	Linn, J. 2016. "The Rebound Effect of Passenger Vehicles," The Energy Journal, Volume 37, Issue Number Two.
1 7
Gillingham, K., 2014. "Identifying the elasticity of driving: Evidence from a gasoline price shock in California".
Regional Science and Urban Economics, 47: pp.13-24.
1 R
Gillingham, K. 2020. The Rebound Effect and the Proposed Rollback of U.S. Fuel Economy Standards, Review
of Environmental Economics and Policy. University of Chicago Press Journal. Volume Number 14, Number 1,
Winter.
19	Gillingham, K. 2011. The Consumer Response to Gasoline Prices: Empirical Evidence and Policy Implications,
Stanford University Ph.D. Dissertation. Accessed at:
https://stacks.stanford.edU/file/druid:wz808zn3318/Gillingham_Dissertation-augmented.pdf
20	Wang, T. and Chen, C. 2014. "Impact of fuel price on vehicle miles traveled (VMT): do the poor respond in the
same way as the rich?" Transportation 41(1): 91-105.
3-48
-------
21	Hymel, K. and Small, K. 2015. "The rebound effect for automobile travel: Asymmetric response to price changes
and novel features of the 2000s," Energy Economics, 49 (2015) 93-103.
22	West, J., Hoekstra, M., Meer, J., Puller, S. 2017. "Vehicle Miles (Not) Traveled: Why Fuel Economy
Requirements Don't Increase Household Driving", Journal of Public Economics, Elsevier, Volume 145(c), pp. 65-
81.
23	Gillingham, K., Jenn, A. and Azevedo, I. 2015. "Heterogeneity in the response to gasoline prices: Evidence from
Pennsylvania and implications for the rebound effect," Energy Economics, Elsevier, vol. 52(S1), pages 41-52.
24
Leung, W. 2015, Three Essays in Energy Economics, University of California San Diego PhD Dissertation,
Accessed at: https://escholarship.org/content/qt3h51364m/qt3h51364m.pdf
25	Langer, A., Maheshri, V. and Winston, C. 2017. "From gallons to miles: A disaggregate analysis of automobile
travel and externality taxes". Journal of Public Economics, 152, pp. 34-46.
26	Knittel, C. and Sandler, R. 2018. The Welfare Impact of Second-Best Uniform-Pigouvian Taxation: Evidence
from Transportation, American Economic Journal: Economic Policy 2018, 10(4): 211-242
https://doi.org/10.1257/pol.20160508 211.
27	De Borger, B. Mulalic, I., and Rouwendal, J. 2016. Measuring the rebound effect with micro data: A first
difference approach, Journal of Environmental Economics and Management, 79, 1-17.
28	Gillingham, K., Rapson, D., Wagner, G. 2016. "The Rebound Effect and Energy Efficiency Policy," Review of
Environmental Economics and Policy, 10 (1), pp. 68-88.
29
Wenzel, T. andFujita, K. 2018. Elasticity of Vehicle Miles of Travel to Changes in the Price of Gasoline and the
Cost of Driving in Texas, Lawrence Berkeley National Laboratory Report LBNL-2001138. Accessed at:
https://eln.lbl.gov/publications/elasticity-vehiclemiles-travel
Gillingham, K. and Munk-Nielsen, A. 2019. A tale of two tails: Commuting and the fuel price response in
driving, Journal of Public Economics, 109, pp. 27-41.
31	Gillingham, K. 2020. The Rebound Effect and the Proposed Rollback of U.S. Fuel Economy Standards, Review
of Environmental Economics and Policy, University of Chicago Press Journal, Volume Number 14, Number 1,
Winter.
32	U.S. Department of Transportation, National Highway Traffic Safety Administration, U.S. Environmental
Protection Agency, Preliminary Regulatory Impact Analysis, The Safer Affordable Fuel-Efficient (SAFE) Vehicles
Rule for Model Year 2021-2026 Passenger Cars and Light Trucks, 2018. July. Updated August 2018; October,
2018.
33	The Safer Affordable Fuel-Efficient (SAFE) Vehicles Rule for Model Years 2021-2026 Passenger Cars and Light
Trucks". 83 Fed. Reg. 42,986 Aug. 24, 2018.
34	Dimitropoulos, A., Questali, W. and Sintek, C. 2018. The rebound effect in road transport: A meta-analysis of
empirical studies, Energy Economics, Volume 75, September, pp. 163 - 179.
35	See pp. 11-351-11-352. 2021. Assessment of Technologies for Improving Light-Duty Vehicle Fuel
Economy-2025-2035, Committee on Assessment of Technologies for Improving Fuel Economy of Light-Duty
Vehicles-Phase 3, Board on Energy and Environmental Systems Division on Engineering and Physical Sciences. A
Consensus Study Report of the National Academies of Sciences, Engineering and Medicine.
36	U.S. EPA Science Advisory Board. Science Advisory Board (SAB) Consideration of the Scientific and Technical
Basis of the EPA's Proposed Rule titled The Safer Affordable Fuel-Efficient (SAFE) Vehicles Rule for Model Years
2021-2026 Passenger Cars and Light Trucks. Document Number EPA-SAB-20-003, February 27, 2020.
37	U.S. Energy Information Administration (EIA), U.S. Department of Energy (DOE), Annual Energy Outlook,
2021, https://www.eia.gov/outlooks/aeo/pdf/00%20AE02021%20Chart%20Library.pdf, see slide entitled, Gross
Domestic Product and population assumptions.
38	Greene, D. 2010. Measuring energy security: Can the United States achieve oil independence? Energy Policy 38,
pp. 1614-1621.
39	Ibid.
3-49
-------
40	Ibid.
41	Cyberattack Forces a Shutdown of a Top U.S. Pipeline. New York Times. May 8th, 2021.
42
U.S. Energy Information Administration. 2021. Total Energy. Annual Energy Review. Table 3.1. Petroleum
Overview. April.
43	Ibid.
44	U.S. Energy Information Administration. 2021. Annual Energy Outlook 2021. Reference Case. Table A11.
Petroleum and Other Liquids Supply and Disposition.
45	See EIA https://www.eia.gov/energyexplained/oil-and-petroleum-products/imports-and-exports.php
46	Bohi, D. and Montgomery, D. 1982. Social Cost of Imported and U.S. Import Policy, Annual Review of Energy,
7:37-60. Energy Modeling Forum, 1981. World Oil, EMF Report 6, Stanford University Press: Stanford 39 CA.
https//emf.stanford.edu/publications/emf-6-world-oil.
47
Plummer, J. (Ed.). 1982. Energy Vulnerability, "Basic Concepts, Assumptions and Numerical Results", pp. 13 -
36, Cambridge MA: Ballinger Publishing Co.
40
Bohi, D. and Montgomery, D. 1982. Social Cost of Imported and U.S. Import Policy, Annual Review of Energy,
7:37-60.
49	Broadman, H. 1986. "The Social Cost of Imported Oil," Energy Policy 14(3):242-252. Broadman H. and Hogan,
W. 1988. "Is an Oil Import Tariff Justified? An American Debate: The Numbers Say 'Yes'." The Energy Journal 9:
7-29.
50	Hogan, W. 1981. "Import Management and Oil Emergencies", Chapter 9 in Deese, 5 David and Joseph Nye, eds.
Energy and Security. Cambridge, MA: Ballinger Publishing Co.
51	Leiby, P., Jones, D., Curlee, R. and Lee, R. 1997. Oil Imports: An Assessment of Benefits and Costs, ORNL-
6851, Oak Ridge National Laboratory, November.
52	Parry, I. and Darmstadter, J. 2004. "The Costs of U.S. Oil Dependency," Resources for the Future, November 17,
2004. Also published as NCEP Technical Appendix Chapter 1: Enhancing Oil Security, the National Commission
on Energy Policy 2004 Ending the Energy Stalemate - A Bipartisan Strategy to Meet America's Energy Challenges.
53	National Research Council. 2009. Hidden Costs of Energy: Unpriced Consequences of Energy Production and
Use. National Academy of Science, Washington, DC.
54	Nordhaus, W. 2007. "Who's Afraid of a Big Bad Oil Shock?," Brookings Papers on Economic Activity,
Economic Studies Program, The Brookings Institution, vol. 38(2), pp. 219-240.
55	Rasmussen, T. and Roitman, A. 2011. Oil Shocks in a Global Perspective: Are We Really That Bad. IMF
Working Paper Series.
56	Hamilton, J. 2012. Oil Prices, Exhaustible Resources, and Economic Growth. In Handbook of Energy and Climate
Change. Retrieved from http://econweb.ucsd.edu/~jhamilto/handbook_climate.pdf
Kim, D. 2012. What is an oil shock? Panel data evidence. Empirical Economics, Volume 43, pp. 121-143.
CO
Engemann, K., Kliesen. K. and Owyang, M. 2011. Do Oil Shocks Drive Business Cycles, Some U.S. and
International Evidence. Federal Reserve Bank of St. Louis, Working Paper Series. No. 2010-007D.
59	Ramey, V. and Vine, D. 2010. "Oil, Automobiles, and the U.S. Economy: How Much have Things Really
Changed?", National Bureau of Economic Research Working Papers, WP 16067 (June). Retrieved from
http://www.nber.org/papers/wl6067.pdf
60	Baumeister C., Peersman, G. and VanRobays, I. 2010. "The Economic Consequences of Oil Shocks: Differences
across Countries and Time," RBA Annual Conference Volume in: Renee Fry & Callum Jones & Christopher Kent
(ed.), Inflation in an Era of Relative Price Shocks, Reserve Bank of Australia.
61	Kilian, L. and Vigfusson, R. 2014. "The role of oil price shocks in causing U.S. recessions," CFS Working Paper
Series 460, Center for Financial Studies.
62	Kilian, L. 2009. "Not All Oil Price Shocks Are Alike: Disentangling Demand and Supply Shocks in the Crude Oil
Market." American Economic Review, 99 (3): pp. 1053-69.
3-50
-------
63	Cashin, P., Mohaddes, K., and Raissi, M. 2014. The Differential Effects of Oil Demand and Supply Shocks on the
Global Economy, Energy Economics, 12 (253), D01:10.1016/j.eneco.2014.03.014
64	Krupnick, A., Morgenstern, R., Balke, N., Brown, S., Herrara, M. and Mohan, S. 2017. "Oil Supply Shocks, U.S.
Gross Domestic Product, and the Oil Security Problem", Resources for the Future Report.
65	Brown, S. 2018. New estimates of the security costs of U.S. oil consumption", Energy Policy, 113 pp. 171-192.
66	Balke, N. and Brown, S. 2018. "Oil Supply Shocks and the U.S. Economy: An Estimated DSGE Model". Energy
Policy, 116, pp. 357-372.
67	Kilian, L. 2009. Not All Oil Price Shocks are Alike: Disentangling Demand and Supply Shocks in the Crude Oil
Market, American Economic Review, 99:3, pp., 1053-1069.; and
68	Kilian, L. and Murphy, D. 2013. "The Role of Inventories and Speculative Trading in the Global Market for
Crude Oil", Journal of Applied Economics, https://doi.org/10.1002/jae.2322.
69	Baumeister, C. and Hamilton, J. 2019. "Structural Interpretation of Vector Autoregressions with Incomplete
Identification: Revisiting the Role of Oil Supply and Demand Shocks", American Economic Review, 109(5),
pp.1873-1910.
70
Mohan, S. 2017. "Oil Price Shocks and the U.S. Economy: An Application of the National Energy Modeling
System". Resources for the Future Report Appendix.
71	U.S. Energy Information Administration. September 23, 2019. "Saudi Arabia crude oil production outage affects
global crude oil and gasoline prices". Today in Energy.
72	Ibid.
73	Ibid.
74	Uria-Martinez, R., Leiby, P., Oladosu, G., Bowman, D., Johnson, M. 2018. Using Meta-Analysis to Estimate
World Oil Demand Elasticity, ORNL Working Paper.
Oladosu, G., Leiby, P., Bowman, D., Uria-Martinez, R., Johnson, M. 2018. Impacts of oil price shocks on the
U.S. economy: a meta-analysis of oil price elasticity of GDP for net oil-importing economies, Energy Policy 115.
pp. 523-544.
76	Union of Concerned Scientist, "What is Tight Oil?". 2015. "Tight oil is a type of oil found in impermeable shale
and limestone rock deposits. Also known as "shale oil," tight oil is processed into gasoline, diesel, and jet fuels—
just like conventional oil—but is extracted using hydraulic fracturing, or 'Tracking."
77	Newell, R. and Prest, B. 2019. The Unconventional Oil Supply Boom: Aggregate Price Response from
Microdata, The Energy Journal, Volume 40, Issue Number 3.
70
Bjornland, H., Nordvik, F. andRohrer, M. 2021. "Supply flexibility in the shale patch: Evidence from North
Dakota," Journal of Applied Economics, February, https://doi.org/10.1002/jae.2808.
79	Crane, K., Goldthau, A., Toman, M., Light, T., Johnson, S., Nader, A., Rabasa, A. and Dogo, H. 2009. Imported
oil and US national security. RAND, 2009. http://www.stormingmedia.us/62/6279/A627994.pdf.
80	Koplow, D. and Martin, A. 1998. Fueling Global Warming: Federal Subsidies to Oil in the United States.
Greenpeace, Washington, DC.
o 1
Moore, J., Behrens, C. and Blodgett, J. 1997. "Oil Imports: An Overview and Update of Economic and Security
Effects." CRS Environment and Natural Resources Policy Division report 98, no. 1: pp. 1-14.
82 Stern, R. 2010. "United States cost of military force projection in the Persian Gulf, 1976-2007." Energy Policy
38, no. 6. June: 2816-2825. doi:10.1016/j.enpol.2010.01.013.
http ://linkinghub. elsevier. com/retrieve/pii/S0301421510000194.
0-5
Delucchi, M. and Murphy, J. 2008. "US military expenditures to protect the use of Persian Gulf oil for motor
vehicles." Energy Policy 36, no. 6. June: 2253-2264. doi:10.1016/j.enpol.2008.03.006,
OA
Securing America's Future Energy. 2018. Issue Brief. The Military Cost of Defending the Global Oil Supply.
85 Ibid.
3-51
-------
86 Crane, K., Goldthau, A., Toman, M., Light, T., Johnson, S., Nader, A., Rabasa, A. and Dogo, H. 2009. Imported
oil and US national security. 2009. RAND. http://www.stormingmedia.us/62/6279/A627994.pdf.
U.S. Energy Information Administration. 2021. Annual Energy Outlook 2021. Reference Case. Table All.
Petroleum and Other Liquids Supply and Disposition.
88	U.S. Energy Information Administration. 2021. Annual Energy Outlook 2021. Reference Case. Table A11.
Petroleum and Other Liquids Supply and Disposition.
89	Leiby, Paul N. 2008. Estimating the Energy Security Benefits of Reduced U.S. Oil Imports, Final Report,
ORNL/TM-2007/028, Oak Ridge National Laboratory, Rev. March 14.
90	Leiby, P., Jones, D., Curlee, R., and Lee, R. 1997. Oil Imports: An Assessment of Benefits and Costs, ORNL-
6851, Oak Ridge National Laboratory, November.
91	Joint Technical Support Document, Final Rulemaking to Establish Light-Duty Vehicle Greenhouse Gas Emission
Standards and Corporate Average Fuel Economy Standards, EPA-420-R-10-901, April 2010. Final Rulemaking for
2017-2025 Light-Duty Vehicle Greenhouse Gas Emission Standards, EPA-420-R-12, August 2012.
92
Final Rulemaking to Establish Greenhouse Gas Emissions Standards and Fuel Efficiency Standards for Medium-
and Heavy-Duty Engines and Vehicles Regulatory Impact Analysis, U.S. EPA/NHTSA, EPA-420-R-11-901,
August 2011.
Greenhouse Gas Emissions and Fuel Efficiency Standards for Medium- and Heavy-Duty Engines and Vehicles -
Phase 2 Regulatory Impact Analysis, U.S. EPA/NHTSA, EPA-420-R-16-900 August 2016.
94	U.S EPA, U.S. DOT/NHTSA, California Air Resources Board. Draft Technical Assessment Report: Midterm
Evaluation of Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average Fuel Economy
Standards for Model Years 2022-2025. EPA-420-D-16-900. July 2016.
95	Regulation of Fuels and Fuel Additives: Changes to Renewable Fuel Standard Program; Proposed Rule, 40
C.F.R. §80 (2010).
96	IEA, Data and Statistics, https://www.iea.org/data-and-
statistics?country=WORLD&fuel=Oil&indicator=OilProductsConsBy Sector; Accessed March 2021.
97
Regulation of Fuels and Fuel Additives: Changes to Renewable Fuel Standard Program; Proposed Rule, 40 C.F.R.
§80. 2010.
98	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: 
-------
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).
107
Intergovernmental Panel on Climate Change (IPCC). 2007. Core Writing Team; Pachauri, R.K; and
Reisinger, A. (ed.), Climate Change 2007: Synthesis Report, Contribution of Working Groups I, II and III to
the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, IPCC, ISBN 92-9169-
122-4.
108	Intergovernmental Panel on Climate Change (IPCC). 2014. Climate Change 2014: Synthesis Report.
Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental
Panel on Climate Change [Core Writing Team, R.K. Pachauri and L.A. Meyer (eds.)]. IPCC, Geneva,
Switzerland, 151 pp.
109	Intergovernmental Panel on Climate Change (IPCC). 2018. Global Warming of 1.5°C. An IPCC Special
Report on the impacts of global warming of 1.5°C above pre-industrial levels and related global
greenhouse gas emission pathways, in the context of strengthening the global response to the threat of
climate change, sustainable development, and efforts to eradicate poverty [Masson-Delmotte, V., P.
Zhai, H.-O. Portner, D. Roberts, J. Skea, P.R. Shukla, A. Pirani, W. Moufouma-Okia, C. Pean, R. Pidcock, S.
Connors, J.B.R. Matthews, Y. Chen, X. Zhou, M.I. Gomis, E. Lonnoy, T. Maycock, M. Tignor, and T.
Waterfield (eds.)].
110	Intergovernmental Panel on Climate Change (IPCC). 2019a. Climate Change and Land: an IPCC special
report on climate change, desertification, land degradation, sustainable land management, food
security, and greenhouse gas fluxes in terrestrial ecosystems [P.R. Shukla, J. Skea, E. Calvo Buendia, V.
Masson-Delmotte, H.-O. Portner, D. C. Roberts, P. Zhai, R. Slade, S. Connors, R. vanDiemen, M. Ferrat, E.
Haughey, S. Luz, S. Neogi, M. Pathak, J. Petzold, J. Portugal Pereira, P. Vyas, E. Huntley, K. Kissick, M.
Belkacemi, J. Malley, (eds.)].
111	Intergovernmental Panel on Climate Change (IPCC). 2019b. IPCC Special Report on the Ocean and
Cryosphere in a Changing Climate [H.-O. Portner, D.C. Roberts, V. Masson-Delmotte, P. Zhai, M. Tignor,
E. Poloczanska, K. Mintenbeck, A. Alegria, M. Nicolai, A. Okem, J. Petzold, B. Rama, N.M. Weyer (eds.)].
112	U.S. Global Change Research Program (USGCRP). 2016. The Impacts of Climate Change on Human Health
in the United States: A Scientific Assessment. Crimmins, A., J. Balbus, J.L. Gamble, C.B. Beard, J.E. Bell, D.
Dodgen, R.J. Eisen, N. Fann, M.D. Hawkins, S.C. Herring, L. Jantarasami, D.M. Mills, S. Saha, M.C. Sarofim,
J. Trtanj, andL. Ziska, Eds. U.S. Global Change Research Program, Washington, DC, 312 pp.
https://dx.dio.org/10.7930/J0R49NQX.
113	U.S. Global Change Research Program (USGCRP). 2018. Impacts, Risks, and Adaptation in the United
States: Fourth National Climate Assessment, Volume II [Reidmiller, D.R., C.W. Avery, D.R. Easterling, K.E.
Kunkel, K.L.M. Lewis, T.K. Maycock, and B.C. Stewart (eds.)]. U.S. Global Change Research Program,
Washington, DC, USA, 1515 pp. doi: 10.7930/NCA4.2018.
114	National Academies of Sciences, Engineering, and Medicine (National Academies). 2016b. Attribution of
Extreme Weather Events in the Context of Climate Change. Washington, DC: The National Academies
Press, https://dio.org/10.17226/21852.
115	National Academies of Sciences, Engineering, and Medicine (National Academies). 2019. Climate Change
and Ecosystems. Washington, DC: The National Academies Press, https://doi.org/10.17226/25504.
116	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.
117	Federal Highway Administration, 1997 Federal Highway Cost Allocation Study,
http://www.fhwa.dot.gov/policy/hcas/final/index.htm, Tables V-22, V-23, and V-24 (last accessed Sept. 9, 2011).
3-53
-------
Chapter 4: Modeling GHG Compliance
4.1 Compliance Modeling, Analytical Updates, and Analytical Revisions
The modeling runs presented within this RIA are not meant to be the sole technical
justification underlying the revisions to the 2023-2026 GHG standards. That justification is also
based upon nearly a decade of analyses presented by EPA in the 2010 and 2012 final rules, the
2016 Draft TAR, and during both the Proposed and Final Determinations.1'2'3'4'5 Please see
Chapter 1.2 for further discussion of these previous EPA analyses. EPA's extensive public
record has made clear that more stringent GHG standards are both feasible at reasonable costs
and result in significant GHG emission reductions and public health and welfare benefits. The
analysis presented here is meant to show that, once again, when assessing standards of similar
stringency to those set forth in the 2012 rule, the results are similar to those presented within
previous EPA analyses. Those previous analytical results are summarized and discussed in
Chapter 1 of this RIA.
To estimate compliance costs and the associated technology pathways that manufacturers
might choose to comply with GHG standards, EPA has traditionally used its Optimization Model
for reducing Emissions of Greenhouse Gases (OMEGA). However, in considering modeling
tools to support the analysis for the proposed and final rulemaking, EPA has chosen to use the
CAFE Compliance and Effects Modeling System (CCEMS) for modeling light-duty GHG
compliance and costs for the revised MYs 2023-2026 GHG standards. As in the NPRM, for the
final rule EPA has also chosen to use the same version of that model used in support of the
SAFE FRM, with updates to inputs as described here. EPA made this choice for the following
reasons:
•	CCEMS has categorizations of technologies and model output formats that are distinct
to the model, so continuing use of CCEMS for this rule has facilitated comparisons to
the SAFE FRM.
•	By using the same modeling tool as used in the SAFE rule, we can more clearly
illustrate the influence of some of the key updates to the inputs used in the SAFE
FRM.
•	EPA considers the SAFE FRM version of the CCEMS model to be an effective
modeling tool for purposes of assessing standards through the MY 2026 timeframe,
along with changes to some of the key inputs as discussed below (see Table 4-2).
To be clear, modeling inputs are critically important to EPA analyses. As long as the
underlying structure of a modeling tool is sound, which is the case with both CCEMS and with
the OMEGA model, then it is not so much the specific tool used by EPA that is of paramount
importance but the inputs for the tools that are of the most importance within our GHG
compliance modeling efforts. This was made clear within the preamble to the SAFE FRM which
stated, "inputs do not define models; models use inputs. Therefore, disagreements about inputs
do not logically extend to disagreements about models. Similarly, while models determine
resulting outputs, they do so based on inputs."6 This statement was a response to public
comments received on the SAFE NPRM, some of which argued that EPA should use its own
modeling tools to support EPA actions. During development of the SAFE FRM, EPA staff had
significant input on the CCEMS and considered the FRM version of the model, given changes
made in response to public comments and EPA staff, to be a suitable modeling tool for that
4-1
-------
analysis. Similarly, we believe the SAFE FRM model and inputs, together with the key changes
we've made since the SAFE FRM and in response to comments and analysis since our August
2021 NPRM, are appropriate for the particular analysis at hand in assessing standards through
2026.
EPA is also currently developing an updated version of OMEGA. In the development of this
updated model, which we refer to as OMEGA2, we are placing emphasis on the treatment of
BEVs, the interaction between consumer and producer decisions including decisions between
cars and light trucks, and the capability to consider a wider range of GHG program options. The
OMEGA2 model is currently under peer review and we expect to make the results public early
next year.
As previously noted, we are using the version of the CCEMS docketed by NHTSA in support
of the SAFE FRM. CCEMS itself has been extensively documented by NHTSA in support of the
SAFE FRM and the documentation used there is applicable to the analysis presented here.7 Table
4-1 shows changes that were made to CCEMS for the NPRM analysis. In addition, the following
changes have been made to the inputs for this analysis for the final rule (see Table 4-2).
For this FRM analysis, EPA is using the MY 2020 base year fleet developed by NHTSA for
their recent NPRM and allowing the model to determine the future fleet based on the consumer
choice model and scrappage models.7'a As such, we have not changed the data contained within
the market file (the base year fleet) from what was used in NHTSA's recent NPRM other than as
described in Table 4-2 and to split the market file into separate framework-OEM and non-
framework-OEM fleets for some model runs to account for the impacts of the California
Framework Agreement.8 Note that the scrappage model received many negative comments
following the SAFE NPRM, but the FRM version of the model incorporated changes such that it
no longer generates the sales and VMT results of the NPRM version which was described by
commenters during the SAFE rule as being inconsistent with economic theory.9 The changes
incorporated are also consistent with recommendations of the EPA Science Advisory Board.10
As mentioned, for some model runs, including the No Action case, we have split the fleet in
two, one fleet consisting of California Framework manufacturers and the other consisting of the
non-Framework manufacturers. This was necessary since, for years that we are modeling
previous to the MY 2023 start of this program, we modeled the Framework manufacturers
meeting the more stringent Framework emission targets (as set in the scenarios file) while having
access to the additional advanced technology incentive multipliers of the Framework. We
modeled the Non-Framework-OEMs meeting less stringent (SAFE) standards while having
access to no advanced technology multipliers. For such model runs, a post-processing step was
necessary to properly sales-weight the two sets of model outputs into a single fleet of results.
This post-processing tool is in the docket for this rule.11
a See Chapter 8.1 for discussion of modeling of vehicle sales, as well as references to reviews of the literature that
EPA has conducted.
4-2
-------
Table 4-1: Changes made to SAFE FRM CCEMS Inputs for NPRM Model Runs
Input file
Changes
Parameters
file
Global social cost of GHG $/ton values used in place of domestic values (see Chapter 3.3).
Inclusion of global social cost of methane (CH4) and nitrous oxide (N20) $/ton values (see Chapter
3.3).
Updated PM2 5 cost factors (benefit per ton values, see Chapter 7)
Rebound effect of -0.10 rather than -0.20 (see Chapter 3.1).
AEO2021 fuel prices (expressed in 2018 dollars) rather than AEO2019.
Update energy security cost per gallon factors (see Chapter 3.2).
Congestion cost factors of 6.34/6.34/5.66 (car/van-SUV/truck) cents/mile rather than
15.4/15/4/13.75 (see Chapter 3.4).
Discounting values to calendar year 2021 rather than calendar year 2019.
The following fuel import and refining inputs have been changed based on AEO2021 (see Chapter
3.2):
Share of fuel savings leading to lower fuel imports:
Gasoline 7%; E85 19%; Diesel 7% rather than 50%; 7.5%; 50%
Share of fuel savings leading to reduced domestic fuel refining:
Gasoline 93%; E85 25.1%; Diesel 93% rather than 50%; 7.5%; 50%
Share of reduced domestic refining from domestic crude:
Gasoline 9%; E85 2.4%; Diesel 9% rather than 10%; 1.5%; 10%
Share of reduced domestic refining from imported crude:
Gasoline 91%; E85 24.6%; Diesel 91% rather than 90%; 13.5%; 90%


Technology
file
High Compression Ratio level 2 (HCR2, sometimes referred to as Atkinson level 2) technology
allowance set to TRUE for all engines beginning in 2018 (see Chapter 2).
Market file
On the Engines sheet, we allow HCR1 and HCR2 technology on all 6-cylinder and smaller engines
rather than allowing it on no engines (see Chapter 2).
Change the off-cycle credit values on the Credits and Adjustments sheet to 15 grams/mile for 2020
through 2026 (for the C ARB -OEM framework) or to 15 gram/mile for 2023 through 2026 (for the
proposed option) depending on the model run.
4-3
-------
Table 4-2: Changes made to EPA NPRM CCEMS Inputs for Final Rule Model Runs
Input file
Changes *
Parameters
file
Updated Gross Domestic Product, Number of Households, VMT growth rates and Historic Fleet
data consistent with updated projections from EI A (insert AEO version).
Updated energy security cost per gallon factors (see preamble Section VII.F).
Updated benefit per ton values and unique values for refinery and electricity generating unit
benefits (see preamble Section V).
Updated tailpipe and upstream emission factors consistent with (insert updated model runs)
Technology
file
High compression ratio level 2 (HCR2, sometimes referred to as Atkinson cycle) technology
allowance set to FALSE thereby making this technology unavailable.
BEV200 phase-in start year set to the same year as the new market file fleet (see below) which,
given the low year-over-year phase-in cap allows for low penetration of BEV200 technology in
favor of BEV300 technology.
Battery cost was reduced by about 25 percent (see preamble Section III. A); battery cost learning
is also held constant (i.e., no further learning) beyond the 2029 model year (see RIA 2.3.4 and
4.1.3).
Market file
The market file has been completely updated to reflect the MY 2020 fleet rather than the MY
2017 fleet used in the SAFE FRM and the EPA proposed rule. This was done by making use of
the market file developed by NHTSA in support of their recent CAFE NPRM (cite). Because the
market files are slightly different between the version of CCEMS we are using and the version
used by NHTSA, the files are not identical. However, the data are the same with the following
exceptions:
-	We have conducted all model runs using EPA Multiplier Mode 2 rather than Mode 1 as used in
the SAFE FRM and our NPRM.
-	We have used projected off-cycle credits as developed by NHTSA in support of their recent
CAFE NRPM rather than modeling all manufacturers as making use of the maximum allowable
off-cycle credits (see RIA 4.1.1.1).
-	We have updated the credit banks to incorporate more up-to-date information from manufacturer
certification and compliance data.
Scenarios
file
The off-cycle credit cap has been set to 10 g/mi even in scenarios and years for which 15 g/mi are
available. In addition, the off-cycle credit cost is set to $0 and is then post-processed back into the
costs calculated within CCEMS itself. See RIA 4.1.1.1 for more detail on why this was done and
the cost per credit that we are using in this final rule.
Runtime
settings
At runtime (in the CCEMS graphical user interface), the "Price Elasticity Multiplier" is now set to
-0.40 rather than the value of -1.00 used in the NPRM analysis.
Notes:
*As noted, we are now using a MY 2020 baseline fleet rather than a MY 2017 baseline fleet. However, since
some date-based data used by the model is hardcoded in the model code, and because we did not want to change
the model code for consistency with the NPRM, we have had to adjust any date-related input data accordingly.
Therefore, the input files we are using have in them headings and date-related identifiers reflecting a MY 2017-
based analysis but the data in the files have been adjusted by 3 years to reflect the fact that anything noted as
2017 is actually 2020. This is most easily understood with respect to the Scenarios input file which specifies the
standards in a year-by-year format. Due to this need to "shift years", the standards for MY 2023 through MY
2026 are actually entered in the columns noted as 2020 through 2023. Importantly, in post-processing of model
results, the "year-shift" is corrected back to reflect the actual years.
Our primary model runs consist of a "no action" case and an "action" case. The results, or
impact of our proposed standards, are measured relative to the no action case. Our no action case
consists of the Framework manufacturers (roughly 28 percent of fleet sales) meeting the
framework while NonFW-OEMs (roughly 72 percent of fleet sales) meet the SAFE FRM
standards. Our action case consists of the whole fleet meeting our standards for model years
2023 and later. Throughout this discussion, our no action case refers to this Framework/Non-
Framework manufacturer compliance split. We provide more detail behind some of the changes
made to the model inputs since our proposed rule in the following section (4.1.1).
4-4
-------
Importantly, CCEMS includes refresh and redesign schedules for all vehicles included in the
MY 2020 base fleet. These schedules are designed to capture the real-world constraints
associated with technology adoption and lead time. Some technologies can be added, or
improved, during a refresh event, while others require a redesign event. All of the electrification
technologies, including start-stop, hybridization and any form of plug-in electrification, require a
redesign event. When these refresh and redesign events are projected to occur impacts the ability
of manufacturers to comply with new standards and also highlights the importance of
flexibilities, especially averaging, banking and trading of credits, which allow manufacturers to
implement technologies on regular refresh/design schedules to minimize costs. The refresh and
redesign schedules we are using are consistent with those used by NHTSA in their recent CAFE
NPRM.12 The MY 2020 base fleet sales and shares of vehicles available for refresh and redesign
are shown in Table 4-3.
Table 4-3 Vehicle Sales Available for Refresh and Redesign
Model
Refresh

Redesign
Year
Sales
Share
Sales
Share
2021
1,783,285
13%
1,642,353
12%
2022
3,818,816
28%
2,572,865
19%
2023
2,288,923
17%
1,308,090
10%
2024
2,396,924
18%
3,422,953
25%
2025
2,084,892
15%
1,643,477
12%
2026
1,291,583
10%
2,379,932
18%
Lastly, to calculate the full program costs, benefits and net benefits, EPA has developed and
made use of an aforementioned post-processing tool.13 For many benefit-cost metrics, the post-
processing tool follows the calculation approach employed within the CCEMS model. For
example, costs associated with application of technology, foregone consumer sales surplus,
congestion, noise, fatalities and non-fatal crashes are all handled within the CCEMS model and
are taken "as-is" in the post-processing tool and transferred through to the final cost-benefit
analysis. However, the calculation of emissions benefits is handled entirely within the post-
processing tool by applying EPA's preferred $/ton benefit values (for both criteria air pollutants
(CAP) and GHGs) and discounting those values exclusively at their internally consistent
discount rates. In other words, the social cost of GHG $/ton values are generated using discount
rates equal to 2.5 percent, 3 percent and 5 percent. Each of those streams of benefit values will
always be discounted, whenever discounting is employed (for net present and/or annualized
valuations) using the internally consistent discount rate. CCEMS uses this same approach.
However, CCEMS can calculate only a single GHG valuation in each run of the model. As such,
to monetize 4 GHG streams (2.5 percent, 3 percent, 5 percent, 3 percent-95th percentile) would
require 4 separate runs of the model despite the fact that the tons do not differ between runs.
Therefore, EPA has chosen to post-process the results such that all 4 streams could be monetized
without re-running the full CCEMS. The post-processing tool also allows for valuation of
upstream CAP benefits separately from tailpipe CAP benefits which the SAFE FRM version of
CCEMS does not allow.
4-5
-------
4.1.1 Changes made to the Model Inputs since the Proposed Rule
4.1.1.1 Off-Cycle Credit Cost and changes since the Proposed Rule
We have updated the cost of off-cycle credits compared to the analysis for the NPRM. In the
NPRM, we used the off-cycle credit costs developed by NHTSA in support of the SAFE FRM.
Those costs are shown in Table 4-4.
Table 4-4: Cost per Off-Cycle Credit used in the NPRM (2018 dollars)

2020
2021
2022
2023
2024
2025
2026
$/gram/mile
83.79
82.21
81.16
79.58
78.52
77.47
76.31
CCEMS applies these costs on a $/gram/mile basis based on the credits entered in the market
input file. As a result, a MY 2026 vehicle adding 15 grams of off-cycle credits (as would have
been the case in our model runs for the proposed rule) would be adding $1,145 dollars in off-
cycle technology (15 x 76.31). In the no action case, most vehicles would have added just 10
grams/mile of off-cycle credit technology for a cost of $763. Therefore, the incremental costs
between the no action and action case vehicles would have automatically been $382 ($1,145 -
$763). Both the cost of these credits and the automatic application of those credits with different
levels of credits in different scenarios was considered inappropriate.
To address this, we ran CCEMS with different levels of off-cycle credit applied (again,
automatically) but with zero cost. For the curve coefficients used, we got the results shown in
Table 4-5.
Table 4-5: Cost per Vehicle relative to No Action at different levels of Zero-Cost Off-cycle Credit (2018
dollars)
Off-cycle credits
2021
2022
2023
2024
2025
2026
2027
0 g/mi
$90
$260
$500
$750
$820
$1,030
$1,110
5 g/mi
$30
$90
$310
$560
$710
$890
$970
10/g/mi
-$20
-$30
$190
$390
$540
$690
$760
15 g/mi
-$20
-$40
$180
$390
$540
$700
$760
As expected, the costs per vehicle are lower with increasing levels of off-cycle credit. This
helps illustrate the value of off-cycle credits relative to other available technologies. We can look
at these data another way, by looking at the incremental costs for the different ranges of off-cycle
credits, as shown in Table 4-6.
Table 4-6: Incremental Off-cycle Credit Cost ($/gram) for Different Levels of Off-cycle Credit (2018 dollars)
Off-cycle credits
2021
2022
2023
2024
2025
2026
2027
0 to 5 g/mi
$12
$34
$38
$38
$22
$28
$28
5 to 10 g/mi
$11
$29
$24
$34
$34
$40
$42
10 to 15 g/mi
$7
$20
$2
$0
$0
-$2
$0
In other words, with 15 grams/mile of free off-cycle credits, the cost per vehicle is essentially
the same as with 10 grams/mile of free off-cycle credits. As a result, the cost/gram of those 5
additional credits is suggested to be essentially $0. Looked at another way, the MY 2026 cost
when no off-cycle credits are made available was $1,030. With 10 grams of free off-cycle credits
available, the costs reduced to $690. This suggests that the value of the 10 grams of off-cycle
4-6
-------
credits were $34/gram ($1030 - $690 = $340; $340/10=$34). Some of those credits were
available at $28/gram in the 0 to 5 g/mile increment and others were available at $40/gram in the
5 to 10 g/mile increment (see Table 4-6 for MY 2026). Given that the largest incremental cost
shown is $42, and the desire to be conservative, we have used a value of $42/grams/mile for off-
cycle credits. Importantly, since identical off-cycle credits are projected in all scenarios in our
modeling, the incremental costs between scenarios is not impacted by this valuation of off-cycle
credits.
4.1.1.2 Battery Costs and Changes since the Proposed Rule
In the proposed rule we used the battery costs and battery cost learning curve used by NHTSA
in the SAFE FRM. This learning curve is part of the technologies input file. As described in
more detail in Section 2.3.4, we wanted to adjust the battery costs applied in CCEMS to better
represent cost savings represented by recent developments, such as increased battery
manufacturing capacity and economies of scale, cathode chemistries with reduced cobalt content,
and also cell capacities and pack topologies that are more consistent with emerging dedicated
BEV platforms (see Chapter 2.3.4). However, those costs are integrated into the executable file
for the model and thus battery pack costs that are used by the model cannot be modified without
recompiling the code, a step we did not want to make since it might imply that the model code
was different and not just the battery cost input file.
An alternative approach was to modify the placement of the learning curve values that are
applied to the battery pack cost inputs as those inputs are entered into the model. By shifting the
battery cost learning curves, we could, in effect, adjust the battery costs prior to initiation of the
compliance modeling.
First, we conducted an assessment to determine by how much the costs should be reduced.
EPA reviewed the battery costs used in the SAFE rulemaking, which had been carried over to the
analysis for EPA's August 2021 Notice of Proposed Rulemaking (NPRM). We considered the
inputs that had previously been used to derive the costs for the SAFE rulemaking, and compared
those costs to the costs that EPA had derived in previous and ongoing analyses. The costs were
also compared to the current and expected future costs of batteries as widely reported in the trade
and academic literature. We concluded that the battery costs used in the proposal were broadly
higher than indicated by this evidence, and that the likely effect of using an updated set of
assumptions would produce projected battery costs significantly lower than those proposed and
more in agreement with emerging consensus on the level and direction of battery costs in the
industry.
In the SAFE FRM, the agencies used the Argonne National Laboratory BatPaC 3.1 model as a
basis for developing battery costs, and chose a set of inputs that included an annual production
volume of 25,000 packs, a BEV battery chemistry of NMC622-G, and a modified cell yield rate
that was lower than the recommended value provided by the authors of BatPaC. Cell capacities
and pack topologies can also have a strong influence on pack cost, and we noted that several
OEMs have begun production of dedicated BEV vehicle platforms that use larger, standardized
cell and module designs.
In previous analyses, EPA estimated battery pack costs using annual production volumes of
50,000, 125,000, 250,000 and 450,000 packs per year. Given the scale and size of battery plants
in operation now and planned for the future, and given the increasing use of relatively
4-7
-------
standardized cells that can be used in more than one capacity of pack, we believe that an annual
volume of 125,000 units is more appropriate to represent the manufacturing capacity of battery
plants over the timeframe covered by the analysis for the Final Rule. Given increased use of low-
cobalt NMC811-G chemistry, and ongoing pressure to reduce battery costs and reliance on
imported cobalt, we believe that it is now a more appropriate choice for BEV battery chemistry.
We do not find evidence that the cell yield rate should be modified from the BatPaC default
value. We also find that cell capacities and pack topologies should be chosen to be consistent
with emerging dedicated BEV platforms such as for example, the GM Ultium platform, the
Volkswagen MEB platform and the Hyundai e-GMP platform which all use relatively large
capacity cells of relatively fixed Ampere-hour capacities in well-defined pack topologies and are
all based upon vehicle platforms shared among multiple vehicle models with the underlying
platforms intended for high-volume vehicle production (i.e., approximately 125,000 vehicles per
year or more). Based on an assessment of the effect of using these updated inputs to the BatPaC
4.0 model in place of those used in the SAFE rulemaking, we found technical justification for
reducing battery costs by approximately 25 percent, as described below.
In the SAFE final rulemaking analysis, NHTSA cited an example 60 kWh BEV battery and
stated a direct manufacturing cost at $178/kWh in 2020 and $141 in 2025, based on BatPaC
outputs using the inputs they had selected. Cost inputs to the CCEMS model are considered
RPE-inclusive costs, and thus direct manufacturing costs derived for the analysis are to be
multiplied by the 1.5 RPE multiplier before being input to the model. The cost for this
hypothetical $178/kWh battery would thus be input to the CCEMS model as $267/kWh, an RPE-
inclusive cost applicable to 2020. Similarly for 2025, this $141/kWh battery (in 2025) would be
input to the CCEMS model as $212/kWh, an RPE-inclusive cost for 2025.
Current trends and broad consensus on the state and direction of battery costs indicate that the
direct manufacturing cost of a 60 kWh BEV battery was likely lower than $178/kWh in 2020 and
will likely be lower than $141/kWh in 2025. Using the updated BatPaC 4.0 and using the input
assumptions described above, BatPaC indicates a cost of about $129/kWh, or an RPE-inclusive
cost of $194/kWh. Because the previous RPE-inclusive cost for this battery was $267/kWh, this
suggests that its cost could be reduced by about 25 percent. Because the indicated change
resulted primarily from factors that would affect batteries across the analysis in a roughly similar
manner (production volume, lower-cobalt chemistry, cell yield, and larger cell capacity) we
concluded that it was reasonable to apply a similar battery cost reduction across the analysis.
As a means to adjust the battery cost inputs to the CCEMS model, we adjusted the battery
cost learning curve such that the curve now applies a learning factor of 1.0 six years earlier than
previously, which in effect results in battery cost inputs to the CCEMS model being reduced by
about 24 percent, consistent with the results of our updated technical assessment. This does not
represent a reconsideration of learning inputs but only acts as a mechanism to apply a correction
factor to the battery costs being input to the CCEMS model.
For the example 60 kWh BEV battery, this results in the cost changes shown in Table 4-7
below.
4-8
-------
Table 4-7. Cost Changes for a 60 kWh BEV Battery

2020
2021
2022
2023
2024
2025
2026
Previous, RPE inclusive
$267
$255
$244
$233
$222
$212
$203
Updated, RPE inclusive
Updated, direct mfg. cost
$203
$135
$193
$129
$185
$123
$178
$119
$168
$112
$151
$107
$154
$102
We note that the resultant direct costs for current and future years within the time frame of the
rule are consistent with the preponderance of publicly available reports and projections in the
industry, which broadly project BEV battery costs reaching approximately $100/kWh at the pack
level by mid-decade. For example, the National Academies of Sciences (NAS) Phase 3 Report
projects pack-level costs between $90-$ 115/kWh by 2025.14 For the year 2020, the updated
estimate of $135/kWh closely corresponds to the volume-averaged OEM price of $137/kWh
reported for that year by the Bloomberg New Energy Finance (BNEF) 2020 Battery Price
Survey.15
We also considered the effect of this reduction on the projected battery costs for future years
beyond the time frame of the rule. Applying the existing learning curve to the downward
adjusted costs past the time frame of the rule would produce costs gradually declining to below
$80 per kWh (for an example 60 kWh battery) in the mid-2030s and to about $75/kWh by the
mid-2040s.
At this time, EPA is uncertain about the potential for battery costs to reach those levels due in
part to uncertainties about the effect of increased demand for critical minerals and other factors,
which we also received comment on, and also because our current battery modeling tools such as
BatPaC 4.0 are unable to generate costs at these levels using inputs that can reasonably be
validated. Moreover, the concept of a learning curve normally applies to "learning by doing,"
that is, it represents savings that result from incremental improvements in manufacturing
processes or small design changes for a specific form of a technology and is not intended to
represent savings that might result from a step change to a different form of the technology.
Many forecasts that anticipate continued lowering of battery costs below a level that can be
technically demonstrated today incorporate the assumption that step changes to the form of the
battery cell, for example, a shift to lithium-metal anodes or solid-state construction, will make
the projected costs possible. Although EPA believes that cost reductions from these new forms
of battery technology are likely to occur in the future, EPA is uncertain if it is appropriate to
account for them by applying a learning curve to costs applicable to the current form of the
technology. Significantly, these new forms of cell technology have not yet been demonstrated in
large volume automotive applications, making it difficult to estimate their cost reduction
potential with a reasonable degree of certainty.
Even though application of a learning curve to a specific technology can be said to assume
unspecified improvements in manufacturing efficiencies or small design changes, the future costs
projected by the curve should still be capable of being validated by reasonable assumptions for
these efficiencies and changes, within reasonable technical boundaries of the specific
technology. Using the current battery cost modeling tools at our disposal, and using a set of
4-9
-------
reasonable assumptions applicable to the prevailing form of lithium-ion chemistries, it is possible
to reasonably validate the potential for costs to be reduced to approximately $90 per kWh for a
60 kWh battery. Due to the widely acknowledged uncertainty of quantitatively projecting
declines in battery costs far into the future, and particularly in the context of the downwardly
adjusted battery costs, we chose to flatten the rate of learning past 2029 so as to prevent future
costs from declining below $90 per kWh for a 60 kWh battery, a level that we can currently
technically validate and which corresponds to the more optimistic end of the NAS estimate for
2025. Although EPA believes that this reflects an appropriate technical application of a learning
curve, it does not represent the potential for cost reductions from step changes to the technology
that may occur in the future. Therefore, in years beyond the time frame of the rule, the adjusted
costs are conservative with respect to the forecasts of BNEF ($58 in 2030) and NAS ($65-$80 by
2030), both of which we believe would require some of their projected cost reduction to result
from significant changes to the technology that have not yet been demonstrated.
With regard to the reliability of battery cost forecasts beyond 2030, NAS on p. 5-139 of the
Phase 3 report states in the context of their own analysis, "as there is higher uncertainty related to
battery technology past 2030, rigorous cost estimates past this point are not attempted."16 We
agree with the implication that battery cost estimates past 2030 are by their nature highly
uncertain and difficult to quantify in a rigorous way. While the NAS Phase 3 report did offer a
forecast of $65 to $80 per kWh for 2030, we also note that this forecast appears to be derived
qualitatively from a number of forecasts gathered from the literature, that vary considerably in
their technical bases and assumptions, and are not focused on recently emergent issues such as
production capacity and mineral demand. Significantly, the list of cited sources omits a 2019
MIT study that explicitly considered the effect of mineral costs on the potential for future
reductions in battery cost to achieve the levels forecast by some of the cited studies.17'18 MIT
concludes that the potential for cost reduction is likely to be limited by the cost of raw materials
and may plateau in a range closer to $100 per kWh for the most widely used family of lithium-
ion chemistries. Our choice of $90 per kWh as a lower limit is thus within the range bounded by
the estimates of NAS and MIT.
It is also important to note that the costs referenced here pertain only to an example 60 kWh
battery. Many EVs are expected to have larger batteries, with 75 to 100 kWh already common in
the market. Battery cost on a dollar per kWh basis tends to decline as battery packs get larger in
capacity. Thus, the equivalent lower limit in our analysis for a 75 kWh or 100 kWh battery that is
presented to the CCEMS model would be less than $90 per kWh in 2029 and later.
As already noted in Section 2.3.4 of this RIA, we believe that holding learning constant after
2029 is likely a conservative assumption, as we continue to expect that continued learning and/or
cost reductions resulting from a change to solid-state or lithium-metal technology or other
developments will occur beyond 2029 but there is uncertainty at this point on what the
appropriate rate of cost reduction resulting from these sources would be. Thus, our battery cost
estimates beyond 2029 in this final rulemaking may be conservatively high.
In consultation with battery experts at the Department of Energy Vehicle Technologies Office
(DOE VTO), we jointly agreed that it is appropriate to be conservative regarding estimates of
future battery costs, including the use of the lower limit as an interim step to represent current
uncertainty about the ability for the cost of current-generation technology to learn down in the
face of uncertain future material costs. Although DOE often makes estimates of future battery
4-10
-------
costs that are nominally below this limit, these estimates are meant to guide research and are
often targets or "stretch goals," not necessarily predictions, as they anticipate future technical
progress which is difficult to predict. For example, the US Advanced Battery Consortium
(USABC) has set a long-term goal of $75 per kWh, which serves as a target that guides ongoing
research by representing a goal that is qualitatively regarded as achievable although not yet
achieved. At the same time, these targets reflect significant promise for future cost reductions by
means of several potential pathways. These include, among others, reduced use or elimination of
critical and expensive materials such as cobalt and nickel, lower material costs through improved
efficiencies and economies of scale, increased energy density through use of next-generation
materials like silicon or lithium metal anodes and higher energy cathodes (resulting in less
material needed per kWh), and more effective material use and recycling practices within the
factory.
DOE actively conducts research and target setting activities that could improve EPA's ability
to better quantify future battery costs in a subsequent rulemaking. DOE VTO currently funds
cost modeling at Argonne National Laboratory, including near-term improvements to the BatPaC
model that will model advanced cell formats, pack formats, and chemistries, including next-
generation materials. In battery research, DOE spends about $10-15 million per year on silicon-
based technologies and cells, $10-15 million per year on lithium-metal technologies, $10-15
million per year on low- or no-cobalt cathodes, and $7-12 million per year on battery recycling
technologies. The Infrastructure Investment and Jobs Act will provide $6 billion in federal
funding to develop a domestic lithium-ion energy storage supply chain and an additional $320
million on recycling and secondary use activities.
EPA continues to study the potential for cost reductions in batteries to occur during and after
the time frame of the rule. For example, we expect that the aforementioned updates to the ANL
BatPaC model, as well as collection of emerging data on forecasts for future mineral prices and
production capacity, will make it possible to characterize the continued declines in battery costs
that we continue to believe will occur after 2029, as well as trends in costs in the nearer term.
These developments are likely to improve our ability to quantify the potential for cost reductions
past 2029, in place of the lower limit we have assumed for this analysis, and we plan to
incorporate this information in the subsequent rulemaking for MYs 2027 and beyond.
4.1.1.3 Restricting HCR2 Technology from the Available Technologies
The HCR2 technology in this version of CCEMS would require a level cylinder deactivation
technology, dynamic cylinder deactivation, that has not yet been added to Atkinson Cycle
Engines either with or without cooled EGR.b HCR1 technologies reflect the effectiveness of
Atkinson Cycle engines with either cooled EGR or fixed cylinder deactivation (however, not
both technologies in combination) and thus also represent a number of high-volume ICE
applications from Mazda, Toyota and Hyundai. The additional step to HCR2 reflected a level of
ICE effectiveness that, while technologically feasible, is not yet within the light-duty vehicle
b Dynamic cylinder deactivation allows any number of cylinders to be deactivated on a cycle-resolved basis. Fixed
cylinder deactivation deactivation deactivates a fixed number of cylinders under certain operating conditions (e.g.,
deactivating 2 out of 4 cylinders). Dynamic cylinder deactivation is currently used by GM in pickup truck and full-
size SUV applications, but has not yet been used in combination with Atkinson Cycle in a production application.
Fixed cylinder deactivation is now standard on Mazda implementations of Atkinson Cycle.
4-11
-------
fleet, and that we also do not anticipate seeing until the later years of this final rule (e.g., MYs
2025-2026). For the final rule, we chose a more conservative approach. We have chosen to set
its availability to FALSE in the technologies input file for the Final Rule and to add a sensitivity
analysis on HCR2 for MYs 2025 and later. This allows us to show two different compliance
paths - one with and one without HCR2. This has two impacts on the results of the CCEMS
analysis for the final rule:
•	Slightly higher technology costs (less than $15 per vehicle) since a highly cost-
effective gasoline technology is not available
•	Slightly higher electrified technology penetrations (BISG, HEV, PHEV, BEV) since
more electrification is required than would otherwise be the case, and because vehicle
electrification costs have been reduced in the final rule due to updated (lower) battery
costs.
For more information on HCR technologies, please refer to Chapter 2.3.2. For results of the
sensitivity analysis using HCR2, please refer to Chapter 4.1.5.1.
4.1.1.4 Shifting of Input File Years due to the Updated Baseline Fleet
As noted in Table 4-2, we have updated the baseline fleet to reflect the MY 2020 fleet rather
than the MY 2017 fleet used in the NPRM. However, due to hardcoded entries in the CCEMS
code, we had to create an offset within the model to allow use of the MY 2020 fleet instead of
the MY 2017 fleet. To do this, every year-based entry in each of the model input files had to be
adjusted by 3 years such that any data were entered in an input column shifted by 3 years. For
example, in the scenarios input file, standard curve coefficients for MYs 2023 through 2026 are
entered in columns with the headers 2020 through 2023. Similarly, banked credits in the market
input file are entered in columns headed by 2012 through 2016 even though those banked credits
are actually 2015 through 2019 vintage credits. This was also done to refresh and redesign
schedules within the market file and for all of the technology costs and learning curve factors in
the technologies file. Importantly, in the post-processing tool used to recombine Framework and
non-Framework manufacturer model runs, the years are shifted back to their proper timing such
that, while the direct model output files present data shifted by 3 years, the post-processed files
reflect actual calendar and model year data.
4.1.2 GHG Targets and Compliance Levels
4.1.2.1 Final Standards
The final standard curve coefficients are presented in Preamble Section II.A.2. Here we
present the fleet targets for each manufacturer. Figure 4-1 depicts the fleet targets of the SAFE
FRM and today's final standards. Also shown are the targets from the 2012 FRM and the
proposed standards from the August 2021 NPRM (Proposal). As can be seen, the final standards
move from the SAFE FRM in the same manner as the Proposal in MYs 2022 and 2023. It then
achieves greater stringency than the Proposal and surpasses the 2012 FRM targets by MY 2026.
4-12
-------
260
240
220
_OJ
bO
(N
O
U
200
180
160

• • • SAFE FRM

• •2012 FRM

•Proposal
X _ 		
^^"Final Standards
• •
• •
• •
140
2020 2021 2022 2023 2024 2025 2026 2027
Model Year
Figure 4-1: Final Fleet-Wide CCh-Equivalent g/mi Compliance Targets (solid black line), Compared to 2012
FRM, SAFE Rule, and Proposal.
These targets are dependent on each manufacturer's car and truck fleets and their sales
weighted footprints. As such, each manufacturer has a set of targets unique to them. Those
targets are shown by manufacturer for MYs 2023 through 2026 in Table 4-8 for cars, Table 4-9
for trucks, and Table 4-10 for the combined fleet.
4-13
-------
Table 4-8: Car Targets (CO2 gram/mile)
Manufacturer
2023
2024
2025
2026
BMW
169
161
152
135
Daimler
174
166
156
139
FCA
176
168
158
140
Ford
170
162
153
136
General Motors
163
155
147
130
Honda
164
156
147
130
Hyundai Kia-H
165
157
148
131
Hyundai Kia-K
163
155
146
129
JLR
171
163
154
136
Mazda
163
155
147
130
Mitsubishi
153
145
137
120
Nissan
166
158
149
132
Subaru
159
152
143
126
Tesla
179
171
161
144
Toyota
164
156
147
130
Volvo
176
168
158
141
VWA
164
156
148
131
TOTAL
166
158
149
132
Table 4-9: Light Truck Targets (CO2 gram/mile)
Manufacturer
2023
2024
2025
2026
BMW
227
216
201
182
Daimler
227
216
201
182
FCA
241
229
213
193
Ford
249
237
220
200
General Motors
252
240
223
203
Honda
216
205
191
172
Hyundai Kia-H
231
219
204
184
Hyundai Kia-K
218
207
193
174
JLR
223
212
197
177
Mazda
206
196
182
163
Mitsubishi
194
184
171
153
Nissan
221
210
195
176
Subaru
202
192
178
160
Tesla
236
224
209
189
Toyota
227
215
201
181
Volvo
222
211
196
176
VWA
214
203
189
170
TOTAL
234
222
207
187
4-14
-------
Table 4-10: Sales Weighted Fleet Targets (CO2 gram/mile)
Manufacturer
2023
2024
2025
2026
BMW
190
181
170
152
Daimler
200
190
177
159
FCA
231
219
204
185
Ford
228
217
202
183
General Motors
221
210
196
177
Honda
186
176
165
147
Hyundai Kia-H
171
163
153
136
Hyundai Kia-K
182
172
161
144
JLR
220
209
195
175
Mazda
184
175
164
146
Mitsubishi
174
165
155
137
Nissan
181
172
162
144
Subaru
191
182
169
151
Tesla
180
172
162
145
Toyota
191
181
169
151
Volvo
210
200
186
167
VWA
193
183
171
153
TOTAL
202
192
179
161
The actual achieved CCte-equivaient (CChe) levels, which include the effect of credit programs on
compliance, are shown in Table 4-11 for cars, Table 4-12 for trucks, and Table 4-13 for the
combined fleets.
Table 4-11: Car Achieved (CChe gram/mile)
Manufacturer
2023
2024
2025
2026
BMW
192
173
138
121
Daimler
171
150
158
155
FCA
160
152
163
149
Ford
158
157
158
146
General Motors
163
158
158
153
Honda
163
153
147
138
Hyundai Kia-H
160
149
134
132
Hyundai Kia-K
166
155
143
142
JLR
224
188
189
189
Mazda
166
146
146
145
Mitsubishi
186
185
127
126
Nissan
170
157
132
132
Subaru
201
189
188
168
Tesla
-10
-10
-10
-10
Toyota
161
138
134
132
Volvo
207
204
198
181
VWA
165
153
156
127
TOTAL
160
148
140
134
4-15
-------
Table 4-12: Light Truck Achieved (CChe gram/mile)
Manufacturer
2023
2024
2025
2026
BMW
197
197
203
203
Daimler
229
229
193
84
FCA
215
212
210
189
Ford
250
222
222
192
General Motors
265
238
217
193
Honda
214
167
163
163
Hyundai Kia-H
268
267
266
127
Hyundai Kia-K
209
188
195
194
JLR
214
203
179
146
Mazda
203
202
177
118
Mitsubishi
227
226
130
130
Nissan
205
200
195
181
Subaru
186
175
167
167
Tesla
-9
-9
-9
-9
Toyota
236
208
216
176
Volvo
158
156
162
161
VWA
213
203
171
147
TOTAL
230
211
203
178
Table 4-13: Sales Weighted Fleet Achieved (CChe gram/mile)
Manufacturer
2023
2024
2025
2026
BMW
194
182
162
151
Daimler
199
188
175
122
FCA
206
202
203
183
Ford
225
205
205
180
General Motors
230
210
196
179
Honda
184
159
153
148
Hyundai Kia-H
171
160
147
131
Hyundai Kia-K
180
166
160
159
JLR
215
203
179
149
Mazda
184
173
161
132
Mitsubishi
207
206
128
128
Nissan
180
169
150
145
Subaru
190
178
173
168
Tesla
-10
-10
-10
-10
Toyota
192
167
168
150
Volvo
170
169
172
166
VWA
193
182
164
139
TOTAL
197
181
173
157
Note that the values shown in Table 4-11 through Table 4-13 are modeled tailpipe
certification values considering use of AJC leakage credits and other off-cycle credits apart from
AJC leakage. This explains the negative 10 grams/mile CChe shown for Tesla cars. That value
reflects 5 grams/mile of AJC efficiency credits and another 5 grams/mile of off-cycle credits.
These are simply the input values for Tesla flowing through the model. To date, Tesla has not
been a major user of the off-cycle credit program given that they make nothing but BEVs.
However, when running the model, we have chosen to apply the off-cycle credit inputs
developed by NHTSA in support of their recent NPRM both on the credit side and the cost side
4-16
-------
for any year in which that credit is available.19 This differs from the proposed rule where we
modeled every manufacturer earning the maximum amount of off-cycle credits available in any
given year and under any given scenario (10 grams/mile under the SAFE FRM standards; 15
grams/mile under the proposed standards). However, this added an incremental cost,
automatically, of nearly $400 to every vehicle because CCEMS does not have the ability to
weigh the application of off-cycle credit technology versus other technologies when making
technology application decisions.
4.1.2.2 Alternatives
Table 4-14, Table 4-15 and Table 4-16 show the car, truck and fleet targets, respectively, for
the alternatives to the final rule. The alternatives that were analyzed include:
•	A less stringent alternative, which were the proposed light-duty GHG standards from
the August 2021 NPRM (Proposal)
•	A more stringent alternative to the final standards having lower CO2 standards for MY
2023 and MY 2024 than what has been finalized (Alternative 2 minus 10)
•	A graphical representation of the fleet average targets is also shown in Figure 4-2.
260
240
220
200
180
160

• SAFE FRM

-2012 FRM
• /
•7
'/
— —Proposal
• • • Alternative 2 minus 10
" v • ¦.

\ \
Final Standards
^ \
N \
140
2020	2021	2022	2023	2024	2025	2026	2027
Model Year
Figure 4-2: Final Rule Fleet Average Targets Compared to the Proposal and Alternative 2 minus 10
4-17
-------
The actual achieved CCtee levels, which include credit programs and their effect on
compliance, are shown in Table 4-17, Table 4-18 and Table 4-19 for cars, trucks and the
combined fleet.
Table 4-14: Car Targets under the Proposal and Alternative 2 minus 10 Standards (CO2 gram/mile)

Proposal Standards
Alternative 2 minus 10 Standards
Manufacturer
2023
2024
2025
2026
2023
2024
2025
2026
BMW
169
161
153
146
167
159
152
135
Daimler
174
166
157
150
171
164
156
139
FCA
176
168
159
151
172
165
158
140
Ford
170
162
154
147
168
160
153
136
General Motors
163
156
148
141
161
154
147
130
Honda
164
156
149
142
162
155
147
130
Hyundai Kia-H
165
157
150
142
163
156
148
131
Hyundai Kia-K
163
155
147
140
160
153
146
129
JLR
171
163
155
147
168
161
154
136
Mazda
163
155
148
141
161
154
147
130
Mitsubishi
153
145
138
132
150
144
137
120
Nissan
166
158
151
143
164
156
149
132
Subaru
159
152
144
137
157
150
143
126
Tesla
179
171
162
155
177
169
161
144
Toyota
164
156
149
142
162
155
147
130
Volvo
176
168
160
152
174
166
158
141
VWA
164
156
149
142
162
155
148
131
TOTAL
166
158
150
143
164
156
149
132
Notes:
Hyundai Kia-H is Hyundai and Hyundai Kia-K is Kia. While these companies are part of Hyundai-Kia, they
operate independently for the purpose of compliance with our GHG program.
Table 4-15: Light Truck Targets under the Proposal and Alternative 2 minus 10 Standards (CO2 gram/mile)

Proposal Standards
Alternative 2 minus 10 Standards
Manufacturer
2023
2024
2025
2026
2023
2024
2025
2026
BMW
227
216
205
195
222
211
201
182
Daimler
227
216
205
195
222
211
201
182
FCA
241
229
217
206
235
224
213
193
Ford
249
237
225
214
243
231
220
200
General Motors
252
240
228
216
246
234
223
203
Honda
216
205
195
185
211
200
191
172
Hyundai Kia-H
231
219
208
198
225
214
204
184
Hyundai Kia-K
218
207
198
188
213
203
193
174
JLR
223
212
201
191
217
207
197
177
Mazda
206
196
186
177
201
191
182
163
Mitsubishi
194
184
175
166
189
180
171
153
Nissan
221
210
200
190
216
205
195
176
Subaru
202
192
182
173
197
187
178
160
Tesla
236
224
213
202
230
219
209
189
Toyota
227
214
205
195
221
211
201
181
Volvo
222
211
200
190
216
206
196
176
VWA
214
203
193
183
209
198
189
170
TOTAL
234
222
211
200
228
217
207
187
4-18
-------
Table 4-16: Fleet Targets under the Proposal and Alternative 2 minus 10 Standards (CO2 gram/mile)

Proposal Standards
Alternative 2 minus 10 Standards
Manufacturer
2023
2024
2025
2026
2023
2024
2025
2026
BMW
190
181
172
164
187
178
170
152
Daimler
200
190
180
171
196
187
177
159
FCA
231
219
208
197
225
215
204
185
Ford
228
217
206
196
223
212
202
183
General Motors
221
210
200
189
217
206
196
177
Honda
186
176
168
160
183
174
165
147
Hyundai Kia-H
171
163
155
147
169
162
153
136
Hyundai Kia-K
182
172
164
156
178
170
162
144
JLR
220
209
198
189
214
205
195
175
Mazda
184
175
166
158
181
172
164
146
Mitsubishi
174
165
157
149
170
163
155
137
Nissan
181
172
164
156
179
170
162
144
Subaru
191
182
172
163
187
178
169
151
Tesla
180
172
163
156
178
170
162
145
Toyota
191
180
172
164
187
179
170
151
Volvo
210
200
190
180
205
196
186
167
VWA
193
183
174
165
189
180
171
153
TOTAL
202
192
182
173
198
189
180
161
Table 4-17: Car Targets Achieved under the Proposal and Alternative 2 minus 10 Standards (CChe
gram/mile)

Proposal Standards
Alternative 2 minus 10 Standards
Manufacturer
2023
2024
2025
2026
2023
2024
2025
2026
BMW
191
171
151
135
189
165
145
130
Daimler
174
153
166
163
180
170
169
168
FCA
153
146
163
149
170
162
161
148
Ford
153
153
158
146
156
156
150
138
General Motors
148
143
150
146
132
126
121
116
Honda
164
162
155
145
164
156
147
138
Hyundai Kia-H
160
150
137
135
159
147
141
139
Hyundai Kia-K
165
155
143
142
167
158
152
151
JLR
224
202
202
202
223
196
196
196
Mazda
164
142
146
145
161
144
138
137
Mitsubishi
185
184
132
131
188
187
127
127
Nissan
169
157
142
141
167
154
128
127
Subaru
201
189
188
180
201
189
188
166
Tesla
-10
-10
-10
-10
-10
-10
-10
-10
Toyota
160
139
136
135
161
144
139
136
Volvo
206
203
198
181
209
205
202
184
VWA
167
155
164
145
166
146
142
117
TOTAL
157
147
143
138
156
145
136
131
4-19
-------
Table 4-18: Light Truck Targets Achieved under the Proposal and Alternative 2 minus 10 Standards (CChe
gram/mile)

Proposal Standards
Alternative 2 minus 10 Standards
Manufacturer
2023
2024
2025
2026
2023
2024
2025
2026
BMW
198
198
208
208
191
191
191
191
Daimler
227
226
194
161
173
172
167
60
FCA
217
214
216
206
225
222
212
191
Ford
249
222
224
211
245
226
222
193
General Motors
268
246
227
213
270
247
235
209
Honda
217
184
180
180
213
174
163
163
Hyundai Kia-H
270
269
268
196
257
256
255
78
Hyundai Kia-K
214
196
204
203
198
182
181
180
JLR
212
204
194
161
214
199
167
135
Mazda
204
203
177
160
201
200
176
132
Mitsubishi
227
226
158
158
227
226
130
130
Nissan
213
211
208
194
208
205
197
183
Subaru
185
176
170
170
186
173
162
162
Tesla
-9
-9
-9
-9
-9
-9
-9
-9
Toyota
239
209
221
202
226
208
208
168
Volvo
169
168
176
175
171
170
165
163
VWA
209
203
181
172
221
203
164
155
TOTAL
232
215
210
198
229
214
204
179
Table 4-19: Fleet Targets Achieved under the Proposal and Alternative 2 minus 10 Standards (CChe
gram/mile)

Proposal Standards
Alternative 2 minus 10 Standards
Manufacturer
2023
2024
2025
2026
2023
2024
2025
2026
BMW
194
181
172
161
190
174
162
152
Daimler
199
188
179
162
176
171
168
117
FCA
207
203
208
197
216
213
204
184
Ford
224
203
206
193
221
207
203
178
General Motors
226
210
200
189
222
205
195
176
Honda
187
171
165
160
185
164
154
149
Hyundai Kia-H
171
161
149
141
168
157
152
134
Hyundai Kia-K
182
169
163
162
177
166
161
161
JLR
213
204
194
164
214
199
168
139
Mazda
183
171
161
152
181
171
156
134
Mitsubishi
207
206
145
145
208
207
129
128
Nissan
182
172
160
155
179
169
147
143
Subaru
189
179
175
173
190
177
169
163
Tesla
-10
-10
-10
-10
-10
-10
-10
-10
Toyota
193
168
171
162
189
171
167
149
Volvo
179
177
182
176
181
179
174
169
VWA
191
182
174
160
198
179
155
138
TOTAL
197
183
178
169
195
182
172
156
4-20
-------
4.1.3 Projected Compliance Costs per Vehicle
4.1.3.1 Final Standards
EPA has performed an updated assessment of the per vehicle costs for manufacturers to meet
the revised MY 2023-2026 standards. Importantly, we applied off-cycle credits at the levels
entered in the market file as projected by NHTSA in their recent NPRM20 and applied costs for
those credits in a post-processing step at the valuation described in Chapter 4.1.1.1 21 The car
costs per vehicle are shown in Table 4-20, Table 4-21 and Table 4-22 for cars, trucks and the
combined fleet, respectively.
Table 4-20: Car Costs/Vehicle Relative to the No Action Scenario (2018 dollars)
Manufacturer
2023
2024
2025
2026
BMW
$8
$112
$840
$762
Daimler
$232
$542
$480
$479
FCA
$253
$212
$158
$329
Ford
$19
$18
$227
$202
General Motors
$577
$546
$651
$669
Honda
$67
$310
$362
$329
Hyundai Kia-H
$92
$132
$756
$790
Hyundai Kia-K
$170
$273
$644
$619
JLR
$26
$619
$581
$547
Mazda
$5
$394
$471
$425
Mitsubishi
$0
$0
$914
$898
Nissan
$228
$327
$1,289
$1,194
Subaru
$18
$18
$17
$209
Tesla
$0
$0
$0
$0
Toyota
$21
$429
$576
$578
Volvo
$0
-$1
$119
$113
VWA
$0
$60
$125
$549
TOTAL
$150
$288
$586
$596
Table 4-21: Light Truck Cost per Vehicle Relative to the No Action Scenario (2018 dollars)
Manufacturer
2023
2024
2025
2026
BMW
$2
$2
$2
$9
Daimler
$35
$34
$725
$3,556
FCA
$1,732
$1,574
$1,465
$1,894
Ford
$39
$477
$428
$754
General Motors
$385
$702
$1,377
$1,746
Honda
$118
$915
$950
$878
Hyundai Kia-H
$45
$44
$43
$4,048
Hyundai Kia-K
$1,194
$1,327
$1,230
$1,144
JLR
$133
$314
$1,321
$1,770
Mazda
$11
$11
$776
$2,500
Mitsubishi
$0
$0
$2,159
$2,028
Nissan
$699
$783
$748
$1,082
Subaru
$2
$27
$57
$57
Tesla
$0
$0
$0
$0
Toyota
$265
$832
$763
$1,537
Volvo
$958
$853
$771
$702
VWA
$0
$125
$461
$856
TOTAL
$485
$732
$909
$1,356
4-21
-------
Table 4-22: Fleet Average Cost per Vehicle Relative to the No Action Scenario (2018 dollars)
Manufacturer
2023
2024
2025
2026
BMW
$6
$72
$538
$489
Daimler
$136
$298
$591
$1,925
FCA
$1,502
$1,355
$1,254
$1,639
Ford
$34
$353
$373
$604
General Motors
$452
$648
$1,123
$1,369
Honda
$88
$563
$606
$557
Hyundai Kia-H
$87
$123
$688
$1,093
Hyundai Kia-K
$518
$624
$840
$797
JLR
$128
$332
$1,283
$1,708
Mazda
$7
$207
$612
$1,411
Mitsubishi
$0
$0
$1,557
$1,482
Nissan
$360
$453
$1,143
$1,166
Subaru
$6
$26
$50
$101
Tesla
$0
$0
$0
$0
Toyota
$125
$597
$655
$978
Volvo
$714
$634
$603
$551
VWA
$0
$97
$318
$727
TOTAL
$330
$524
$759
$1,000
Overall, EPA estimates the costs of the final standards at $1,000 per vehicle relative to the no
action scenario. The increase in costs between MYs 2024 and 2025 under the rule is a result of
the elimination of advanced technology multiplier credits in combination with the increased
stringency between MY 2024 and MY 2025.
Of note is the difference in costs per vehicle for the Framework manufacturers (BMW, Ford,
Honda, Volvo and VWA) and the non-Framework manufacturers (with the exception of Subaru)
with the Framework manufacturers showing several hundreds of dollars lower costs. Since the
Framework manufacturers are incurring costs associated with the Framework, their incremental
costs to meet the final standards, which are more stringent than the Framework, are lower than
for those non-Framework manufacturers that have chosen to comply with the SAFE FRM.
The MY 2026 projected cost per vehicle is roughly the same as was previously estimated for
the proposed standards in the NPRM despite the final standards being more stringent (RIA
4.1.3.2 presents the updated costs per vehicle for the proposed standards as an alternative). This
is due primarily to two factors: lower battery costs resulting in more BEVs and, in turn, less
technology applied to gasoline vehicles. The first factor is the updated battery costs, which are
about 24 percent lower than those estimated in the NPRM. This reduces the per vehicle cost of
electrified vehicles, which in turn increases the technology penetration of electrified
technologies. The increased penetration of electrified vehicle technologies, and especially BEV
technology (which with a 0 g/mile compliance value having such a large impact), results in less
technology application on conventional, non-electrified gasoline and diesel vehicles. These two
factors, lower battery costs resulting in more BEVs and less technology applied to gasoline
vehicles explains how, on average, the final rule's per vehicle costs are so similar to the
proposal's per vehicle costs. In Section 4.1.5 where we present our sensitivity results, we present
more information surrounding this BEV penetration impact on ICE technology costs. We also
present results with lower battery costs and higher battery costs than used in our primary analysis
4-22
-------
to present what costs could be if battery costs are lower or higher than we have projected. We
also present costs that exclude hybridization to show what might happen if industry chooses to
rely largely on ICE and plug-in vehicles rather than investing in hybridization too.
Costs per vehicle continue to rise through MY 2028 before starting to decrease as shown in
Table 4-23. Importantly, the final rule results in higher BEV penetration estimates than those
estimated in the NPRM. Also, we have chosen to stop the rate of learning after 2029, unlike in
the NPRM, where we allowed battery costs to continue to decline. Applying the existing learning
rates to the downward adjusted costs past the time frame of the rule would have produced costs
gradually declining to levels that are difficult to empirically support at this time, due in part to
uncertainties about the effect of increased demand for critical minerals and other factors, which
we also received comment on. To account for this uncertainty in the context of the downward
adjusted battery costs, we held the battery cost learning curve constant after 2029 to prevent
projected future reductions in cost to exceed what we can currently technically demonstrate. This
has the effect of reducing the rate of per-vehicle cost reductions year-over-year in the outer years
out to 2050, compared to the NPRM. As previously explained in Section 2.3.4 and 4.1.1.2, we
believe that holding learning constant after 2029 is likely a conservative assumption, as we
would expect some level of continued learning beyond 2029 but there is uncertainty at this point
on what the appropriate level of learning would be.
4-23
-------
Table 4-23 Costs per Vehicle Projected through 2050 for the Final Standards (2018 dollars per vehicle)
Year
Car
Light Truck
Total
2023
$150
$485
$330
2024
$288
$732
$524
2025
$586
$909
$759
2026
$596
$1,356
$1,000
2027
$802
$1,469
$1,159
2028
$908
$1,462
$1,207
2029
$839
$1,381
$1,132
2030
$823
$1,377
$1,117
2031
$816
$1,374
$1,110
2032
$807
$1,366
$1,098
2033
$796
$1,353
$1,085
2034
$783
$1,341
$1,071
2035
$772
$1,324
$1,055
2036
$766
$1,307
$1,042
2037
$755
$1,286
$1,025
2038
$748
$1,279
$1,017
2039
$742
$1,273
$1,010
2040
$736
$1,269
$1,004
2041
$731
$1,266
$998
2042
$726
$1,261
$992
2043
$723
$1,260
$990
2044
$720
$1,247
$981
2045
$718
$1,246
$980
2046
$715
$1,244
$977
2047
$713
$1,243
$974
2048
$714
$1,242
$974
2049
$714
$1,237
$972
2050
$712
$1,233
$969
4.1.3.2 Alternatives
Car, truck, and fleet average vehicle costs for the Proposal and Alternative 2 minus 10
standards relative to the no action scenario (framework OEMs meeting the framework, non-
framework OEMs meeting SAFE) are summarized in Table 4-24, Table 4-25 and Table 4-26.
4-24
-------
Table 4-24: Car Average Cost per Vehicle for the Proposal and Alternative 2 minus 10 Standards Relative to
the No Action Scenario (2018 dollars)

Proposal Standards
Alternative 2 minus 10 Standards
Manufacturer
2023
2024
2025
2026
2023
2024
2025
2026
BMW
-$9
$90
$460
$416
$138
$535
$634
$563
Daimler
$53
$343
$302
$319
$241
$272
$243
$219
FCA
$250
$202
$153
$328
$270
$239
$193
$349
Ford
$21
$22
$231
$208
$361
$338
$532
$494
General Motors
$941
$848
$923
$888
$2,217
$2,022
$1,968
$1,856
Honda
$24
$26
$112
$104
$48
$297
$339
$300
Hyundai Kia-H
$24
$61
$689
$665
$483
$498
$633
$606
Hyundai Kia-K
$126
$229
$635
$600
$204
$252
$328
$315
JLR
$25
$256
$251
$246
$47
$468
$452
$436
Mazda
$5
$394
$471
$427
$326
$769
$820
$745
Mitsubishi
$0
$0
$682
$676
$0
$0
$916
$899
Nissan
$251
$308
$939
$865
$398
$486
$1,430
$1,330
Subaru
$18
$18
$17
$14
$18
$18
$18
$251
Tesla
$0
$0
$0
$0
$0
$0
$0
$0
Toyota
$20
$374
$530
$503
$37
$323
$479
$497
Volvo
$0
$0
$120
$118
$0
-$1
-$2
-$7
VWA
-$175
-$111
-$86
$143
$327
$615
$591
$877
TOTAL
$171
$257
$506
$493
$465
$561
$741
$724
Table 4-25: Light Truck Average Cost per Vehicle for the Proposal and Alternative 2 minus 10 Standards
Relative to the No Action Scenario (2018 dollars)

Proposal Standards
Alternative 2 minus 10 Standards
Manufacturer
2023
2024
2025
2026
2023
2024
2025
2026
BMW
-$154
-$152
-$131
-$130
$441
$410
$380
$351
Daimler
$6
$6
$697
$1,383
$2,514
$2,234
$2,069
$4,502
FCA
$1,490
$1,356
$1,266
$1,378
$1,714
$1,558
$1,409
$1,840
Ford
$38
$378
$361
$278
$224
$497
$460
$775
General Motors
$235
$338
$1,064
$1,188
$349
$656
$842
$1,322
Honda
-$11
$288
$370
$342
$128
$969
$985
$911
Hyundai Kia-H
$0
$0
$0
$1,999
$553
$542
$532
$5,400
Hyundai Kia-K
$851
$932
$860
$798
$2,096
$1,973
$1,808
$1,663
JLR
$26
$111
$821
$1,292
$429
$713
$1,565
$2,048
Mazda
$0
$0
$764
$1,315
$77
$76
$792
$2,089
Mitsubishi
$0
$0
$1,068
$1,031
$0
$0
$2,160
$2,028
Nissan
$475
$469
$451
$785
$699
$703
$677
$1,020
Subaru
$0
$9
$9
$8
$2
$85
$176
$175
Tesla
$0
$0
$0
$0
$0
$0
$0
$0
Toyota
$164
$661
$601
$628
$667
$1,127
$1,037
$1,692
Volvo
$472
$426
$392
$363
$871
$778
$704
$641
VWA
-$8
-$7
$121
$159
$986
$1,322
$1,566
$1,555
TOTAL
$364
$500
$679
$775
$689
$893
$938
$1,369
4-25
-------
Table 4-26: Fleet Average Cost per Vehicle for the Proposal and Alternative 2 minus 10 Standards Relative to
the No Action Scenario (2018 dollars)

Proposal Standards
Alternative 2 minus 10 Standards
Manufacturer
2023
2024
2025
2026
2023
2024
2025
2026
BMW
-$62
$2
$247
$220
$252
$491
$543
$484
Daimler
$29
$180
$484
$813
$1,347
$1,210
$1,108
$2,240
FCA
$1,297
$1,170
$1,085
$1,205
$1,489
$1,345
$1,212
$1,597
Ford
$33
$282
$325
$259
$260
$453
$478
$698
General Motors
$481
$519
$1,016
$1,085
$996
$1,136
$1,241
$1,516
Honda
$9
$134
$218
$202
$82
$580
$611
$555
Hyundai Kia-H
$21
$55
$623
$788
$488
$500
$621
$1,055
Hyundai Kia-K
$373
$464
$712
$669
$856
$835
$826
$766
JLR
$26
$120
$793
$1,238
$411
$703
$1,509
$1,965
Mazda
$1
$202
$607
$844
$201
$427
$799
$1,382
Mitsubishi
$0
$0
$882
$859
$0
$0
$1,562
$1,484
Nissan
$314
$353
$808
$846
$484
$547
$1,225
$1,249
Subaru
$5
$12
$14
$13
$7
$71
$139
$199
Tesla
$0
$0
$0
$0
$0
$0
$0
$0
Toyota
$82
$494
$561
$559
$308
$664
$715
$997
Volvo
$351
$316
$321
$300
$650
$580
$525
$476
VWA
-$78
-$52
$31
$154
$707
$1,021
$1,152
$1,268
TOTAL
$275
$386
$598
$644
$586
$740
$850
$1,070
4.1.4 Technology Penetration Rates
4.1.4.1 Final Rule
Many manufacturers have projected aggressive moves toward electrification in the coming
years, with several manufacturers projecting a complete transition to plug-in vehicles by 2030 or
2035, as previously described in section 0. Today's rule sets new standards through 2026,
however it is intended to begin a future transition toward electrification. Table 4-27 shows the
penetration rate of BEV+PHEV technology under the No Action scenario. Table 4-28, Table
Table 4-29, and Table 4-30 show the penetration rate of BEV+PHEV technology with the
remaining share being traditional ICE and/or advanced ICE technology under today's final
standards. Values shown reflect fleet penetration and are not increments from the SAFE
standards or other standards. The combined fleet technology penetrations for ICE vehicles are
shown in Table 4-31.
4-26
-------
Table 4-27 BEV+PHEV Penetration Rates under the No Action Scenario
Manufacturer
Car
Light Truck
Fleet
2023
2024
2025
2026
2023
2024
2025
2026
2023
2024
2025
2026
BMW
4%
8%
10%
17%
10%
10%
10%
10%
6%
9%
10%
14%
Daimler
12%
12%
13%
13%
8%
8%
8%
8%
10%
10%
10%
11%
FCA
17%
18%
19%
19%
1%
1%
1%
1%
3%
4%
4%
4%
Ford
13%
13%
14%
19%
1%
1%
3%
6%
5%
5%
6%
10%
General Motors
5%
5%
5%
6%
0%
1%
1%
1%
2%
2%
3%
3%
Honda
1%
1%
3%
7%
0%
6%
8%
8%
1%
3%
6%
8%
Hyundai Kia-H
8%
7%
7%
7%
0%
0%
0%
0%
7%
7%
7%
7%
Hyundai Kia-K
2%
1%
1%
2%
0%
0%
0%
0%
1%
1%
1%
1%
JLR
0%
0%
0%
0%
16%
16%
16%
16%
15%
15%
15%
15%
Mazda
7%
7%
7%
7%
0%
0%
0%
0%
3%
4%
4%
4%
Mitsubishi
3%
3%
3%
3%
0%
0%
0%
0%
2%
2%
2%
2%
Nissan
2%
2%
2%
2%
0%
0%
0%
0%
1%
1%
1%
1%
Subaru
0%
0%
0%
0%
1%
1%
1%
1%
0%
0%
0%
0%
Tesla
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
Toyota
2%
2%
3%
3%
0%
0%
0%
0%
1%
1%
2%
2%
Volvo
3%
3%
4%
11%
13%
13%
14%
14%
10%
10%
12%
14%
VWA
16%
17%
17%
17%
11%
11%
11%
11%
13%
13%
13%
14%
TOTAL
9%
9%
9%
10%
2%
2%
3%
3%
5%
5%
6%
7%
Table 4-28: Car BEV+PHEV Penetration Rates under the Final Standards
Manufacturer
2023
2024
2025
2026
BMW
4%
9%
22%
29%
Daimler
15%
18%
18%
19%
FCA
20%
22%
22%
22%
Ford
13%
13%
16%
21%
General Motors
11%
11%
11%
13%
Honda
2%
5%
8%
12%
Hyundai Kia-H
10%
10%
18%
18%
Hyundai Kia-K
3%
3%
8%
8%
JLR
0%
3%
3%
3%
Mazda
7%
13%
13%
13%
Mitsubishi
3%
3%
3%
3%
Nissan
3%
3%
17%
17%
Subaru
0%
0%
0%
3%
Tesla
100%
100%
100%
100%
Toyota
2%
6%
9%
9%
Volvo
3%
3%
4%
11%
VWA
16%
17%
17%
25%
TOTAL
10%
12%
16%
17%
4-27
-------
Table 4-29: Light Truck BEV+PHEV Penetration Rates under the Final Standards
Manufacturer
2023
2024
2025
2026
BMW
10%
10%
10%
10%
Daimler
8%
8%
21%
56%
FCA
13%
13%
13%
18%
Ford
1%
7%
8%
17%
General Motors
4%
8%
14%
18%
Honda
0%
13%
17%
17%
Hyundai Kia-H
0%
0%
0%
23%
Hyundai Kia-K
11%
11%
11%
11%
JLR
16%
16%
28%
35%
Mazda
0%
0%
0%
21%
Mitsubishi
0%
0%
16%
16%
Nissan
4%
5%
5%
9%
Subaru
1%
1%
1%
1%
Tesla
100%
100%
100%
100%
Toyota
1%
12%
12%
16%
Volvo
22%
22%
23%
23%
VWA
11%
12%
12%
18%
TOTAL
5%
9%
11%
17%
Table 4-30: Fleet BEV+PHEV Penetration Rates under the Final Standards
Manufacturer
2023
2024
2025
2026
BMW
6%
10%
18%
22%
Daimler
12%
14%
20%
36%
FCA
14%
15%
15%
18%
Ford
5%
9%
10%
18%
General Motors
6%
9%
13%
16%
Honda
1%
8%
12%
14%
Hyundai Kia-H
9%
9%
17%
19%
Hyundai Kia-K
6%
6%
9%
9%
JLR
15%
15%
26%
34%
Mazda
3%
7%
7%
17%
Mitsubishi
2%
2%
10%
10%
Nissan
3%
4%
14%
15%
Subaru
0%
0%
0%
1%
Tesla
100%
100%
100%
100%
Toyota
2%
9%
10%
12%
Volvo
17%
17%
18%
20%
VWA
13%
14%
14%
21%
TOTAL
7%
10%
14%
17%
4-28
-------
Table 4-31: Fleet ICE Technology Penetration Rates under the Final Standards
Technology
2023
2024
2025
2026
No-action
Final
No-action
Final
No-action
Final
No-action
Final
Gasoline Direct Injection
(without turbo, HCR, HEV,
etc.)
21%
20%
15%
12%
13%
8%
9%
5%
Cylinder deactivation
9%
8%
8%
7%
6%
5%
6%
6%
Turbocharging level 1
25%
23%
23%
22%
22%
18%
20%
15%
Turbocharging level 2
0%
0%
0%
0%
0%
0%
0%
0%
Cooled EGR
0%
0%
0%
0%
0%
0%
0%
0%
High compression ratio level 0
4%
4%
2%
2%
2%
2%
3%
2%
High compression ratio level 1
18%
21%
26%
30%
28%
33%
32%
36%
High compression ratio level 2
0%
0%
0%
0%
0%
0%
0%
0%
Mild hybrid
3%
3%
3%
4%
3%
4%
5%
5%
Strong hybrid P2
2%
3%
2%
3%
2%
3%
2%
6%
Strong hybrid Powersplit
3%
3%
3%
2%
3%
1%
2%
1%
PHEV
1%
1%
1%
1%
1%
1%
0%
1%
BEV
4%
7%
5%
10%
5%
13%
6%
17%
Mass reduction 0
21%
19%
20%
13%
17%
9%
13%
4%
Mass reduction 1
18%
16%
19%
20%
21%
22%
25%
24%
Mass reduction 2
20%
18%
19%
10%
19%
9%
18%
8%
Mass reduction 3
31%
37%
32%
45%
32%
47%
33%
49%
Mass reduction 4
10%
10%
11%
11%
11%
12%
11%
14%
Mass reduction 5
0%
0%
0%
0%
0%
0%
0%
0%
Mass reduction 6
0%
0%
0%
0%
0%
0%
0%
0%
Curb Weight reduction
(relative to MRO)
5.0%
5.2%
5.1%
5.7%
5.3%
5.9%
5.4%
6.3%
This final rule includes advanced technology multipliers. A recent working paper by
Gillingham (2021) uses a stylized model to examine the effects of EV multipliers on EV
adoption and conventional vehicle emission reductions.22 He finds that, under some conditions,
multipliers may reduce EV adoption and increase vehicle emissions; under other conditions, they
may increase EV adoption and decrease vehicle emissions. In particular, under the conditions of
low levels of EV market share and EV costs higher than those of conventional vehicles, EV
multiplier incentives are expected to increase EV penetration. Gillingham (2021) states that
tightening the standards in addition to allowing multipliers will "offset the standard-weakening
effect of the generous crediting." Gillingham acknowledges the stylized nature of his model and
suggests examining the effectiveness of advanced technology multipliers with more detailed
models in regulatory analyses, citing as an example the NHTSA CAFE model. As part of the
analysis for this proposed rule, EPA has estimated the benefits and costs of this rule with and
without the advanced technology multipliers in a memo to the docket. For reasons discussed in
Preamble Section II.B. 1, we are finalizing the limited use of multipliers to promote
commercialization of advanced technologies and to provide compliance flexibility.
To help shed light on the impact of advanced technology multipliers on the penetration rates
of BEVs and PHEVs, we conducted model runs without the multipliers. Those results along with
the results of the runs with multipliers are shown in Table 4-32. The results presented in this
table suggest that the advanced technology multipliers are not expected to have a large impact on
BEV and PHEV technology penetration.
4-29
-------
Table 4-32: Impact of Advanced Technology Multipliers on the Penetration of BEV and PHEV Technology

2023
2024
2025
2026
Final standards, with multipliers
7%
10%
14%
17%
Final standards, without multipliers
8%
11%
13%
17%
4.1.4.2 Alternatives
Table 4-33, Table 4-34, and Table 4-35 show the penetration rate of BEV+PHEV technology
with the remaining share being traditional ICE and/or advanced ICE technology for the Proposal
standards and the Alternative 2 minus 10 standards. Values shown reflect fleet penetration and
are not increments from the SAFE standards or other standards.
Table 4-33: Car BEV+PHEV Penetration Rates under the Proposal and Alternative 2 minus 10 Standards
Manufacturer
Proposal Standards
Alternative 2 minus 10 Standards
2023
2024
2025
2026
2023
2024
2025
2026
BMW
4%
9%
17%
23%
4%
13%
17%
24%
Daimler
13%
16%
16%
17%
15%
15%
15%
15%
FCA
20%
22%
22%
22%
20%
21%
22%
22%
Ford
13%
13%
16%
21%
13%
13%
16%
21%
General Motors
16%
16%
16%
17%
29%
30%
30%
32%
Honda
1%
1%
4%
8%
2%
5%
7%
11%
Hyundai Kia-H
8%
8%
16%
16%
14%
14%
16%
16%
Hyundai Kia-K
3%
3%
8%
8%
3%
3%
4%
4%
JLR
0%
0%
0%
0%
0%
0%
0%
0%
Mazda
7%
13%
13%
13%
11%
17%
17%
17%
Mitsubishi
3%
3%
3%
3%
4%
3%
3%
3%
Nissan
4%
5%
14%
14%
4%
4%
18%
18%
Subaru
0%
0%
0%
0%
0%
0%
0%
3%
Tesla
100%
100%
100%
100%
100%
100%
100%
100%
Toyota
2%
4%
7%
7%
3%
3%
5%
6%
Volvo
3%
3%
4%
11%
3%
3%
4%
11%
VWA
14%
15%
15%
19%
20%
23%
23%
28%
TOTAL
10%
11%
15%
16%
13%
14%
17%
19%
4-30
-------
Table 4-34: Light Truck BEV+PHEV Penetration Rates under the Proposal and Alternative 2 minus 10
Standards
Manufacturer
Proposal Standards
Alternative 2 minus 10 Standards
2023
2024
2025
2026
2023
2024
2025
2026
BMW
10%
10%
10%
10%
14%
14%
14%
14%
Daimler
8%
8%
21%
27%
33%
33%
33%
68%
FCA
11%
11%
11%
12%
13%
13%
13%
17%
Ford
1%
6%
7%
9%
1%
5%
6%
15%
General Motors
3%
7%
13%
13%
3%
5%
6%
11%
Honda
0%
8%
12%
12%
0%
13%
17%
17%
Hyundai Kia-H
0%
0%
0%
0%
0%
0%
0%
46%
Hyundai Kia-K
8%
8%
8%
8%
20%
20%
20%
20%
JLR
16%
16%
24%
31%
18%
19%
30%
38%
Mazda
0%
0%
0%
0%
0%
0%
0%
14%
Mitsubishi
0%
0%
4%
4%
0%
0%
16%
16%
Nissan
4%
4%
4%
7%
4%
4%
4%
7%
Subaru
1%
1%
1%
1%
1%
1%
1%
1%
Tesla
100%
100%
100%
100%
100%
100%
100%
100%
Toyota
1%
11%
11%
12%
4%
14%
14%
21%
Volvo
16%
16%
18%
18%
21%
21%
22%
22%
VWA
11%
11%
11%
11%
11%
15%
15%
15%
TOTAL
4%
8%
10%
11%
6%
9%
10%
16%
Table 4-35: Fleet BEV+PHEV Penetration Rates under the Proposal and Alternative 2 minus 10 Standards
Manufacturer
Proposal Standards
Alternative 2 minus 10 Standards
2023
2024
2025
2026
2023
2024
2025
2026
BMW
6%
10%
14%
19%
8%
13%
16%
20%
Daimler
10%
12%
19%
22%
24%
24%
24%
40%
FCA
12%
12%
13%
14%
14%
14%
14%
18%
Ford
5%
8%
10%
12%
4%
7%
9%
16%
General Motors
8%
10%
14%
15%
12%
14%
14%
18%
Honda
1%
4%
7%
10%
1%
8%
11%
14%
Hyundai Kia-H
7%
7%
15%
15%
13%
13%
15%
19%
Hyundai Kia-K
5%
5%
8%
8%
9%
9%
9%
9%
JLR
15%
15%
22%
30%
17%
18%
29%
36%
Mazda
3%
7%
7%
7%
5%
9%
9%
16%
Mitsubishi
2%
2%
4%
4%
2%
2%
10%
10%
Nissan
4%
4%
11%
12%
4%
4%
14%
15%
Subaru
0%
0%
0%
0%
0%
0%
0%
1%
Tesla
100%
100%
100%
100%
100%
100%
100%
100%
Toyota
2%
7%
9%
9%
3%
8%
9%
12%
Volvo
13%
13%
14%
16%
16%
16%
18%
19%
VWA
12%
13%
13%
14%
15%
18%
18%
21%
TOTAL
7%
9%
12%
13%
9%
12%
14%
17%
Note that Alternative 2 minus 10 has slightly higher BEV+PHEV penetration in the early
years but then lower BEV+PHEV penetration in the later years, although these differences are
very small and likely within the model's variability. This can be explained by, at least, three
important considerations. The first of these being the advanced technology multipliers in the
4-31
-------
final standards in both MYs 2023 and 2024 which Alternative 2 minus 10 does not have. Those
multipliers serve to hinder slightly the BEV+PHEV penetration due to their multiplicative effect.
The reverse is then true in the later years. Having introduced more BEVs and PHEVs in the early
years under Alternative 2 minus 10, slightly fewer are needed in later years to "make up" for the
slower pace of technology introduction in those earlier years where multipliers are provided.
This characterization of technology penetration, of course, ignores the impacts of multipliers on
costs where multipliers provide manufacturers with more flexibility in achieving compliance
which serves to reduce costs in the early years.
4.1.4.3 Fleet Mix
The version of CCEMS used by EPA makes use of a dynamic fleet share model that
estimates, separately, the shares of passenger cars and light trucks based on vehicle
characteristics, and then adjusts them so that the market shares sum to one. As such, fleet mix
can change depending on the standards within a given modeled scenario. Table 4-36 shows the
fleet mix projections for the final standards and each of the alternatives.
Table 4-36 Fleet Mix Projections for the Final Standards, Proposal and Alternative 2 minus 10
Model
Final Standards

Proposal
Alternative 2 minus 10
Year
Car
Light Truck
Car
Light Truck
Car
Light Truck
2020
44%
56%
44%
56%
44%
56%
2021
44%
56%
44%
56%
44%
56%
2022
46%
54%
46%
54%
46%
54%
2023
46%
54%
46%
54%
46%
54%
2024
47%
53%
47%
53%
47%
53%
2025
47%
53%
47%
53%
47%
53%
2026
47%
53%
48%
52%
47%
53%
The net benefits for a given set of standards depend in large part on how those standards
affect the fleet mix. In this rule, as discussed in Chapter 8.1.2 and Preamble Section VII.B,
CCEMS uses the dynamic fleet share modeling from DOE's National Energy Modeling System
(NEMS). EPA will continue to assess fleet mix questions in subsequent rulemakings, including
consumer and producer decisions between cars and light trucks, any possible "upsizing" effect of
the standards, and questions about how modeled fleet share results compare with observed trends
(e.g., current and future AEO estimates).
4.1.5 Sensitivities
We have conducted the following sensitivities:
•	AEO high oil price (AEO high)
•	AEO low oil price (AEO low)
•	Allow HCR2 in MY 2025 and later (Allow HCR2)
•	Battery costs higher
•	Battery costs lower (battery costs roughly 24 percent lower than the updated FRM
costs)
•	Sales demand elasticity of-0.15
•	Sales demand elasticity of-1.0
4-32
-------
•	Mass safety coefficients at the 5th percentile (Mass safety 5th pctile)
•	Mass safety coefficients at the 95th percentile (Mass safety 95th pctile)
•	No further application of mild or strong hybrid technology (no hybrids)
•	Perfect trading, which allows perfect trading of CO2 credits between manufacturers0
•	Rebound rate of -5 percent
•	Rebound rate of -15 percent
Each sensitivity is compared to its own no action scenario. In other words, the no action
standards were used but the no action scenario was run using the same set of sensitivity
parameters as used for the action scenario.
The high and low battery cost sensitivity cases were selected to evaluate the effect of
variations in battery cost. Figure 4-3 illustrates the range of costs using an example of a 60 kWh
battery and its direct manufacturing cost (DMC) under each sensitivity case. The solid line
depicts the costs applicable to the primary case, which as described in Sections 2.3.4 and 4.1.1.2,
reduced costs during the time frame of the rule by about 24 percent from the proposal, and
flattened learning beginning in MY 2029. For the high cost case, we used the battery costs from
the proposal, until MY 2035 when we merged them with the primary case. For the low cost case,
we reduced costs during the time frame of the rule by about 33 percent from the proposal, and
allowed learning to proceed at the rates defined by the battery learning curve that was used in the
proposal.
$200
$180
$160
jz $140
H $120
V> $100
$80
$80
$40
2020 2025 2030 2035 2040 2045 2050
Model year
Figure 4-3. Battery cost sensitivity cases in terms of $/kWh DMC for a representative 60 kWh battery
• High
¦	Primary
¦	Low
c To simulate perfect trading, the entire fleet is attributed to a single manufacturer, dubbed "Industry," in the market
input file.
4-33
-------
The high oil price and low oil price sensitivities use the fuel prices shown in Figure 4-4.
AEO 2021 Gasoline Prices
5
AEO 2021 Electricity Prices
45 		
0.14
1 3.5 "
S-:;
= 0.125 / ^—High Oil Price
^ Reference Case
n 1 •>
fU f ——
bo 3 X
co Reference Case
o 1 5
^ Low Oil Price
1
o o.iz
cm Low Oil Price
0.115
0.5
0

2020
2022
2024
2026
2028
2030
r>
5L 2032
0>
g. 2034
o>
-1 2036
-<
£ 2038
2040
2042
2044
2046
2048
2050
c
2020
2022
2024
2026
2028
2030
r>
SL 2032
0>
= 2034
o>
-1 2036
-<
£ 2038
2040
2042
2044
2046
2048
2050
Figure 4-4 Gasoline and Electricity Prices used in the Primary and Sensitivity Analyses
4.1.5.1 Compliance Costs per Vehicle and Technology Penetration
The per-vehicle compliance costs for the final rule and for each of the analyzed sensitivities
are shown in Table 4-37. The technology penetration rates for the final rule and for each of the
analyzed sensitivities are shown in Table 4-38 and Table 4-39. Note that the costs per vehicle
and the technology penetration rates for the mass safety and rebound sensitivities are identical to
those for the final standards since those sensitivities have no impact on compliance; therefore,
those results are not shown in the tables that follow.
Table 4-37: Costs per Vehicle for the Final Standards and Sensitivities relative to their No Action Scenarios
(2018 dollars)*
Model
Year
Final
AEO
high
AEO
low
Allow
HCR2
Battery
costs
higher
Battery
costs
lower
Demand
elasticity
-0.15'
Demand
elasticity
-1.0 '
No
hybrids
Perfect
trading
2021
$72
$65
$76
$72
$101
$40
$72
$72
$78
$6
2022
$185
$156
$211
$185
$223
$151
$185
$185
$185
$21
2023
$330
$297
$380
$331
$445
$288
$329
$328
$315
$147
2024
$524
$480
$563
$524
$760
$454
$522
$521
$510
$360
2025
$759
$753
$805
$760
$1,092
$708
$759
$758
$826
$772
2026
$1,000
$951
$1,040
$986
$1,398
$909
$1,000
$999
$1,023
$1,061
4-34
-------
Table 4-38: MY 2026 Technology Penetration Rates for the No-Action and Final Standards in the AEO High,
AEO Low, Allow HCR2 and No Hybrids Sensitivities

Final
AEO high
AEO low
Allow HCR2
No hybrids
Technology
No-
Action
Final
No-
Action
Final
No-
Action
Final
No-
Action
Final
No-
Action
Final
DEAC
6%
6%
7%
5%
6%
6%
6%
6%
8%
6%
TURBO1
20%
15%
18%
15%
20%
14%
20%
15%
19%
16%
HCR1
32%
36%
37%
37%
30%
35%
31%
32%
32%
35%
HCR2
0%
0%
0%
0%
0%
0%
1%
4%
0%
0%
AT8
26%
21%
25%
21%
27%
22%
26%
22%
26%
23%
AT8L2
7%
3%
4%
2%
7%
3%
7%
3%
8%
4%
AT8L3
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
AT9L2
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
AT10L2
22%
17%
22%
15%
23%
18%
22%
18%
24%
22%
AT10L3
2%
3%
2%
3%
2%
3%
2%
3%
2%
3%
SS12V
40%
31%
36%
28%
42%
32%
40%
32%
46%
42%
BISG
5%
5%
5%
5%
5%
5%
5%
5%
2%
1%
SHEVP2
2%
6%
2%
5%
2%
5%
2%
4%
0%
0%
SHEVPS
2%
1%
2%
1%
2%
1%
2%
1%
2%
1%
P2HCR1
2%
4%
2%
3%
2%
4%
2%
4%
0%
0%
PHEV
0%
1%
1%
1%
0%
1%
0%
1%
0%
1%
BEV
6%
17%
7%
17%
6%
17%
6%
17%
8%
19%
MRO
13%
4%
5%
4%
15%
5%
13%
4%
13%
4%
MR1
25%
24%
31%
24%
23%
22%
25%
25%
23%
22%
MR2
18%
8%
18%
7%
18%
8%
18%
8%
18%
8%
MR3
33%
49%
35%
51%
33%
52%
33%
50%
32%
50%
MR4
11%
14%
11%
12%
12%
14%
11%
12%
14%
16%
MR5
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
MR6
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
4-35
-------
Table 4-39 MY 2026 Technology Penetration Rates for the No-Action and Final Standards in the Battery
Costs Higher and Lower, Demand Elasticity of -0.15 and -1.0 and the Perfect Trading Sensitivities

Battery costs
higher
Battery costs
lower
Demand
elasticity -0.15
Demand
elasticity -1.0
Perfect trading
Technology
No-
Action
Final
No-
Action
Final
No-
Action
Final
No-
Action
Final
No-
Action
Final
DEAC
6%
5%
8%
6%
6%
6%
6%
6%
10%
10%
TURBO1
18%
13%
21%
16%
20%
15%
20%
15%
30%
23%
HCR1
31%
36%
32%
34%
32%
36%
32%
36%
25%
26%
HCR2
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
AT8
25%
12%
26%
23%
26%
21%
26%
21%
36%
29%
AT8L2
5%
3%
7%
4%
7%
3%
7%
3%
2%
1%
AT8L3
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
AT9L2
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
AT10L2
25%
17%
22%
17%
22%
17%
21%
17%
22%
18%
AT10L3
3%
7%
2%
3%
2%
3%
2%
3%
0%
1%
SS12V
41%
27%
40%
32%
40%
31%
40%
31%
46%
37%
BISG
4%
5%
6%
6%
5%
5%
5%
5%
5%
5%
SHEVP2
2%
10%
1%
3%
2%
6%
2%
6%
0%
3%
SHEVPS
3%
2%
2%
1%
2%
1%
2%
1%
2%
1%
P2HCR1
2%
6%
2%
3%
2%
4%
2%
4%
2%
3%
PHEV
2%
5%
0%
0%
0%
1%
0%
1%
0%
1%
BEV
5%
11%
7%
20%
6%
17%
6%
17%
7%
16%
MR0
13%
4%
13%
4%
13%
4%
13%
4%
12%
5%
MR1
20%
15%
26%
33%
25%
24%
25%
24%
27%
27%
MR2
16%
9%
18%
11%
18%
8%
18%
8%
17%
11%
MR3
33%
36%
33%
41%
33%
49%
33%
49%
32%
45%
MR4
18%
36%
10%
11%
11%
14%
11%
14%
12%
12%
MR5
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
MR6
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
4.1.5.2 How Battery Costs Impact Non-BEV Vehicle Costs in our Modeling
In Chapter 4.1.3, we noted that the reduced battery costs had the effect of not only reducing
average per vehicle costs, but that their lower costs combined with the higher BEV penetration
has the effect of lowering the technology needed on other vehicles. Likewise, higher battery
costs have the effect of reducing BEV penetration and increasing the technology (and costs) of
other vehicles. This is illustrated in Table 4-40 which compares the contribution to the average
cost per vehicle by vehicles having the various powertrains in MY 2026 for our Final standards
to the costs of those same standards using our higher battery cost estimates.
4-36
-------
Table 4-40 Cost Contributions Comparing our Primary Battery Costs to the Higher Battery Costs (2018
dollars per Vehicle)

Final
Final with Higher Battery Costs
Powertrain
Share of Vehicles in
Contribution to
Share of Vehicles
Contribution to

Powertrain Category
$/vehicle
in Powertrain
Category
$/vehicle
Conventional
36%
$296
34%
$334
12V Start-Stop
31%
$188
27%
$202
Mild HEV
5%
$47
5%
$49
Strong HEV
10%
$280
18%
$573
PHEV
1%
$11
5%
$259
BEV
17%
$1,052
11%
$1,033
Fuel Cell
0%
$0
0%
$0
Sum*
100%
$1,873
100%
$2,450
* Note that costs presented here are not marginal costs relative to a "no action" scenario so will not be reflected
elsewhere in this RIA.



Two things stand out in this table:
•	The lower BEV penetration when battery costs are higher (11 percent vs. 17 percent)
results in stronger HEV penetration (18 percent vs. 10 percent) and, due to the higher
battery costs, $293 greater contribution to the $/vehicle ($573 minus $280) in the
higher battery cost case. The same is true of PHEVs, with the results being $248
greater contribution to the $/vehicle in the higher battery cost case.
•	Non-electrified vehicle share actually decreases in the higher battery cost case, but
their contribution to $/vehicle increases. With fewer BEVs, Conventional (or ICE) and
start-stop vehicles must add more (and/or more costly) technology than they do in the
primary case ($334 plus $202 giving $536 in the higher battery cost case, vs. $484 in
the primary case).
Table 4-41 compares the contributions by vehicles having the various powertrains in MY
2026 for our Final standards to the costs of those same standards using our lower battery cost
estimates.
Table 4-41 Cost Contributions Comparing our Primary Battery Costs to the Lower Battery Costs (2018
dollars per Vehicle)

Final
Final with Lower Battery Costs
Powertrain
Share of Vehicles in
Contribution to
Share of Vehicles in
Contribution to

Powertrain Category
$/vehicle
Powertrain Category
$/vehicle
Conventional
36%
$296
36%
$281
12V Start-Stop
31%
$188
32%
$168
Mild HEV
5%
$47
6%
$44
Strong HEV
10%
$280
6%
$156
PHEV
1%
$11
0%
$2
BEV
17%
$1,052
20%
$1,002
Fuel Cell
0%
$0
0%
$0
Sum*
100%
$1,873
100%
$1,653
* Note that costs presented here are not marginal costs relative to a "no action" scenario so will not be reflected
elsewhere in this RIA.



4-37
-------
As with Table 4-40, the scenario with lower battery costs results, as expected, in more BEVs
and even lower contributions from strong HEV and PHEV technologies to the average per
vehicle cost driven in part by their lower battery costs but also their lower shares. The same
Conventional and start-stop trend is shown here too with slightly less (and/or less costly)
technology being added in the lower battery cost case.
These results support our statement that our primary battery costs, when compared to our
higher battery costs, which are more similar to those used in our proposal, have a dual impact on
the average per vehicle costs in that they result in higher BEV penetrations which then lowers the
additional technology costs of other vehicles. Similarly, with lower battery costs than in our
primary case, this impact becomes more pronounced.
4.2 Estimates of Fuel Economy Impacts
4.2.1 Final Rule
The estimated impacts on fuel economy associated with our No Action scenario and the final
standards are shown in Table 4-42. Importantly, these fuel economy values are based on the
standards and the model's estimated achieved levels, or rating, and therefore do not consider use
of AC leakage credits. The fuel economy values are estimated using the average CO2 content of
the gasoline used for compliance testing (8887 grams CChper gallon of certification gasoline).
Table 4-43 presents the fuel economy values assuming full use of AC leakage credits where we
have calculated the fuel economy again using the CO2 content of certification gasoline and
adding to the values shown in Table 4-42 the AC leakage credit. Because we expect full use of
the AC leakage credit, the values shown in Table 4-43 are considered to be more indicative of
the actual fuel economy values in compliance testing.
Perhaps of most interest are the estimated fuel economy impacts on-the-road, or the expected
"EPA label values." Those fuel economy values are shown in Table 4-44 where we have
multiplied the values shown in Table 4-43 by the anticipated "gap" of 0.8 to reflect the estimated
real-world values relative to the test cycle values.
4-38
-------
Table 4-42: Fuel Economy (MPG) Estimates based on the GHG Standards*
Regulatory
Class
MY
No Action Scenario
Final Rule
Standard
Rating
Standard
Rating
Car
2023
51
53
54
56
2024
52
55
56
60
2025
53
56
60
63
2026
54
57
67
66
Truck
2023
36
36
38
39
2024
36
37
40
42
2025
37
38
43
44
2026
38
39
48
50
Combined
2023
41
43
44
45
2024
42
44
46
49
2025
43
45
50
51
2026
45
46
55
57
* Calculated as 8887 divided by CChe. Note that the "rating" is the estimated compliance value and,
as such, includes possible under compliance due to use of banked credits and/or over compliance for
earning credits for future use.
Table 4-43: Fuel Economy (MPG) Estimates assuming full use of AC Leakage Credits*
Regulatory
Class
MY
No Action Scenario
Final Rule
Standard
Rating
Standard
Rating
Car
2023
47
49
50
51
2024
48
51
52
55
2025
49
52
55
58
2026
50
53
61
60
Truck
2023
33
34
35
36
2024
34
35
37
39
2025
35
36
40
40
2026
36
37
44
46
Combined
2023
39
40
41
42
2024
39
41
43
45
2025
40
42
46
47
2026
41
43
50
52
* Calculated as 8887 divided by (C02e + AC Leakage Credit). Note that the "rating" is the
estimated compliance value and, as such, includes possible under compliance due to use of banked
credits and/or over compliance for earning credits for future use.
4-39
-------
Table 4-44: Fuel Economy (MPG) Estimated "Label Value"*
Regulatory
Class
MY
No Action Scenario
Final Rule
Standard
Rating
Standard
Rating
Car
2023
38
39
40
41
2024
38
41
42
44
2025
39
41
44
46
2026
40
42
49
48
Truck
2023
27
27
28
29
2024
27
28
30
31
2025
28
29
32
32
2026
28
29
35
36
Combined
2023
31
32
33
33
2024
32
33
34
36
2025
32
34
37
38
2026
33
34
40
41
* Calculated as 8887 divided by (CChe + AC Leakage Credit) then multiplied by 0.8. Note that the
"rating" is the estimated compliance value and, as such, includes possible under compliance due
to use of banked credits and/or over compliance for earning credits for future use.
4.2.2 Alternatives
Here we present the analogous series of tables presented in Chapter 4.2.1 but for each of the
alternatives (i.e., the Proposal and Alternative).
Table 4-45: Fuel Economy (MPG) Estimates based on the GHG Standards for the Proposal Standards*
Regulatory
Class
MY
No Action Scenario
Proposal
Standard
Rating
Standard
Rating
Car
2023
51
53
54
57
2024
52
55
56
61
2025
53
56
59
62
2026
54
57
62
64
Truck
2023
36
36
38
38
2024
36
37
40
41
2025
37
38
42
42
2026
38
39
44
45
Combined
2023
41
43
44
45
2024
42
44
46
49
2025
43
45
49
50
2026
45
46
51
52
* Calculated as 8887 divided by CC>2e. Note that the "rating" is the estimated compliance value
and, as such, includes possible under compliance due to use of banked credits and/or over
compliance for earning credits for future use.
4-40
-------
Table 4-46: Fuel Economy (MPG) Estimates Assuming Full Use of AC Leakage Credits for the Proposal
Standards*
Regulatory Class
MY
No Action Scenario
Proposal
Standard
Rating
Standard
Rating
Car
2023
47
49
50
52
2024
48
51
52
56
2025
49
52
54
57
2026
50
53
57
59
Truck
2023
33
34
35
36
2024
34
35
37
38
2025
35
36
39
39
2026
36
37
41
41
Combined
2023
39
40
41
42
2024
39
41
43
45
2025
40
42
45
46
2026
41
43
47
48
* Calculated as 8887 divided by (CChe + AC Leakage Credit). Note that the "rating" is the
estimated compliance value and, as such, includes possible under compliance due to use of
banked credits and/or over compliance for earning credits for future use.
Table 4-47: Fuel Economy (MPG) Estimated "Label Value" Under the Proposal Standards*
Regulatory Class
MY
No Action Scenario
Proposal
Standard
Rating
Standard
Rating
Car
2023
38
39
40
42
2024
38
41
42
44
2025
39
41
43
46
2026
40
42
45
47
Truck
2023
27
27
28
29
2024
27
28
30
31
2025
28
29
31
31
2026
28
29
33
33
Combined
2023
31
32
33
33
2024
32
33
34
36
2025
32
34
36
37
2026
33
34
38
39
* Calculated as 8887 divided by (C02e+ AC Leakage Credit) then multiplied by 0.8. Note
that the "rating" is the estimated compliance value and, as such, includes possible under
compliance due to use of banked credits and/or over compliance for earning credits for
future use.
4-41
-------
Table 4-48: Fuel Economy (MPG) Estimates Based on the GHG Standards of the Alternative 2 minus 10
Standards*
Regulatory Class
MY
No Action Scenario
Alternative 2 minus 10
Standard
Rating
Standard
Rating
Car
2023
51
53
54
57
2024
52
55
57
61
2025
53
56
60
65
2026
54
57
67
68
Truck
2023
36
36
39
39
2024
36
37
41
42
2025
37
38
43
44
2026
38
39
48
50
Combined
2023
41
43
45
45
2024
42
44
47
49
2025
43
45
50
52
2026
45
46
55
57
* Calculated as 8887 divided by CC>2e. Note that the "rating" is the estimated compliance
value and, as such, includes possible under compliance due to use of banked credits and/or
over compliance for earning credits for future use.
Table 4-49: Fuel Economy (MPG) Estimates Assuming Full Use of AC Leakage Credits in the Alternative 2
minus 10 Standards*
Regulatory Class
MY
No Action Scenario
Alternative 2 minus 10
Standard
Rating
Standard
Rating
Car
2023
47
49
50
52
2024
48
51
52
56
2025
49
52
55
59
2026
50
53
61
62
Truck
2023
33
34
36
36
2024
34
35
38
39
2025
35
36
40
40
2026
36
37
44
45
Combined
2023
39
40
42
42
2024
39
41
44
45
2025
40
42
46
47
2026
41
43
50
52
* Calculated as 8887 divided by (CChe + AC Leakage Credit). Note that the "rating" is the
estimated compliance value and, as such, includes possible under compliance due to use of
banked credits and/or over compliance for earning credits for future use.
4-42
-------
Table 4-50: Fuel Economy (MPG) Estimated "Label Value" Under the Alternative 2 minus 10 Standards*
Regulatory Class
MY
No Action Scenario
Alternative 2 minus 10


Standard
Rating
Standard
Rating
Car
2023
38
39
40
42

2024
38
41
42
45

2025
39
41
44
47

2026
40
42
49
49
Truck
2023
27
27
29
29

2024
27
28
30
31

2025
28
29
32
32

2026
28
29
35
36
Combined
2023
31
32
33
34

2024
32
33
35
36

2025
32
34
36
38

2026
33
34
40
41
* Calculated as 8887 divided by (C02e+ AC Leakage Credit) then multiplied by 0.8. Note
that the "rating" is the estimated compliance value and, as such, includes possible under
compliance due to use of banked credits and/or over compliance for earning credits for
future use.





4-43
-------
References for Chapter 4
1	75 FR 25324.
2	77 FR 62624.
3	EPA-420-D-16-900, July 2016.
4	EPA-420-R-16-020, November 2016.
5	EPA-420-R-17-001, January 2017.
6	85 FR 24218.
7	See 86 FR 49602 and CAFE Model Documentation, August 2021.
8	California Air Resources Board. Framework Agreements on Clean Cars. August 17, 2020. Last accessed on the
Internet on 5/25/2021 at the following URL: https://ww2.arb.ca.gov/sites/default/files/2020-08/clean-car-
framework-documents-all-bmw-ford-honda-volvo-vw.pdf
9	85 FR 24647.
10	Science Advisory Board (SAB) Consideration of the Scientific and Technical Basis of the EPA's Proposed Rule
titled The Safer Affordable Fuel-Efficient (SAFE) Vehicles Rule for Model Years 2021-2026 Passenger Cars and
Light Trucks, February 27, 2020.
11	EPA CCEMS Post-Processing Tool, available in the docket and at
https://github.com/USEPA/EPA_CCEMS_PostProcessingTool
12	85 FR 24174.
13	EPA CCEMS Post-Processing Tool, available in the docket and at
https://github.com/USEPA/EPA_CCEMS_PostProcessingTool.
14	National Academies of Sciences, Engineering, and Medicine 2021. "Assessment of Technologies for Improving
Light-Duty Vehicle Fuel Economy 2025-2035". Washington, DC: The National Academies Press.
https://doi.org/10.17226/26092
15	Bloomberg New Energy Finance, "Battery Pack Prices Cited Below $100/kWh for the First Time in 2020, While
Market Average Sits at $137/kWh," accessed on October 30, 2021 at https://about.bnef.com/blog/battery-pack-
prices-cited-below-100-kwh-for-the-first-time-in-2020-while-market-average-sits-at-137-kwh/
16	National Academies of Sciences, Engineering, and Medicine 2021. "Assessment of Technologies for Improving
Light-Duty Vehicle Fuel Economy 2025-2035". Washington, DC: The National Academies Press.
https://doi.org/10.17226/26092, at p. 5-139.
17	Massachusetts Institute of Technology, "Insights into Future Mobility," MIT Energy Initiative (2019).
1 R
Hsieh, I-Yun Lisa et al., "Learning only buys you so much: Practical limits on battery price reduction." Applied
Energy, 239 (April 2019): 218-224.
19	86 FR 49602.
20	86 FR 49602.
21	EPA CCEMS Post-Processing Tool, available in the docket and at
https://github.com/USEPA/EPA_CCEMS_PostProcessingTool.
99
Gillingham, K. (2021). "Designing Fuel-Economy Standards in Light of Electric Vehicles." NBER working paper
#29067.
4-44
-------
Chapter 5: Projected Impacts on Emissions, Fuel Consumption, and Safety
This chapter documents EPA's analysis of the emission, fuel consumption and safety impacts
of the emission standards for light-duty vehicles. Light-duty vehicles include passenger vehicles
such as cars, sport utility vehicles, vans, and pickup trucks. Such vehicles are used for both
commercial and personal uses and are significant contributors to the total United States (U.S.)
GHG emission inventory.
5.1 Projected Emissions Impacts
5.1.1 Greenhouse Gas Emissions
5.1.1.1 Final Rule
EPA estimated the GHG emissions impacts associated with the Final standards, including
impacts on tailpipe emissions from light-duty cars and trucks and the upstream emissions
associated with the fuels used to power those vehicles (both at the refinery and the electricity
generating unit). The tailpipe emissions of CO2 are estimated internal to the model based on the
policy scenario(s) being run (as controlled via the scenarios input file) and the projected
compliance pathway which impacts the projected technology mix. The tailpipe emissions of CH4
and N2O make use of vehicle emission factors estimated by EPA's MOVES model.21'1'2 The
upstream emissions are then calculated using emission factors applied to the gallons of liquid
fuels projected to be consumed and the kilowatt hours of electricity projected to be consumed.
The upstream emission factors used in this final rule the modeling have been updated since
EPA's proposed rule. The updated upstream emission factors are consistent with those used in
the recent NHTSA CAFE proposal and were generated using the DOE/Argonne GREET model.3
In this final rule we have not attempted to change past practices by projecting the final
destinations of BEV's and PHEV's to estimate regional emission inventory impacts, however, we
are performing additional research to potentially add that capability in future rulemakings. See
Figure 5-1 through Figure 5-3 for a comparison of how our upstream emission factors have
changed since the NPRM.
a EPA used identical CH4 and N20 vehicle tailpipe emission factors to those used by NHTSA in their August
CAFE NPRM. See 86 FR 49602.
5-1
-------


C02; EGU


C02; Refinery


140000
120000


140000
120000



-



c
100000
80000
	
c
100000
80000



NPRM



I
60000
40000
20000
Q
1M r r\ 1VI
I
60000
40000
20000
Q

IMTIMVI
E
nj
FRM
E
nj

FRM
Oil

Oil









2020 2025 2030 2035 2040 2045 2050


2020 2025 2030 2035 2040 2045 2050

Figure 5-1 Electricity Generating Unit (EGU) and Refinery Emission Factors for CO2
CH4; Refinery
iNPRM
NPRM
Figure 5-2 Electricity Generating Unit (EGU) and Refinery Emission Factors for CELt
N20; Refinery
'NPRM
NPRM
Figure 5-3 Electricity Generating Unit (EGU) and Refinery Emission Factors for N2O
5-2
-------
Table 5-1: Impacts on GHG Emissions under the Final Standards Relative to the No Action Scenario
Year
CO2 Upstream
(MMT)
CH4 Upstream
(metric tons)
N20 Upstream
(metric tons)
CO2 Tailpipe
(MMT)
CH4 Tailpipe
(metric tons)
N20 Tailpipe
(metric tons)
2023
0
-5,149
-120
-5
-11
-26
2024
1
-10,097
-233
-11
-23
-60
2025
1
-17,342
-402
-18
-43
-112
2026
2
-27,311
-636
-29
-71
-182
2027
2
-39,613
-919
-41
-103
-256
2028
2
-52,768
-1,221
-54
-145
-337
2029
3
-64,891
-1,499
-66
-192
-416
2030
3
-76,665
-1,769
-77
-244
-494
2031
3
-87,836
-2,024
-88
-292
-568
2032
4
-98,675
-2,271
-99
-342
-641
2033
4
-108,880
-2,504
-109
-392
-710
2034
4
-118,279
-2,722
-118
-440
-775
2035
4
-126,826
-2,921
-127
-571
-835
2036
5
-134,419
-3,099
-134
-619
-890
2037
5
-140,938
-3,253
-141
-662
-941
2038
5
-146,592
-3,387
-146
-701
-984
2039
5
-151,747
-3,506
-151
-735
-1,023
2040
5
-156,120
-3,607
-155
-764
-1,055
2041
5
-159,800
-3,691
-159
-787
-1,082
2042
5
-162,773
-3,759
-161
-807
-1,104
2043
4
-165,254
-3,815
-163
-823
-1,122
2044
4
-167,458
-3,863
-165
-836
-1,136
2045
4
-169,301
-3,902
-166
-846
-1,147
2046
3
-170,811
-3,933
-167
-854
-1,157
2047
3
-172,003
-3,957
-167
-861
-1,165
2048
3
-173,079
-3,978
-167
-866
-1,172
2049
2
-175,319
-3,992
-168
-869
-1,177
2050
2
-177,519
-4,005
-168
-872
-1,181
Sum
93
-3,257,463
-74,989
-3,219
-14,771
-21,746
5-3
-------
5.1.1.2 Alternatives
EPA estimated the GHG emissions impacts associated with the Proposal and Alternative 2
minus 10 standards, including impacts on tailpipe emissions from light-duty cars and trucks and
the upstream emissions associated with the fuels used to power those vehicles (both at the
refinery and the electricity generating unit). The tailpipe emissions of CO2 are estimated internal
to the model based on the policy scenario(s) being run (as controlled via the scenarios input file)
and the projected compliance pathway which impacts the projected technology mix. The tailpipe
emissions of CH4 and N2O make use of vehicle emission factors estimated by EPA's MOVES
model.b'4'5 The upstream emissions are then calculated using emission factors applied to the
gallons of liquid fuels projected to be consumed and the kilowatt hours of electricity projected to
be consumed. The upstream emission factors used in this final rule have been updated since
EPA's proposed rule. The updated upstream emission factors are identical to those used in the
recent NHTSA CAFE proposal and were generated using the DOE/Argonne GREET model.6
b EPA used identical CH4 and N20 vehicle tailpipe emission factors to those used by NHTSA in their August
CAFENPRM. See 86 FR 49602.
5-4
-------
Table 5-2: Impacts on GHG Emissions under the Proposal Standards Relative to the No Action Scenario
Year
CO2 Upstream
(MMT)
CH4 Upstream
(metric tons)
N20 Upstream
(metric tons)
CO2 Tailpipe
(MMT)
CH4 Tailpipe
(metric tons)
N20 Tailpipe
(metric tons)
2023
0
-3,350
-83
-4
-11
-24
2024
1
-6,787
-163
-8
-22
-51
2025
2
-12,403
-294
-14
-39
-94
2026
2
-18,942
-444
-21
-55
-134
2027
2
-27,167
-629
-29
-80
-181
2028
2
-35,787
-823
-37
-107
-229
2029
2
-43,622
-999
-44
-138
-276
2030
2
-51,167
-1,169
-52
-170
-321
2031
2
-58,239
-1,327
-59
-200
-362
2032
2
-65,114
-1,481
-65
-230
-405
2033
2
-71,682
-1,629
-72
-262
-446
2034
3
-77,674
-1,766
-77
-291
-484
2035
3
-83,132
-1,891
-83
-374
-519
2036
3
-87,868
-2,001
-87
-402
-551
2037
3
-91,701
-2,091
-91
-426
-578
2038
3
-95,015
-2,168
-94
-447
-601
2039
3
-97,966
-2,235
-97
-465
-621
2040
3
-100,475
-2,293
-99
-481
-639
2041
3
-102,563
-2,341
-101
-494
-654
2042
3
-104,143
-2,376
-103
-505
-665
2043
3
-105,813
-2,415
-104
-516
-678
2044
2
-107,247
-2,448
-105
-526
-689
2045
2
-108,374
-2,474
-106
-533
-697
2046
2
-109,333
-2,496
-106
-540
-704
2047
2
-110,095
-2,512
-107
-545
-710
2048
2
-110,800
-2,528
-107
-550
-716
2049
1
-112,412
-2,543
-107
-555
-723
2050
1
-114,017
-2,558
-108
-560
-729
Sum
59
-2,112,887
-48,177
-2,086
-9,525
-13,484
5-5
-------
Table 5-3: Impacts on GHG Emissions under the Alternative 2 minus 10 Standards Relative to the No Action
Scenario
Year
CO2 Upstream
(MMT)
CH4 Upstream
(metric tons)
N20 Upstream
(metric tons)
CO2 Tailpipe
(MMT)
CH4 Tailpipe
(metric tons)
N20 Tailpipe
(metric tons)
2023
0
-7,751
-180
-8
-19
-40
2024
1
-14,238
-328
-15
-34
-80
2025
1
-22,073
-506
-23
-54
-130
2026
2
-32,434
-748
-33
-85
-202
2027
2
-44,564
-1,027
-45
-121
-277
2028
2
-57,775
-1,332
-58
-169
-364
2029
3
-69,705
-1,608
-70
-219
-447
2030
3
-81,328
-1,878
-81
-276
-532
2031
3
-92,290
-2,131
-92
-328
-610
2032
4
-102,768
-2,373
-103
-382
-686
2033
4
-112,594
-2,599
-112
-434
-756
2034
4
-121,539
-2,808
-121
-484
-822
2035
5
-129,681
-2,999
-129
-623
-882
2036
5
-136,890
-3,170
-137
-672
-938
2037
5
-142,869
-3,312
-143
-714
-986
2038
5
-148,053
-3,436
-148
-752
-1,028
2039
5
-152,748
-3,545
-152
-784
-1,065
2040
5
-156,641
-3,637
-156
-813
-1,096
2041
5
-159,903
-3,714
-159
-835
-1,121
2042
5
-162,696
-3,780
-161
-856
-1,145
2043
5
-164,960
-3,833
-163
-873
-1,164
2044
5
-167,076
-3,882
-165
-888
-1,181
2045
4
-168,784
-3,919
-166
-899
-1,194
2046
4
-170,221
-3,951
-167
-908
-1,206
2047
4
-171,604
-3,982
-168
-918
-1,220
2048
3
-172,872
-4,010
-168
-928
-1,234
2049
3
-175,336
-4,031
-169
-936
-1,246
2050
2
-177,862
-4,052
-169
-943
-1,257
Sum
99
-3,317,253
-76,770
-3,282
-15,948
-22,910
The cumulative CO2 emission reductions relative to the No Action scenario are shown in Figure
5-4.
5-6
-------
3,500
| 3,000
Alternative 2 minus 10
2,500
Final Standards
---2012 FRM
2,000
Proposed Standards
1,500
1,000
500
0
2020 2023 2026 2029 2032 2035 2038 2041 2044 2047 2050
Calendar Year
Figure 5-4 Cumulative CO2 Reductions relative to the No Action scenario (Million Metric Tons CO2)
5.1.2 Non-Greenhouse Gas Emissions
5.1.2.1 Final Rule
The model runs that EPA conducted estimated the inventories under the Final standards of
non-GHG air pollutants resulting from tailpipe emissions from light-duty cars and trucks, and the
upstream emissions associated with the fuels used to power those vehicles (both at the refinery
and the electricity generating unit).0 The tailpipe emissions of PM2.5, NOx, VOCs, CO and SO2
are estimated using emission factors from EPA's MOVES model. The tailpipe emission factors
used are identical to those used in NHTSA's recent NPRM.7 The upstream emissions are then
calculated using emission factors applied to the gallons of liquid fuels projected to be consumed
and the kilowatt hours of electricity projected to be consumed. The upstream emission factors
used in this final rule modeling have been updated since EPA's proposed rule. The updated
upstream emission factors are identical to those used in the recent NHTSA CAFE proposal and
were generated using the DOE/Argonne GREET model.8 See Figure 5-4 through Figure 5-7 for a
comparison of how some of our upstream non-GHG emission factors have changed since the
NPRM. See Figure 5-8 for how some of our tailpipe emission factors have changed since the
NPRM.
c Section V.C of the preamble includes more information on why we were not able to perform air quality modeling
for the non-GHG impacts of this rulemaking. We are considering how to project air quality impacts from changes in
non-GHG emissions in future LD GHG rulemakings.
5-7
-------
120
100
80
60
40
20
0
10
9
8
7
6
5
4
3
2
1
0
180
160
140
120
100
80
60
40
20
0
NOx; EGU
NOx; Refinery
-FRM
-NPRM
£ 80
c
0
1	60
| 40
ab
20
0
2020 2025 2030 2035 2040 2045 2050
2020 2025 2030 2035 2040 2045 2050
Figure 5-5 Electricity Generating Unit (EGU) and Refinery Emission Factors for NOx
PM2.5; EGU
PM2.5; Refinery
-FRM
-NPRM
2020 2025 2030 2035 2040 2045 2050
10
9
8
7
6
5
' 4
, 3
2
1
0
2020 2025 2030 2035 2040 2045 2050
Figure 5-6 Electricity Generating Unit (EGU) and Refinery Emission Factors for PM2.5
SOx; EGU
-NPRM
-FRM
2020 2025 2030 2035 2040 2045 2050

180

160

140


1—
CD
120
c

o
100
I
80
£

TO
60
Qfl


40

20

0
SOx; Refinery
2020 2025 2030 2035 2040 2045 2050
Figure 5-7 Electricity Generating Unit (EGU) and Refinery Emission Factors for SOx
5-8
-------
VOC; EGU
30
28
26
£ 24
CO
c 22
o
120
>18
g, 16
14
12
10
-NPRM
¦FRM
2020 2025 2030 2035 2040 2045 2050
VOC; Refinery
NPRM
Figure 5-8 Electricity Generating Unit (EGU) and Refinery Emission Factors for VOC
NOx, Tailpipe, MY2023
oj o.l
' NPRM, Truck
• NPRM, Car
' FRM, Truck
¦FRM, Car
PM2.5, Tailpipe, MY2023
= 0.006
0.004
0.002
' NPRM, Truck
¦	NPRM, Car
' FRM, Truck
¦	FRM, Car
0 3 6 9 12 15 18 21 24 27 30 33 36 39
0 3 6 9 12 15 18 21 24 27 30 33 36 39
NOx, Tailpipe, MY2026
v 0.1
0.008
> NPRM, Truck =| a006
¦NPRM, Car	*E
2 0.004
- FRM, Truck	00
¦FRM, Car	0,002
PM2.5, Tailpipe, MY2026
i NPRM,Truck
¦	NPRM, Car
' FRM, Truck
¦	FRM, Car
0 3 6 9 12 15 18 21 24 27 30 33 36 39
0 3 6 9 12 15 18 21 24 27 30 33 36 39
0.16
0.14
0.12
^ 0.1
"E
-£• 0.08
ro
ab 0.06
0.04
0.02
NOx, Tailpipe, MY2035
' NPRM, Truck
¦	NPRM, Car
' FRM, Truck
¦	FRM, Car
PM2.5, Tailpipe, MY2035
— 0.006
J.004
- NPRM, Truck
¦	NPRM, Car
• FRM, Truck
¦	FRM, Car
0 3 6 9 12 15 18 21 24 27 30 33 36 39
0 3 6 9 12 15 18 21 24 27 30 33 36 39
Figure 5-9 Tailpipe Emission Factors for MYs 2023,2026 and 2035 for NOx and PM2.5
5-9
-------
Table 5-4: Impacts on Upstream Non-GHG Emissions Under the Final Standards Relative to the No Action
Scenario

PM2.5 (US tons)
NOx (US tons)
SO2 (US tons)
VOC (US tons)
CO (US tons)
Year
EGU
Refinery
EGU
Refinery
EGU
Refinery
EGU
Refinery
EGU
Refinery
2023
111
-110
1,320
-1,226
1,154
-558
197
-1,941
699
-688
2024
244
-222
2,898
-2,471
2,512
-1,118
437
-3,899
1,551
-1,392
2025
417
-380
4,957
-4,231
4,260
-1,911
756
-6,713
2,681
-2,391
2026
640
-595
7,601
-6,607
6,473
-2,984
1,174
-10,560
4,158
-3,745
2027
857
-842
10,172
-9,329
8,577
-4,214
1,592
-15,010
5,632
-5,302
2028
1,067
-1,099
12,667
-12,161
10,565
-5,494
2,011
-19,700
7,105
-6,930
2029
1,291
-1,344
15,275
-14,850
12,836
-6,731
2,425
-24,132
8,571
-8,475
2030
1,506
-1,581
17,773
-17,440
15,045
-7,930
2,821
-28,421
9,976
-9,968
2031
1,704
-1,802
20,057
-19,858
17,106
-9,057
3,183
-32,456
11,262
-11,368
2032
1,898
-2,018
22,283
-22,197
19,147
-10,154
3,536
-36,385
12,517
-12,729
2033
2,078
-2,219
24,324
-24,373
21,060
-11,181
3,859
-40,068
13,669
-14,000
2034
2,243
-2,408
26,254
-26,430
22,645
-12,139
4,187
-43,508
14,818
-15,196
2035
2,389
-2,579
27,964
-28,286
24,029
-13,006
4,483
-46,623
15,853
-16,278
2036
2,521
-2,732
29,497
-29,940
25,249
-13,781
4,753
-49,415
16,797
-17,247
2037
2,636
-2,864
30,849
-31,373
26,304
-14,456
4,997
-51,846
17,646
-18,089
2038
2,735
-2,979
31,996
-32,607
27,175
-15,040
5,210
-53,952
18,384
-18,819
2039
2,806
-3,077
32,826
-33,659
27,772
-15,529
5,368
-55,763
18,930
-19,443
2040
2,862
-3,159
33,480
-34,535
28,215
-15,938
5,498
-57,286
19,380
-19,966
2041
2,900
-3,226
33,932
-35,240
28,481
-16,267
5,596
-58,526
19,716
-20,391
2042
2,924
-3,277
34,212
-35,780
28,598
-16,520
5,667
-59,496
19,955
-20,721
2043
2,939
-3,318
34,384
-36,211
28,621
-16,722
5,721
-60,285
20,134
-20,989
2044
2,933
-3,349
34,312
-36,539
28,528
-16,869
5,719
-60,881
20,122
-21,179
2045
2,921
-3,372
34,165
-36,788
28,371
-16,979
5,704
-61,342
20,067
-21,323
2046
2,905
-3,389
33,977
-36,973
28,180
-17,058
5,682
-61,694
19,988
-21,430
2047
2,883
-3,399
33,714
-37,083
27,927
-17,103
5,648
-61,923
19,866
-21,495
2048
2,860
-3,407
33,436
-37,170
27,660
-17,137
5,612
-62,111
19,734
-21,545
2049
2,851
-3,431
33,350
-37,475
27,512
-17,308
5,606
-62,238
19,706
-21,633
2050
2,841
-3,454
33,249
-37,769
27,351
-17,473
5,597
-62,347
19,669
-21,713
5-10
-------
Table 5-5: Estimated Non-GHG Emission Impacts of the Final Standards Relative to the No Action Scenario

Upstream Emissions (US tons)
Tailpipe Emissions (US tons)
Year
pm25
NOx
S02
voc
CO
pm25
NOx
S02
VOC
CO
2023
1
94
596
-1,744
12
7
717
-37
1,003
6,505
2024
22
427
1,394
-3,462
159
9
1,173
-77
1,693
10,048
2025
37
726
2,349
-5,957
290
8
1,645
-133
2,424
13,248
2026
45
994
3,490
-9,386
413
4
2,090
-208
3,149
15,356
2027
15
843
4,363
-13,418
331
-4
2,399
-295
3,702
15,150
2028
-32
505
5,072
-17,689
174
-21
2,383
-386
3,820
9,475
2029
-53
425
6,105
-21,707
96
-46
2,108
-471
3,566
-474
2030
-75
333
7,115
-25,601
8
-77
1,588
-554
2,962
-14,786
2031
-99
199
8,049
-29,273
-106
-106
1,167
-633
2,469
-27,521
2032
-120
85
8,994
-32,849
-212
-137
699
-709
1,896
-41,484
2033
-141
-49
9,878
-36,209
-331
-168
228
-780
1,287
-55,715
2034
-165
-177
10,506
-39,321
-377
-199
-241
-846
666
-70,103
2035
-190
-322
11,023
-42,140
-425
-287
-1,250
-906
-2,905
-92,848
2036
-211
-443
11,468
-44,661
-449
-321
-1,693
-959
-3,647
-106,860
2037
-228
-524
11,848
-46,849
-444
-353
-2,079
-1,006
-4,323
-119,740
2038
-244
-610
12,135
-48,742
-435
-383
-2,419
-1,046
-4,946
-131,691
2039
-271
-833
12,243
-50,395
-512
-409
-2,698
-1,081
-5,495
-142,121
2040
-297
-1,055
12,277
-51,788
-586
-434
-2,943
-1,110
-5,993
-151,549
2041
-325
-1,308
12,214
-52,930
-674
-455
-3,138
-1,134
-6,422
-159,628
2042
-353
-1,568
12,078
-53,829
-766
-473
-3,290
-1,153
-6,784
-166,420
2043
-379
-1,827
11,899
-54,564
-855
-490
-3,416
-1,168
-7,117
-172,314
2044
-415
-2,227
11,659
-55,162
-1,057
-503
-3,508
-1,178
-7,402
-177,017
2045
-451
-2,624
11,392
-55,638
-1,256
-514
-3,575
-1,185
-7,660
-180,783
2046
-483
-2,995
11,122
-56,012
-1,442
-523
-3,633
-1,191
-7,914
-184,085
2047
-516
-3,368
10,823
-56,274
-1,629
-531
-3,675
-1,194
-8,135
-186,783
2048
-548
-3,734
10,523
-56,499
-1,811
-538
-3,708
-1,196
-8,332
-189,005
2049
-580
-4,124
10,204
-56,633
-1,926
-543
-3,729
-1,197
-8,488
-190,712
2050
-613
-4,519
9,878
-56,749
-2,044
-547
-3,745
-1,198
-8,619
-192,095
5-11
-------
5.1.2.2 Alternatives
The model runs that EPA conducted estimated the inventories under the Proposal and
Alternative 2 minus 10 standards of non-GHG air pollutants resulting from tailpipe emissions
from light-duty cars and trucks, and the upstream emissions associated with the fuels used to
power those vehicles (both at the refinery and the electricity generating unit). The tailpipe
emissions of PM2.5, NOx, VOCs, CO and SO2 are estimated using emission factors from EPA's
MOVES model. The tailpipe emission factors used are identical to those used in NHTSA's recent
NPRM.9 The upstream emissions are then calculated using emission factors applied to the
gallons of liquid fuels projected to be consumed and the kilowatt hours of electricity projected to
be consumed. The upstream emission factors used in this final rule modeling have been updated
since EPA's proposed rule. The updated upstream emission factors are identical to those used in
the recent NHTSA CAFE proposal and were generated using the DOE/Argonne GREET
model.10
Table 5-6: Impacts on Upstream Non-GHG Emissions under the Proposal Standards Relative to the No
Action Scenario

PM2.5 (US tons)
NOx (US tons)
SO2 (US tons)
VOC (US tons)
CO (US tons)
Year
EGU
Refinery
EGU
Refinery
EGU
Refinery
EGU
Refinery
EGU
Refinery
2023
100
-81
1,190
-902
1,040
-409
177
-1,430
630
-506
2024
206
-162
2,450
-1,811
2,124
-817
369
-2,851
1,311
-1,021
2025
350
-288
4,163
-3,209
3,578
-1,446
635
-5,079
2,252
-1,815
2026
483
-424
5,735
-4,707
4,884
-2,120
886
-7,501
3,137
-2,670
2027
618
-585
7,337
-6,482
6,187
-2,921
1,148
-10,398
4,063
-3,687
2028
744
-748
8,825
-8,282
7,361
-3,733
1,401
-13,377
4,950
-4,724
2029
879
-903
10,399
-9,976
8,738
-4,512
1,651
-16,165
5,835
-5,698
2030
1,007
-1,051
11,889
-11,591
10,065
-5,259
1,887
-18,837
6,674
-6,630
2031
1,123
-1,187
13,223
-13,077
11,277
-5,951
2,098
-21,315
7,425
-7,492
2032
1,240
-1,321
14,553
-14,527
12,506
-6,631
2,309
-23,748
8,175
-8,337
2033
1,352
-1,449
15,831
-15,909
13,706
-7,283
2,512
-26,086
8,896
-9,146
2034
1,452
-1,567
17,001
-17,197
14,664
-7,882
2,711
-28,236
9,596
-9,895
2035
1,541
-1,674
18,039
-18,360
15,501
-8,425
2,892
-30,188
10,227
-10,574
2036
1,619
-1,768
18,950
-19,374
16,221
-8,900
3,054
-31,899
10,791
-11,169
2037
1,683
-1,844
19,699
-20,198
16,796
-9,288
3,191
-33,299
11,267
-11,655
2038
1,737
-1,910
20,326
-20,904
17,263
-9,623
3,310
-34,507
11,678
-12,074
2039
1,773
-1,964
20,747
-21,486
17,553
-9,894
3,392
-35,513
11,964
-12,420
2040
1,803
-2,011
21,096
-21,979
17,778
-10,123
3,464
-36,374
12,211
-12,716
2041
1,824
-2,048
21,340
-22,374
17,912
-10,308
3,520
-37,074
12,400
-12,955
2042
1,834
-2,074
21,456
-22,649
17,935
-10,437
3,554
-37,576
12,515
-13,126
2043
1,848
-2,104
21,620
-22,955
17,997
-10,580
3,597
-38,131
12,660
-13,315
2044
1,849
-2,125
21,622
-23,190
17,977
-10,687
3,604
-38,561
12,681
-13,451
2045
1,842
-2,140
21,544
-23,351
17,890
-10,759
3,597
-38,867
12,654
-13,544
2046
1,835
-2,153
21,455
-23,483
17,795
-10,817
3,588
-39,122
12,622
-13,620
2047
1,823
-2,160
21,314
-23,567
17,655
-10,853
3,571
-39,296
12,559
-13,669
2048
1,811
-2,167
21,175
-23,642
17,517
-10,884
3,554
-39,456
12,498
-13,713
2049
1,812
-2,188
21,201
-23,896
17,490
-11,021
3,564
-39,641
12,528
-13,802
2050
1,812
-2,208
21,211
-24,142
17,448
-11,154
3,571
-39,814
12,548
-13,887
5-12
-------
Table 5-7: Impacts on Non-GHG Emissions under the Proposal Standards Relative to the No Action Scenario

Upstream Emissions (US tons)
Tailpipe Emissions (US tons)
Year
pm25
NOx
S02
voc
CO
pm25
NOx
S02
VOC
CO
2023
19
288
630
-1,252
124
5
551
-27
778
4,643
2024
44
640
1,307
-2,482
290
7
948
-57
1,380
7,860
2025
62
955
2,132
-4,444
437
4
1,246
-102
1,849
9,188
2026
59
1,028
2,764
-6,615
467
3
1,584
-150
2,402
11,572
2027
33
855
3,265
-9,250
376
-7
1,588
-207
2,485
8,257
2028
-5
543
3,628
-11,976
226
-20
1,512
-265
2,467
3,668
2029
-24
423
4,226
-14,514
137
-37
1,286
-320
2,233
-3,534
2030
-43
298
4,806
-16,950
43
-56
932
-372
1,810
-12,887
2031
-64
146
5,326
-19,216
-68
-74
652
-421
1,473
-21,185
2032
-81
26
5,875
-21,438
-162
-93
348
-468
1,090
-30,146
2033
-96
-78
6,423
-23,574
-249
-113
39
-514
682
-39,590
2034
-115
-196
6,782
-25,525
-299
-132
-257
-556
279
-48,723
2035
-133
-321
7,076
-27,296
-347
-187
-901
-593
-1,997
-63,422
2036
-148
-424
7,321
-28,845
-377
-207
-1,169
-626
-2,452
-71,908
2037
-161
-500
7,508
-30,109
-387
-225
-1,394
-653
-2,855
-79,513
2038
-173
-578
7,640
-31,198
-395
-242
-1,589
-677
-3,227
-86,536
2039
-191
-740
7,659
-32,121
-456
-257
-1,746
-696
-3,549
-92,586
2040
-207
-883
7,654
-32,909
-505
-272
-1,885
-713
-3,851
-98,171
2041
-224
-1,033
7,604
-33,554
-556
-284
-1,995
-726
-4,101
-102,802
2042
-240
-1,193
7,498
-34,022
-611
-294
-2,078
-736
-4,311
-106,641
2043
-256
-1,335
7,416
-34,534
-655
-305
-2,160
-746
-4,533
-110,466
2044
-277
-1,568
7,291
-34,958
-770
-313
-2,222
-753
-4,722
-113,561
2045
-298
-1,807
7,132
-35,270
-890
-320
-2,265
-758
-4,895
-115,937
2046
-318
-2,028
6,978
-35,534
-998
-326
-2,305
-761
-5,047
-118,042
2047
-338
-2,253
6,803
-35,726
-1,110
-332
-2,335
-763
-5,182
-119,820
2048
-356
-2,467
6,633
-35,902
-1,215
-336
-2,360
-765
-5,305
-121,407
2049
-376
-2,694
6,469
-36,078
-1,274
-340
-2,384
-768
-5,415
-122,928
2050
-396
-2,932
6,294
-36,243
-1,340
-344
-2,403
-770
-5,509
-124,246
5-13
-------
Table 5-8: Impacts on Upstream Non-GHG Emissions under the Alternative 2 minus 10 Standards Relative
to the No Action Scenario

PM2.5 (US tons)
NOx (US tons)
SO2 (US tons)
VOC (US tons)
CO (US tons)
Year
EGU
Refinery
EGU
Refinery
EGU
Refinery
EGU
Refinery
EGU
Refinery
2023
165
-165
1,967
-1,844
1,719
-840
293
-2,915
1,042
-1,033
2024
317
-305
3,766
-3,402
3,264
-1,543
568
-5,383
2,015
-1,914
2025
475
-468
5,651
-5,202
4,857
-2,356
862
-8,272
3,057
-2,937
2026
690
-685
8,193
-7,610
6,977
-3,444
1,265
-12,181
4,482
-4,310
2027
899
-927
10,678
-10,280
9,004
-4,652
1,671
-16,559
5,913
-5,838
2028
1,114
-1,187
13,220
-13,140
11,027
-5,945
2,099
-21,305
7,415
-7,484
2029
1,341
-1,431
15,879
-15,818
13,343
-7,179
2,520
-25,724
8,909
-9,022
2030
1,565
-1,669
18,472
-18,420
15,638
-8,386
2,932
-30,043
10,369
-10,523
2031
1,771
-1,891
20,846
-20,840
17,780
-9,516
3,308
-34,088
11,706
-11,925
2032
1,966
-2,102
23,079
-23,130
19,832
-10,593
3,662
-37,945
12,965
-13,257
2033
2,145
-2,299
25,118
-25,249
21,747
-11,597
3,985
-41,543
14,115
-14,497
2034
2,308
-2,480
27,012
-27,228
23,299
-12,519
4,308
-44,855
15,246
-15,648
2035
2,453
-2,646
28,707
-29,020
24,668
-13,357
4,602
-47,867
16,275
-16,694
2036
2,583
-2,793
30,229
-30,615
25,876
-14,106
4,871
-50,562
17,214
-17,628
2037
2,691
-2,916
31,492
-31,938
26,853
-14,731
5,101
-52,813
18,013
-18,408
2038
2,783
-3,022
32,564
-33,078
27,657
-15,272
5,302
-54,765
18,710
-19,083
2039
2,848
-3,111
33,320
-34,037
28,190
-15,718
5,448
-56,422
19,215
-19,653
2040
2,896
-3,185
33,881
-34,818
28,553
-16,084
5,564
-57,796
19,612
-20,121
2041
2,928
-3,244
34,258
-35,442
28,754
-16,377
5,650
-58,911
19,905
-20,499
2042
2,955
-3,294
34,574
-35,973
28,901
-16,627
5,727
-59,873
20,167
-20,824
2043
2,972
-3,334
34,762
-36,381
28,936
-16,820
5,784
-60,632
20,355
-21,078
2044
2,970
-3,365
34,741
-36,719
28,884
-16,972
5,790
-61,248
20,374
-21,274
2045
2,959
-3,387
34,605
-36,956
28,737
-17,077
5,777
-61,694
20,326
-21,411
2046
2,945
-3,404
34,439
-37,140
28,563
-17,157
5,759
-62,052
20,260
-21,517
2047
2,933
-3,421
34,303
-37,326
28,414
-17,237
5,747
-62,412
20,213
-21,625
2048
2,921
-3,436
34,153
-37,489
28,253
-17,307
5,732
-62,731
20,157
-21,720
2049
2,923
-3,468
34,199
-37,876
28,212
-17,516
5,748
-62,997
20,208
-21,853
2050
2,928
-3,500
34,263
-38,270
28,185
-17,727
5,768
-63,260
20,269
-21,991
5-14
-------
Table 5-9: Impacts on Non-GHG Emissions under the Alternative 2 minus 10 Standards Relative to the No
Action Scenario

Upstream Emissions (US tons)
Tailpipe Emissions (US tons)
pm25
NOx
S02
voc
CO
pm25
NOx
S02
VOC
CO
2023
0
123
878
-2,622
9
13
1,252
-55
1,752
11,518
2024
12
364
1,721
-4,815
101
19
2,004
-105
2,902
18,564
2025
8
449
2,501
-7,410
119
17
2,365
-162
3,509
20,694
2026
4
583
3,534
-10,915
172
7
2,572
-238
3,910
19,264
2027
-28
398
4,352
-14,888
74
-3
2,780
-323
4,332
17,328
2028
-73
80
5,082
-19,206
-69
-25
2,632
-415
4,272
9,005
2029
-90
61
6,164
-23,203
-113
-51
2,297
-500
3,941
-2,175
2030
-105
52
7,251
-27,111
-154
-86
1,694
-583
3,226
-18,312
2031
-121
7
8,264
-30,780
-219
-117
1,213
-661
2,653
-32,396
2032
-137
-50
9,240
-34,282
-293
-150
693
-735
2,002
-47,700
2033
-153
-131
10,151
-37,557
-382
-183
178
-805
1,322
-63,181
2034
-173
-217
10,780
-40,548
-402
-216
-324
-868
643
-78,463
2035
-193
-313
11,311
-43,265
-419
-308
-1,402
-926
-3,144
-102,406
2036
-210
-386
11,770
-45,691
-414
-343
-1,867
-978
-3,924
-116,830
2037
-224
-446
12,122
-47,712
-394
-375
-2,259
-1,021
-4,620
-129,816
2038
-239
-514
12,386
-49,462
-373
-405
-2,602
-1,058
-5,269
-142,066
2039
-263
-717
12,472
-50,974
-438
-432
-2,881
-1,090
-5,832
-152,616
2040
-289
-937
12,469
-52,232
-509
-456
-3,124
-1,115
-6,323
-161,876
2041
-316
-1,184
12,377
-53,261
-593
-477
-3,315
-1,136
-6,747
-169,779
2042
-339
-1,399
12,274
-54,145
-657
-496
-3,474
-1,154
-7,133
-176,790
2043
-362
-1,619
12,116
-54,848
-723
-512
-3,604
-1,168
-7,500
-182,886
2044
-395
-1,978
11,912
-55,458
-900
-526
-3,702
-1,178
-7,812
-187,838
2045
-428
-2,351
11,660
-55,917
-1,085
-537
-3,772
-1,185
-8,072
-191,655
2046
-459
-2,702
11,406
-56,293
-1,257
-548
-3,836
-1,190
-8,334
-195,157
2047
-488
-3,023
11,177
-56,665
-1,412
-558
-3,894
-1,195
-8,585
-198,407
2048
-515
-3,336
10,946
-56,999
-1,562
-566
-3,944
-1,200
-8,822
-201,332
2049
-544
-3,677
10,696
-57,248
-1,646
-574
-3,982
-1,203
-9,014
-203,699
2050
-572
-4,007
10,458
-57,492
-1,722
-580
-4,016
-1,208
-9,191
-205,912
5.2 Projected Fuel Consumption
5.2.1 Final Rule
The revised standards will reduce not only greenhouse gas emissions but also fuel
consumption. Reducing fuel consumption is one of, although not the only, means of reducing
greenhouse gas emissions from the transportation fleet.
Table 5-10 shows the estimated fuel consumption changes, including rebound effects, credit
usage and advanced technology multiplier use, under the final standards relative to the no action
scenario.
5-15
-------
Table 5-10: Impacts on Fuel Consumption for the Final Standards Relative to the No Action Scenario

Gasoline
% of 2020
Oil
Electricity
% of 2020
Year
Equivalents
(Million Gallons)
US Gasoline
Consumption
(Million
BBL)
(Gigawatt
hours)
US Electricity
Consumption
2023
-582
-0.5%
-11
3,631
0.1%
2024
-1,197
-1.0%
-23
8,241
0.2%
2025
-2,067
-1.7%
-39
14,593
0.4%
2026
-3,245
-2.6%
-61
23,196
0.6%
2027
-4,603
-3.7%
-87
32,224
0.8%
2028
-6,031
-4.9%
-114
41,712
1.1%
2029
-7,376
-6.0%
-139
50,609
1.3%
2030
-8,680
-7.0%
-164
59,241
1.6%
2031
-9,906
-8.0%
-187
67,266
1.8%
2032
-11,100
-9.0%
-209
75,194
2.0%
2033
-12,219
-9.9%
-231
82,594
2.2%
2034
-13,260
-10.7%
-250
89,541
2.4%
2035
-14,203
-11.5%
-268
95,798
2.5%
2036
-15,046
-12.2%
-284
101,503
2.7%
2037
-15,781
-12.8%
-298
106,634
2.8%
2038
-16,417
-13.3%
-310
111,098
2.9%
2039
-16,964
-13.7%
-320
114,939
3.0%
2040
-17,424
-14.1%
-329
118,225
3.1%
2041
-17,798
-14.4%
-336
120,847
3.2%
2042
-18,091
-14.6%
-341
122,895
3.2%
2043
-18,329
-14.8%
-346
124,591
3.3%
2044
-18,494
-14.9%
-349
125,718
3.3%
2045
-18,620
-15.0%
-351
126,590
3.3%
2046
-18,714
-15.1%
-353
127,333
3.4%
2047
-18,772
-15.2%
-354
127,808
3.4%
2048
-18,818
-15.2%
-355
128,233
3.4%
2049
-18,842
-15.2%
-356
128,458
3.4%
2050
-18,860
-15.2%
-356
128,625
3.4%
Sum
-361,438

-6,821
2,457,336

Notes:





The CCEMS effects reports (i.e., the model output files) report all liquid fuels as gasoline equivalents; to
determine barrels of oil, the gasoline equivalents have been treated as retail gasoline having 90 percent pure
gasoline which have then been adjusted using the ratio of the energy densities of pure gasoline to oil
(114,200/129,670 both in BTU/gallon, GREET 2017) which is then divided by 42 gallons in a barrel of oil;
according to the Energy Information Administration (EIA), US gasoline consumption in 2020 was 123.73 billion
gallons, roughly 16 percent less (due to the coronavirus pandemic) than the highest consumption on record
(2018).11 According to the Department of Energy, there are 33.7 kWh of electricity per gallon gasoline
equivalent, the metric reported by CCEMS for electricity consumption and used here to convert to kWh.
According to EIA, the US consumed 3,800,000 gigawatt hours of electricity in 2020.

5-16
-------
5.2.2 Alternatives
Table 5-1 land Table 5-12 show the estimated fuel consumption changes, including rebound
effects, credit usage and advanced technology multiplier use, under the Proposal and Alternative
2 minus 10 standards, respectively, relative to the no action scenario.
Table 5-11: Impacts on Fuel Consumption for the Proposal Standards Relative to the No Action Scenario
Year
Gasoline
Equivalents
(Million Gallons)
% of 2020
US Gasoline
Consumption
Oil
(Million
BBL)
Electricity
(Gigawatt
hours)
% of 2020
US Electricity
Consumption
2023
-429
-0.3%
-8
3,272
0.1%
2024
-883
-0.7%
-17
6,967
0.2%
2025
-1,576
-1.3%
-30
12,256
0.3%
2026
-2,325
-1.9%
-44
17,501
0.5%
2027
-3,216
-2.6%
-61
23,243
0.6%
2028
-4,130
-3.3%
-78
29,062
0.8%
2029
-4,982
-4.0%
-94
34,452
0.9%
2030
-5,800
-4.7%
-109
39,631
1.0%
2031
-6,558
-5.3%
-124
44,346
1.2%
2032
-7,303
-5.9%
-138
49,111
1.3%
2033
-8,017
-6.5%
-151
53,754
1.4%
2034
-8,671
-7.0%
-164
57,983
1.5%
2035
-9,265
-7.5%
-175
61,799
1.6%
2036
-9,784
-7.9%
-185
65,210
1.7%
2037
-10,208
-8.3%
-193
68,090
1.8%
2038
-10,575
-8.5%
-200
70,575
1.9%
2039
-10,880
-8.8%
-205
72,643
1.9%
2040
-11,141
-9.0%
-210
74,492
2.0%
2041
-11,352
-9.2%
-214
76,002
2.0%
2042
-11,504
-9.3%
-217
77,073
2.0%
2043
-11,672
-9.4%
-220
78,341
2.1%
2044
-11,787
-9.5%
-222
79,224
2.1%
2045
-11,866
-9.6%
-224
79,828
2.1%
2046
-11,931
-9.6%
-225
80,407
2.1%
2047
-11,971
-9.7%
-226
80,800
2.1%
2048
-12,008
-9.7%
-227
81,211
2.1%
2049
-12,051
-9.7%
-227
81,663
2.1%
2050
-12,089
-9.8%
-228
82,053
2.2%
Sum
-233,975

-4,416
1,580,986

Notes:
See prior table.
5-17
-------
Table 5-12: Impacts on Fuel Consumption for Alternative 2 minus 10 Relative to the No Action Scenario
Year
Gasoline
Equivalents
(Million Gallons)
% of 2020
US Gasoline
Consumption
Oil
(Million
BBL)
Electricity
(Gigawatt
hours)
% of 2020
US Electricity
Consumption
2023
-873
-0.7%
-16
5,409
0.1%
2024
-1,638
-1.3%
-31
10,708
0.3%
2025
-2,528
-2.0%
-48
16,637
0.4%
2026
-3,721
-3.0%
-70
25,003
0.7%
2027
-5,055
-4.1%
-95
33,826
0.9%
2028
-6,497
-5.3%
-123
43,535
1.1%
2029
-7,836
-6.3%
-148
52,607
1.4%
2030
-9,145
-7.4%
-173
61,574
1.6%
2031
-10,370
-8.4%
-196
69,915
1.8%
2032
-11,539
-9.3%
-218
77,883
2.0%
2033
-12,629
-10.2%
-238
85,290
2.2%
2034
-13,631
-11.0%
-257
92,126
2.4%
2035
-14,541
-11.8%
-274
98,345
2.6%
2036
-15,354
-12.4%
-290
104,023
2.7%
2037
-16,033
-13.0%
-303
108,856
2.9%
2038
-16,622
-13.4%
-314
113,069
3.0%
2039
-17,123
-13.8%
-323
116,668
3.1%
2040
-17,530
-14.2%
-331
119,642
3.1%
2041
-17,860
-14.4%
-337
122,006
3.2%
2042
-18,145
-14.7%
-342
124,197
3.3%
2043
-18,369
-14.8%
-347
125,961
3.3%
2044
-18,537
-15.0%
-350
127,288
3.3%
2045
-18,654
-15.1%
-352
128,223
3.4%
2046
-18,746
-15.2%
-354
129,064
3.4%
2047
-18,840
-15.2%
-356
130,040
3.4%
2048
-18,922
-15.3%
-357
130,982
3.4%
2049
-18,985
-15.3%
-358
131,726
3.5%
2050
-19,055
-15.4%
-360
132,547
3.5%
Sum
-368,780

-6,960
2,517,150

Notes:
See prior table.
5.3 Projected Safety Impacts
EPA has long considered the safety implications of its emission standards. With respect to its
light-duty greenhouse gas emission regulations, EPA has historically considered the potential
impacts of GHG standards on safety including: the 2010 rule which established the 2012-2016
light-duty vehicle GHG standards, the 2012 rule which previously established 2017-2025 light-
duty vehicle GHG standards, the 2017 MTE Proposed Determination and the 2020 SAFE
Rulemaking. In addition, section 202(a)(4) of the Clean Air Act specifically prohibits the use of
an emission control device, system or element of design that will cause or contribute to an
unreasonable risk to public health, welfare, or safety.
The potential relationship between GHG emissions standards and safety is multi-faceted, and
can be influenced not only by control technologies, but also by consumer decisions about vehicle
ownership and use. EPA has estimated the impacts of this rule on safety by accounting for
changes in new vehicle purchase, changes in vehicle scrappage, fleet turnover and VMT, and
changes in vehicle weight as an emissions control strategy. Safety impacts relate to changes in
5-18
-------
the use of vehicles in the fleet, relative mass changes, and the turnover of fleet to newer and safer
vehicles fleet turnover have been estimated and considered in the standard setting process.
The GHG emissions standards are attribute-based standards, using vehicle footprint as the
attribute. Footprint is defined as a vehicle's wheelbase multiplied by its average track width—in
other words, the area enclosed by the points at which the wheels meet the ground. The standards
are therefore generally based on a vehicle's size: larger vehicles have numerically higher GHG
emissions targets and smaller vehicles have numerically lower GHG emissions targets.
Footprint-based standards help to distribute the burden of compliance across all vehicle
footprints and across all manufacturers. Manufacturers are not compelled to build vehicles of any
particular size or type, and each manufacturer has its own fleetwide standard for its car and truck
fleets in each year that reflects the light-duty vehicles it chooses to produce.
Consistent with previous light-duty GHG analyses, EPA assessed the potential of these final
MY 2023-2026 standards to affect vehicle safety. EPA applied the same historical relationships
between mass, size, and fatality risk that were established and documented in the SAFE
rulemaking. These relationships are based on the statistical analysis of historical crash data,
which included an analysis performed by using the most recently available crash studies based
on data for model years 2007 to 2011. EPA used the findings of this analysis to estimate safety
impacts of the modeled mass reductions over the lifetimes of new vehicles in response to MYs
2023-2026 standards. As in initially promulgating the GHG standards, the MTE Proposed
Determination and this rule, EPA's assessment is that manufacturers can achieve the MYs 2023-
2026 standards while using modest levels of mass reduction as one technology option among
many. On the whole, EPA considers safety impacts in the context of all projected health impacts
from the rule including public health benefits from the projected reductions in air pollution.
The projected change in risk of fatal and non-fatal injuries is influenced by changes in fleet
mix (car/truck share), vehicle scrappage rates, distribution of VMT among vehicles in the fleet
and vehicle mass. EPA estimates that these factors together will result in an average 0.06 percent
increase (with results from sensitivity cases ranging from a decrease of 0.25 percent to an
increase of 0.36 percent) in the annual fatalities per billion miles driven through 2050.13 In
addition to changes in risk, EPA also considered the projected impact of the final standards on
the absolute number of fatal and non-fatal injuries. The majority of the fatalities projected would
result from the projected increased driving - i.e., people choosing to drive more due to the lower
operating costs of more efficient vehicles. Our cost-benefit analysis accounts for both the value
of this additional driving and its associated risk, which we assume are considerations in the
decision to drive. The risk valuation associated with this increase in driving partially offsets the
associated increase in societal costs due to increased fatalities and non-fatal injuries.
This analysis projects that there will be an increase in vehicle miles traveled (VMT) under the
revised standards of 304 billion miles compared to the No Action case through 2050 (an increase
of about 0.3 percent). EPA estimates that vehicle safety, in terms of risk measured as the total
fatalities per the total distance travelled over this period, will remain almost unchanged at 5.012
fatalities per billion miles under the rule, compared to 5.010 fatalities per billion miles for the no-
action case. EPA has also estimated, over the same 30-year period, that total fatalities will
increase by 1,780, with 1,348 deaths attributed to increased driving and 432 deaths attributed to
the increase in fatality risk. In other words, approximately 75 percent of the change in fatalities
under these revised standards is due to projected increases in VMT and mobility (i.e., people
5-19
-------
driving more). Our analysis also considered the increase in non-fatal injuries. Consistent with the
SAFE FRM, EPA assumed that non-fatal injuries scale with fatal injuries.
EPA also estimated the societal costs of these safety impacts using assumptions consistent
with the SAFE FRM (see Table 10-1). Specifically, we are continuing to use the cost associated
with each fatality of $10.4 million. We have also continued to use a scalar of approximately 1.6
applied to fatality costs to estimate non-fatal injury costs. In addition, we have accounted for the
driver's inherent valuation of risk when making the decision to drive more due to rebound. This
risk valuation partially offsets the fatal and non-fatal injury costs described above, and,
consistent with the SAFE FRM, is calculated as 90 percent of the fatal and non-fatal injury costs
due to rebound to reflect the fact that consumers do not fully evaluate the risks associated with
this additional driving.
5-20
-------
References for Chapter 5
1	U.S. EPA. Overview of EPA's MOtor Vehicle Emission Simulator (MOVES3). EPA-420-R-21-004. March 2021.
2	86 FR 49602
3	U.S. Department of Energy, Argonne National Laboratory, Greenhouse gases, Regulated Emissions,
and Energy use in Transportation (GREET) Model, Last Update: 9 Oct. 2020, https://greet.es.anl.gov/.
4	U.S. EPA. Overview of EPA's MOtor Vehicle Emission Simulator (MOVES3). EPA-420-R-21-004. March 2021.
5	86 FR 49602
6	U.S. Department of Energy, Argonne National Laboratory, Greenhouse gases, Regulated Emissions,
and Energy use in Transportation (GREET) Model, Last Update: 9 Oct. 2020, https://greet.es.anl.gov/.
7	86 FR 49602.
8	U.S. Department of Energy, Argonne National Laboratory, Greenhouse gases, Regulated Emissions,
and Energy use in Transportation (GREET) Model, Last Update: 9 Oct. 2020, https://greet.es.anl.gov/.
9	86 FR 49602.
10	U.S. Department of Energy, Argonne National Laboratory, Greenhouse gases, Regulated Emissions,
and Energy use in Transportation (GREET) Model, Last Update: 9 Oct. 2020, https://greet.es.anl.gov/.
11	www.eia.gov/tools/faqs, accessed on 11/5/2021, see US_gasoline_consumption_2020.pdf contained in the docket
fortius rule.
12	www.eia.gov/energyexplained/electricity/use-of-electricity.php, accessed on 11/5/2021, see
US_electricity_consumption_2020.pdf contained in the docket for this rule.
13	This range of fatality risk values is based on a sensitivity analysis using the 5% to 95% confidence interval of
mass-safety coefficients presented in the SAFE FRM.
5-21
-------
Chapter 6: Vehicle Costs, Fuel Savings and Non-Emission Benefits
In this chapter, EPA presents our estimate of the costs, the fuel savings and the non-emission
benefits associated with the revised standards.
6.1 Costs
The presentation here summarizes the vehicle level costs associated with the new
technologies expected to be added to meet the MY 2023 and later GHG standards, including
hardware costs to comply with the AJC credit program. The analysis summarized here also
provides costs associated with congestion and noise (see Chapter 3), and for fatalities and non-
fatal crashes and includes rebound effects.
For our analysis of safety impacts and how they are reflected in the benefit cost analysis,
having used the CCEMS, we have also used the safety-related inputs consistent with the NPRM.
For example, we have used the costs associated with fatalities of $10.4 million, as was done in
the NPRM and the SAFE FRM. We have also used, as mentioned above, the scaler of
approximately 1.6 applied to fatality costs to estimate non-fatal crash costs. In addition, we have
offset the fatality costs with a fatality risk value calculated as the fatalities due to rebound driving
multiplied by the fatality costs scaled by 90 percent to reflect the fact that consumers do not fully
evaluate the risks associated with driving. The same non-fatal crash risk scaler was applied to the
fatality risk value to estimate the non-fatal crash risk value. All of this is done exactly as was
done in the NPRM and in the SAFE FRM with the exception that, rather than presenting fatality
costs and non-fatal crash costs as "Costs" and fatality risk value and non-fatal crash risk value as
"Benefits," we have calculated the net of these and present the net result as a "Cost."
6-1
-------
6.1.1 Final Rule
Table 6-1: Costs Associated with the Final Program Relative to the No Action Scenario (SBillions of 2018
dollars)
Calendar
Year
Foregone
Consumer
Sales Surplus
Technology
Costs
Congestion
Noise
Fatality Costs
Non-fatal
Crash Costs
Total Costs
2023
0.029
5.6
0.03
0.00045
0.13
0.23
6.1
2024
0.045
8.9
0.048
0.00075
0.23
0.38
9.6
2025
0.079
13
0.082
0.0013
0.32
0.54
14
2026
0.11
16
0.12
0.002
0.42
0.7
17
2027
0.13
18
0.19
0.0031
0.49
0.82
20
2028
0.12
19
0.26
0.0043
0.51
0.85
20
2029
0.1
17
0.32
0.0054
0.49
0.82
19
2030
0.093
17
0.4
0.0067
0.44
0.73
19
2031
0.093
17
0.46
0.0077
0.41
0.67
19
2032
0.088
17
0.53
0.0088
0.37
0.61
19
2033
0.086
17
0.58
0.0097
0.34
0.55
19
2034
0.082
17
0.64
0.011
0.3
0.49
19
2035
0.078
17
0.68
0.011
0.27
0.44
19
2036
0.074
17
0.73
0.012
0.23
0.38
18
2037
0.071
16
0.76
0.013
0.21
0.34
18
2038
0.068
16
0.79
0.013
0.18
0.3
17
2039
0.066
16
0.81
0.014
0.17
0.27
17
2040
0.063
16
0.84
0.014
0.15
0.25
17
2041
0.061
16
0.86
0.014
0.15
0.24
17
2042
0.06
16
0.87
0.015
0.14
0.23
17
2043
0.058
16
0.88
0.015
0.14
0.22
17
2044
0.057
15
0.89
0.015
0.14
0.22
17
2045
0.057
15
0.89
0.015
0.14
0.23
17
2046
0.055
15
0.9
0.015
0.14
0.23
16
2047
0.054
15
0.9
0.015
0.15
0.24
16
2048
0.053
15
0.91
0.015
0.15
0.24
16
2049
0.053
15
0.9
0.015
0.15
0.25
16
2050
0.052
15
0.9
0.015
0.16
0.25
16
PV, 3%
$1.3
$280
$9.6
$0.16
$4.9
$8.1
$300
PV, 7%
$0.84
$160
$4.8
$0.08
$3.2
$5.3
$180
Annualized,
3%
$0,069
$14
$0.49
$0.0082
$0.25
$0.42
$15
Annualized,
7%
$0,068
$13
$0.39
$0.0065
$0.26
$0.43
$14
Notes:
"Foregone Consumer Sales Surplus" refers to the difference between a vehicle's price and the buyer's willingness
to pay for the new vehicle; the impact reflects the reduction in new vehicle sales described in Chapter 8.1.
6-2
-------
6.1.2 Alternatives
Table 6-2: Costs Associated with the Proposal Relative to the No Action Scenario (SBillions of 2018 dollars)
Calendar
Year
Foregone
Consumer Sales
Surplus
Technology
Costs
Congestion
Noise
Fatality
Costs
Non-fatal
Crash
Costs
Total
Costs
2023
0.025
4.7
0.014
0.00021
0.11
0.18
5
2024
0.032
6.5
0.026
0.0004
0.19
0.32
7.1
2025
0.06
9.9
0.052
0.00083
0.25
0.42
11
2026
0.061
10
0.086
0.0014
0.32
0.54
11
2027
0.066
12
0.14
0.0022
0.34
0.56
13
2028
0.062
12
0.19
0.0031
0.34
0.56
13
2029
0.051
11
0.24
0.0039
0.32
0.53
12
2030
0.046
11
0.3
0.0049
0.28
0.46
12
2031
0.046
11
0.34
0.0056
0.26
0.42
12
2032
0.043
11
0.39
0.0064
0.23
0.38
12
2033
0.042
11
0.43
0.007
0.21
0.34
12
2034
0.04
11
0.47
0.0077
0.18
0.3
12
2035
0.038
11
0.5
0.0082
0.16
0.27
12
2036
0.037
10
0.53
0.0087
0.14
0.23
11
2037
0.035
10
0.56
0.0091
0.12
0.2
11
2038
0.033
9.8
0.58
0.0094
0.11
0.18
11
2039
0.032
9.7
0.59
0.0096
0.097
0.16
11
2040
0.031
9.7
0.61
0.0099
0.089
0.14
11
2041
0.03
9.6
0.62
0.01
0.082
0.13
11
2042
0.029
9.6
0.62
0.01
0.078
0.13
10
2043
0.029
9.7
0.63
0.01
0.075
0.12
11
2044
0.028
9.6
0.64
0.01
0.071
0.12
10
2045
0.028
9.4
0.64
0.01
0.071
0.12
10
2046
0.027
9.3
0.64
0.011
0.072
0.12
10
2047
0.027
9.2
0.65
0.011
0.072
0.12
10
2048
0.026
9.1
0.65
0.011
0.072
0.12
10
2049
0.026
9.1
0.65
0.011
0.072
0.12
10
2050
0.025
9.1
0.65
0.011
0.072
0.12
9.9
PV, 3%
$0.72
$170
$7
$0.11
$3.2
$5.3
$190
PV, 7%
$0.46
$100
$3.5
$0,057
$2.1
$3.5
$110
Annualized,
3%
$0,037
$8.9
$0.35
$0.0058
$0.16
$0.27
$9.8
Annualized,
7%
$0,037
$8.4
$0.28
$0.0046
$0.17
$0.29
$9.2
Notes:
"Foregone Consumer Sales Surplus" refers to the difference between a vehicle's price and the buyer's willingness
to pay for the new vehicle; the impact reflects the reduction in new vehicle sales described in Chapter 8.1.
6-3
-------
Table 6-3: Costs Associated with Alternative 2 minus 10 Relative to the No Action Scenario (SBillions of 2018
dollars)
Calendar
Year
Foregone
Consumer Sales
Surplus
Technology
Costs
Congestion
Noise
Fatality
Costs
Non-fatal
Crash
Costs
Total
Costs
2023
$0.06
$10
$0,051
$0.0008
$0.24
$0.4
$11
2024
$0,073
$12
$0,085
$0.0013
$0.38
$0.64
$14
2025
$0,091
$14
$0.13
$0.0021
$0.45
$0.75
$15
2026
$0.12
$17
$0.18
$0.0029
$0.5
$0.83
$19
2027
$0.13
$18
$0.25
$0,004
$0.55
$0.91
$20
2028
$0.13
$19
$0.32
$0.0053
$0.55
$0.91
$21
2029
$0.11
$18
$0.38
$0.0064
$0.52
$0.87
$20
2030
$0,096
$18
$0.45
$0.0076
$0.46
$0.75
$19
2031
$0,096
$18
$0.51
$0.0086
$0.41
$0.68
$19
2032
$0,091
$18
$0.57
$0.0096
$0.37
$0.61
$19
2033
$0,089
$18
$0.63
$0.01
$0.33
$0.54
$19
2034
$0,085
$18
$0.68
$0,011
$0.29
$0.47
$19
2035
$0,081
$17
$0.72
$0,012
$0.25
$0.41
$19
2036
$0,077
$17
$0.76
$0,013
$0.22
$0.35
$19
2037
$0,074
$17
$0.79
$0,013
$0.19
$0.31
$18
2038
$0.07
$16
$0.82
$0,014
$0.16
$0.27
$18
2039
$0,068
$16
$0.84
$0,014
$0.15
$0.24
$18
2040
$0,066
$16
$0.86
$0,014
$0.13
$0.22
$17
2041
$0,064
$16
$0.88
$0,015
$0.12
$0.2
$17
2042
$0,062
$16
$0.89
$0,015
$0.12
$0.19
$17
2043
$0,061
$16
$0.9
$0,015
$0.11
$0.18
$17
2044
$0.06
$16
$0.9
$0,015
$0.11
$0.18
$17
2045
$0.06
$16
$0.9
$0,015
$0.11
$0.18
$17
2046
$0,058
$15
$0.91
$0,015
$0.12
$0.19
$17
2047
$0,056
$15
$0.91
$0,015
$0.12
$0.19
$17
2048
$0,055
$15
$0.91
$0,015
$0.12
$0.2
$16
2049
$0,055
$15
$0.9
$0,015
$0.12
$0.2
$16
2050
$0,054
$15
$0.9
$0,015
$0.13
$0.21
$16
PV, 3%
$1.5
$290
$10
$0.17
$5.3
$8.7
$320
PV, 7%
$0.94
$180
$5.2
$0,088
$3.6
$5.9
$190
Annualized,
3%
$0,075
$15
$0.52
$0.0087
$0.27
$0.44
$16
Annualized,
7%
$0,076
$14
$0.42
$0.0071
$0.29
$0.48
$15
Notes:
"Foregone Consumer Sales Surplus" refers to the difference between a vehicle's price and the buyer's willingness
to pay for the new vehicle; the impact reflects the reduction in new vehicle sales described in Chapter 8.1.
6-4
-------
6.2 Fuel Savings
Fuel savings presented here include fuel expenditure impacts for all fuels, including increased
expenditures on electricity and include rebound effects, credit usage and advanced technology
multiplier use. The net benefits calculations use the aggregate value of fuel savings (calculated
using pre-tax fuel prices) since savings in fuel taxes do not represent a reduction in the value of
economic resources utilized in producing and consuming fuel.
6.2.1 Final Rule
Table 6-4: Fuel Savings Associated with the Final Program ($Billions of 2018 dollars)
Calendar Year
Retail Fuel Savings
Fuel Tax Savings
Pre-Tax Fuel Savings
2023
$0.94
$0.31
$0.62
2024
$1.9
$0.64
$1.2
2025
$3.2
$1.1
$2.1
2026
$5.1
$1.7
$3.3
2027
$7.4
$2.4
$4.9
2028
$10
$3.2
$6.9
2029
$13
$3.8
00
00
&
2030
$16
$4.5
$12
2031
$18
$5.1
$13
2032
$21
$5.6
$16
2033
$24
$6.2
$17
2034
$26
$6.7
$19
2035
$28
$7.1
$21
2036
$30
$7.5
$23
2037
$32
$7.8
$24
2038
$34
$8.1
$26
2039
$35
$8.3
$27
2040
$37
$8.5
$29
2041
$38
$8.6
$30
2042
$39
$8.7
$31
2043
$40
$8.8
$31
2044
$41
$8.8
$32
2045
$41
$8.8
$32
2046
$42
$8.8
$33
2047
$42
$8.8
$33
2048
$42
$8.7
$33
2049
$42
$8.7
$33
2050
$42
$8.6
$33
PV, 3%
$420
$100
$320
PV, 7%
$210
$51
$150
Annualized, 3%
$21
$5.1
$16
Annualized, 7%
$17
$4.1
$12
6-5
-------
6.2.2 Alternatives
Table 6-5: Fuel Savings Associated with the Proposal (SBillions of 2018 dollars)
Calendar Year
Retail Fuel Savings
Fuel Tax Savings
Pre-Tax Fuel Savings
2023
$0.62
$0.23
$0.39
2024
$1.3
$0.47
$0.78
2025
$2.3
$0.84
$1.5
2026
$3.5
$1.2
$2.3
2027
$5.1
$1.7
$3.4
2028
00
VO
&
$2.2
$4.7
2029
$8.5
$2.6
$6
2030
$11
$3
$7.7
2031
$12
$3.4
$8.9
2032
$14
$3.7
$10
2033
$16
$4
$11
2034
$17
$4.4
$13
2035
$18
$4.6
$14
2036
$20
$4.9
$15
2037
$21
$5
$16
2038
$22
$5.2
$17
2039
$23
$5.3
$18
2040
$24
$5.4
$18
2041
$25
$5.5
$19
2042
$25
$5.5
$20
2043
$26
$5.6
$20
2044
$26
$5.6
$21
2045
$26
$5.6
$20
2046
$27
$5.6
$21
2047
$27
$5.6
$21
2048
$27
$5.6
$21
2049
$27
$5.6
$21
2050
$27
$5.5
$22
PV, 3%
$270
$65
$210
PV, 7%
$130
$33
$100
Annualized, 3%
$14
$3.3
$11
Annualized, 7%
$11
$2.7
$8.2
6-6
-------
Table 6-6: Fuel Savings Associated with Alternative 2 minus 10 ($Billions of 2018 dollars)
Calendar Year
Retail Fuel Savings
Fuel Tax Savings
Pre-Tax Fuel Savings
2023
$1.4
$0.47
$0.94
2024
$2.6
$0.88
$1.7
2025
$4.1
$1.4
$2.7
2026
$6
$2
$4
2027
$8.3
$2.7
$5.6
2028
$11
$3.4
$7.6
2029
$14
$4.1
$9.5
2030
$17
$4.7
$12
2031
$19
$5.3
$14
2032
$22
$5.9
$16
2033
$24
$6.4
$18
2034
$27
$6.8
$20
2035
$29
$7.3
$21
2036
$31
$7.6
$23
2037
$33
$7.9
$25
2038
$34
$8.2
$26
2039
$36
$8.4
$27
2040
$37
$8.5
$29
2041
$38
$8.6
$30
2042
$39
$8.7
$31
2043
$40
$8.8
$31
2044
$41
$8.8
$32
2045
$40
$8.8
$32
2046
$41
$8.8
$33
2047
$42
$8.8
$33
2048
$42
$8.8
$33
2049
$42
$8.7
$33
2050
$42
$8.7
$33
PV, 3%
$430
$100
$320
PV, 7%
$210
$53
$160
Annualized, 3%
$22
$5.2
$16
Annualized, 7%
$17
$4.2
$13
6-7
-------
6.3 Non-Emission Benefits
Non-emission benefits include the drive value, or drive surplus (see Chapter 3.4) and the
energy security benefits (see Chapter 3.2). With changes in fuel consumption, there are also
associated changes in the amount of time spent refueling vehicles. Consistent with the
assumptions used in the NPRM (and presented in Table 6-7 and Table 6-8), the costs of time
spent refueling are calculated as the total amount of time the driver of a typical vehicle would
spend refueling multiplied by the value of their time. If less time is spent refueling vehicles
under the revised standards, then a refueling time savings would be incurred and vice versa.
Table 6-7: CCEMS Inputs used to Estimate Refueling Time Costs

Cars
Vans/SUVs
Pickups
Fixed Component of Average Refueling Time in Minutes (by Fuel Type)
Gasoline
3.5
3.5
3.5
Ethanol-85
3.5
3.5
3.5
Diesel
3.5
3.5
3.5
Electricity
3.5
3.5
3.5
Hydrogen
0
0
0
Compressed Natural Gas
0
0
0
Average Tank Volume Refueled
65%
65%
65%
Value of Travel Time per Vehicle (2018 $/hour)
20.46
20.79
20.79
Table 6-8: CCEMS Inputs used to Estimate Electric Refueling Time Costs

Cars
Vans/SUVs
Pickups
Electric Vehicle Recharge Thresholds (BEV200)
Miles until mid-trip charging event
2,000
1,500
1,600
Share of miles charged mid-trip
6.00%
9.00%
8.00%
Charge rate (miles/hour)
67
67
67
Electric Vehicle Recharge Thresholds (BEV300)
Miles until mid-trip charging event
5,200
3,500
3,800
Share of miles charged mid-trip
3.00%
4.00%
4.00%
Charge rate (miles/hour)
100
100
100
6-8
-------
6.3.1 Non-emission Benefits of the Final Rule
Table 6-9: Benefits from Non-Emission Sources under the Final Rule (SBillions of 2018 dollars)
Calendar Year
Drive
Refueling Time
Energy Security
Total Non-Emission

Value
Savings
Benefits
Benefits
2023
$0,035
-$0.0052
$0,031
$0,061
2024
$0,055
-$0.03
$0,065
$0.09
2025
$0,091
-$0.07
$0.11
$0.13
2026
$0.14
-$0.12
$0.18
$0.2
2027
$0.22
-$0.15
$0.26
$0.33
2028
$0.32
-$0.19
$0.34
$0.47
2029
$0.42
-$0.23
$0.43
$0.61
2030
$0.55
-$0.27
$0.51
$0.79
2031
$0.64
-$0.29
$0.59
$0.94
2032
$0.74
-$0.34
$0.68
$1.1
2033
$0.83
-$0.38
$0.76
$1.2
2034
$0.93
-$0.43
$0.84
$1.3
2035
$1
-$0.47
$0.92
$1.5
2036
$1.1
-$0.51
$0.99
$1.6
2037
$1.2
-$0.56
$1.1
$1.7
2038
$1.2
-$0.6
$1.1
$1.8
2039
$1.3
-$0.63
$1.2
$1.8
2040
$1.3
-$0.67
$1.3
$1.9
2041
$1.4
-$0.69
$1.3
$2
2042
$1.4
-$0.7
$1.3
$2
2043
$1.4
-$0.72
$1.4
$2.1
2044
$1.4
-$0.73
$1.4
$2.1
2045
$1.4
-$0.75
$1.5
$2.1
2046
$1.5
-$0.77
$1.5
$2.2
2047
$1.5
-$0.78
$1.5
$2.2
2048
$1.5
-$0.81
$1.5
$2.2
2049
$1.5
-$0.82
$1.6
$2.2
2050
$1.5
-$0.83
$1.6
$2.3
PV, 3%
$15
$-7.4
$14
$21
PV, 7%
$7.2
$-3.6
$7
$11
Annualized,
$0.75
$-0.38
$0.73
$1.1
3%




Annualized,
$0.58
$-0.29
$0.56
$0.85
7%




6-9
-------
6.3.2 Non-emission Benefits of the Proposal and Alternative
Table 6-10: Benefits from Non-Emission Sources Under the Proposal (SBillions of 2018 dollars)
Calendar Year
Drive
Refueling Time
Energy Security
Total Non-Emission

Value
Savings
Benefits
Benefits
2023
$0,013
-$0,019
$0,023
$0,017
2024
$0,021
-$0.05
$0,048
$0,019
2025
$0,045
-$0,091
$0,086
$0.04
2026
$0,085
-$0.12
$0.13
$0,094
2027
$0.15
-$0.14
$0.18
$0.19
2028
$0.22
-$0.16
$0.24
$0.29
2029
$0.29
-$0.18
$0.29
$0.4
2030
$0.38
-$0.19
$0.34
$0.54
2031
$0.45
-$0.19
$0.39
$0.66
2032
$0.53
-$0.21
$0.45
$0.77
2033
$0.59
-$0.24
$0.5
$0.85
2034
$0.66
-$0.26
$0.55
$0.95
2035
$0.72
-$0.29
$0.6
$1
2036
$0.77
-$0.31
$0.65
$1.1
2037
$0.82
-$0.32
$0.69
$1.2
2038
$0.86
-$0.32
$0.73
$1.3
2039
$0.88
-$0.33
$0.77
$1.3
2040
$0.93
-$0.34
$0.81
$1.4
2041
$0.95
-$0.35
$0.83
$1.4
2042
$0.97
-$0.34
$0.86
$1.5
2043
$0.99
-$0.36
$0.88
$1.5
2044
$1
-$0.38
$0.91
$1.5
2045
$0.99
-$0.39
$0.92
$1.5
2046
$1
-$0.41
$0.95
$1.6
2047
$1
-$0.42
$0.97
$1.6
2048
$1
-$0.44
$0.99
$1.6
2049
$1
-$0.47
$1
$1.6
2050
$1
-$0.49
$1
$1.6
PV, 3%
$10
$-4.4
$9.3
$15
PV, 7%
$5
$-2.3
$4.6
$7.3
Annualized,
$0.52
$-0.22
$0.47
$0.77
3%




Annualized,
$0.4
$-0.18
$0.37
$0.59
7%




6-10
-------
Table 6-11: Benefits from Non-Emission Sources under Alternative 2 minus 10 ($Billions of 2018 dollars)
Calendar Year
Drive
Refueling Time
Energy Security
Total Non-Emission

Value
Savings
Benefits
Benefits
2023
$0.06
-$0,015
$0,047
$0,092
2024
$0,099
-$0,036
$0,089
$0.15
2025
$0.15
-$0,057
$0.14
$0.24
2026
$0.22
-$0,094
$0.21
$0.33
2027
$0.3
-$0.13
$0.28
$0.46
2028
$0.41
-$0.18
$0.37
$0.6
2029
$0.5
-$0.22
$0.45
$0.74
2030
$0.63
-$0.27
$0.54
$0.9
2031
$0.72
-$0.31
$0.62
$1
2032
$0.83
-$0.35
$0.7
$1.2
2033
$0.92
-$0.39
$0.78
$1.3
2034
$1
-$0.42
$0.86
$1.5
2035
$1.1
-$0.45
$0.94
$1.6
2036
$1.2
-$0.49
$1
$1.7
2037
$1.2
-$0.51
$1.1
$1.8
2038
$1.3
-$0.52
$1.2
$1.9
2039
$1.3
-$0.54
$1.2
$2
2040
$1.4
-$0.55
$1.3
$2.1
2041
$1.4
-$0.55
$1.3
$2.2
2042
$1.5
-$0.57
$1.4
$2.2
2043
$1.5
-$0.59
$1.4
$2.3
2044
$1.5
-$0.61
$1.4
$2.3
2045
$1.5
-$0.62
$1.5
$2.3
2046
$1.5
-$0.64
$1.5
$2.4
2047
$1.5
-$0.69
$1.5
$2.4
2048
$1.5
-$0.76
$1.6
$2.3
2049
$1.5
-$0.81
$1.6
$2.3
2050
$1.5
-$0.86
$1.6
$2.3
PV, 3%
$16
$-6.7
$15
$24
PV, 7%
$7.9
$-3.3
$7.2
$12
Annualized,
$0.81
$-0.34
$0.75
$1.2
3%




Annualized,
$0.64
$-0.27
$0.58
$0.95
7%




6-11
-------
Chapter 7: Non-GHG Health and Environmental Impacts
In this chapter we discuss the health effects associated with non-GHG pollutants, specifically:
particulate matter, ozone, nitrogen oxides (NOx), sulfur oxides (SOx), carbon monoxide and air
toxics. These pollutants will not be directly regulated by the standards, but the standards will
affect emissions of these pollutants and precursors.
7.1 Health and Environmental Impacts of Non-GHG Pollutants
7.1.1 Background on Non-GHG Pollutants Impacted by the Final Standards
7.1.1.1 Particulate Matter
Particulate matter (PM) is a complex mixture of solid particles and liquid droplets distributed
among numerous atmospheric gases which interact with solid and liquid phases. Particles in the
atmosphere range in size from less than 0.01 to more than 10 micrometers (|im) in diameter.1
Atmospheric particles can be grouped into several classes according to their aerodynamic
diameter and physical sizes. Generally, the three broad classes of particles include ultrafine
particles (UFPs, generally considered as particles with a diameter less than or equal to 0.1 |im
[typically based on physical size, thermal diffusivity or electrical mobility]), "fine" particles
(PM2.5; particles with a nominal mean aerodynamic diameter less than or equal to 2.5 |im), and
"thoracic" particles (PM10; particles with a nominal mean aerodynamic diameter less than or
equal to 10 |im). Particles that fall within the size range between PM2.5 and PM10, are referred to
as "thoracic coarse particles" (PM10-2.5, particles with a nominal mean aerodynamic diameter
greater than 2.5 |im and less than or equal to 10 |im). EPA currently has NAAQS for PM2.5 and
PMio.a
Most particles are found in the lower troposphere, where they can have residence times
ranging from a few hours to weeks. Particles are removed from the atmosphere by wet
deposition, such as when they are carried by rain or snow, or by dry deposition, when particles
settle out of suspension due to gravity. Atmospheric lifetimes are generally longest for PM2.5,
which often remains in the atmosphere for days to weeks before being removed by wet or dry
deposition.2 In contrast, atmospheric lifetimes for UFP and PM10-2.5 are shorter. Within hours,
UFP can undergo coagulation and condensation that lead to formation of larger particles in the
accumulation mode, or can be removed from the atmosphere by evaporation, deposition, or
reactions with other atmospheric components. PM10-2.5 are also generally removed from the
atmosphere within hours, through wet or dry deposition.3
Particulate matter consists of both primary and secondary particles. Primary particles are
emitted directly from sources, such as combustion-related activities (e.g., industrial activities,
motor vehicle operation, biomass burning), while secondary particles are formed through
atmospheric chemical reactions of gaseous precursors (e.g., sulfur oxides (SOx), nitrogen oxides
(NOx) and volatile organic compounds (VOCs)). From 2000 to 2017, national annual average
a Regulatory definitions of PM size fractions, and information on reference and equivalent methods for measuring
PM in ambient air, are provided in 40 CFR Parts 50, 53, and 58. With regard to national ambient air quality
standards (NAAQS) which provide protection against health and welfare effects, the 24-hour PM10 standard
provides protection against effects associated with short-term exposure to thoracic coarse particles (i.e., PM10-2.5).
7-1
-------
ambient PM2.5 concentrations have declined by over 40 percent,4 largely reflecting reductions in
emissions of precursor gases.
7.1.1.2	Ozone
Ground-level ozone pollution forms in areas with high concentrations of ambient NOx and
VOCs when solar radiation is strong. Major U.S. sources of NOx are highway and nonroad
motor vehicles, engines, power plants and other industrial sources, with natural sources, such as
soil, vegetation, and lightning, serving as smaller sources. Vegetation is the dominant source of
VOCs in the U.S. Volatile consumer and commercial products, such as propellants and solvents,
highway and nonroad vehicles, engines, fires, and industrial sources also contribute to the
atmospheric burden of VOCs at ground-level.
The processes underlying ozone formation, transport, and accumulation are complex. Ground-
level ozone is produced and destroyed by an interwoven network of free radical reactions
involving the hydroxyl radical (OH), NO, NO2, and complex reaction intermediates derived from
VOCs. Many of these reactions are sensitive to temperature and available sunlight. High ozone
events most often occur when ambient temperatures and sunlight intensities remain high for
several days under stagnant conditions. Ozone and its precursors can also be transported
hundreds of miles downwind which can lead to elevated ozone levels in areas with otherwise low
VOC or NOx emissions. As an air mass moves and is exposed to changing ambient
concentrations of NOx and VOCs, the ozone photochemical regime (relative sensitivity of ozone
formation to NOx and VOC emissions) can change.
When ambient VOC concentrations are high, comparatively small amounts of NOx catalyze
rapid ozone formation. Without available NOx, ground-level ozone production is severely
limited and VOC reductions will have little impact on ozone concentrations. Photochemistry
under these conditions is said to be "NOx-limited." When NOx levels are sufficiently high,
faster NO2 oxidation consumes more radicals, dampening ozone production. Under these "VOC-
limited" conditions (also referred to as "NOx-saturated" conditions), VOC reductions are
effective in reducing ozone, and NOx can react directly with ozone resulting in suppressed ozone
concentrations near NOx emission sources. Under these NOx-saturated conditions, NOx
reductions can actually increase local ozone under certain circumstances, but overall ozone
production (considering downwind formation) decreases and even in VOC-limited areas, NOx
reductions are not expected to increase ozone levels if the NOx reductions are sufficiently large -
large enough to become NOx-limited.
7.1.1.3	Nitrogen Oxides
Oxides of nitrogen (NOx) refers to nitric oxide (NO) and nitrogen dioxide (NO2). MostN02
is formed in the air through the oxidation of nitric oxide (NO) emitted when fuel is burned at a
high temperature. NOx is a criteria pollutant, regulated for its adverse effects on public health
and the environment, and highway vehicles are an important contributor to NOx emissions. NOx,
along with VOCs, are the two major precursors of ozone and NOx is also a major contributor to
secondary PM2.5 formation.
7.1.1.4	Sulfur Oxides
Sulfur dioxide (SO2), a member of the sulfur oxide (SOx) family of gases, is formed from
burning fuels containing sulfur (e.g., coal or oil), extracting gasoline from oil, or extracting
7-2
-------
metals from ore. SO2 and its gas phase oxidation products can dissolve in water droplets and
further oxidize to form sulfuric acid which reacts with ammonia to form sulfates, which are
important components of ambient PM.
7.1.1.5	Carbon Monoxide
Carbon monoxide (CO) is a colorless, odorless gas emitted from combustion processes.
Nationally, particularly in urban areas, the majority of CO emissions to ambient air come from
mobile sources.5
7.1.1.6	Air Toxics
Light-duty engine emissions contribute to ambient levels of air toxics that are known or
suspected human or animal carcinogens, or that have noncancer health effects. These compounds
include, but are not limited to, benzene, formaldehyde, acetaldehyde, naphthalene, and 1,3-
butadiene. These compounds were identified as national or regional risk drivers or contributors
in the 2014 National-scale Air Toxics Assessment and have significant inventory contributions
from mobile sources.6'7
7.1.2 Health Effects Associated with Exposure to Non-GHG Pollutants
7.1.2.1 Particulate Matter
Scientific evidence spanning animal toxicological, controlled human exposure, and
epidemiologic studies shows that exposure to ambient PM is associated with a broad range of
health effects. These health effects are discussed in detail in the Integrated Science Assessment
for Particulate Matter (PM ISA), which was finalized in December 2019.8 The PM ISA
characterizes the causal nature of relationships between PM exposure and broad health categories
(e.g., cardiovascular effects, respiratory effects, etc.) using a weight-of-evidence approach.b>9
Within this characterization, the PM ISA summarizes the health effects evidence for short- and
long-term exposures to PM2.5, PM10-2.5, and ultrafine particles, and concludes that human
exposures to ambient PM2.5 are associated with a number of adverse health effects. The
discussion below highlights the PM ISA's conclusions pertaining to the health effects evidence
for both short- and long-term PM exposures. Further discussion of PM-related health effects can
also be found in the 2020 Policy Assessment for the review of the PM NAAQS.10
EPA has concluded that recent evidence in combination with evidence evaluated in the 2009
PM ISA supports a "causal relationship" between both long- and short-term exposures to PM2.5
and premature mortality and cardiovascular effects and a "likely to be causal relationship"
between long- and short-term PM2.5 exposures and respiratory effects.11 Additionally, recent
experimental and epidemiologic studies provide evidence supporting a "likely to be causal
b The causal framework draws upon the assessment and integration of evidence from across scientific disciplines,
spanning atmospheric chemistry, exposure, dosimetry and health effects studies (i.e., epidemiologic, controlled
human exposure, and animal toxicological studies), and assess the related uncertainties and limitations that
ultimately influence our understanding of the evidence. This framework employs a five-level hierarchy that
classifies the overall weight-of-evidence with respect to the causal nature of relationships between criteria pollutant
exposures and health and welfare effects using the following categorizations: causal relationship; likely to be causal
relationship; suggestive of, but not sufficient to infer, a causal relationship; inadequate to infer the presence or
absence of a causal relationship; and not likely to be a causal relationship.
7-3
-------
relationship" between long-term PM2.5 exposure and nervous system effects, and long-term
PM2.5 exposure and cancer. In contrast, EPA determined that the more limited and uncertain
evidence was "suggestive of, but not sufficient to infer, a causal relationship" for long-term
PM2.5 exposure and reproductive and developmental effects (i.e., male/female reproduction and
fertility; pregnancy and birth outcomes), long- and short-term exposures and metabolic effects,
and short-term exposure and nervous system effects.
As discussed extensively in the 2019 PM ISA, recent studies continue to support and extend
the evidence base linking short- and long-term PM2.5 exposures and mortality.8 For short-term
PM2.5 exposure, recent multi-city studies, in combination with single- and multi-city studies
evaluated in the 2009 PM ISA, provide evidence of consistent, positive associations across
studies conducted in different geographic locations, populations with different demographic
characteristics, and studies using different exposure assignment techniques. Additionally, the
consistent and coherent evidence across scientific disciplines for cardiovascular morbidity,
particularly ischemic events and heart failure, and to a lesser degree for respiratory morbidity,
with the strongest evidence for exacerbations of chronic obstructive pulmonary disease (COPD)
and asthma, provide biological plausibility for cause-specific mortality and ultimately total
mortality.
In addition to re-analyses and extensions of the American Cancer Society (ACS) and Harvard
Six Cities (HSC) cohorts, multiple new cohort studies conducted in the U.S. and Canada
consisting of people employed in a specific job (e.g., teacher, nurse), and that apply different
exposure assignment techniques provide evidence of positive associations between long-term
PM2.5 exposure and mortality. Biological plausibility for mortality due to long-term PM2.5
exposure is provided by the coherence of effects across scientific disciplines for cardiovascular
morbidity, particularly for coronary heart disease (CHD), stroke and atherosclerosis, and for
respiratory morbidity, particularly for the development of COPD. Additionally, recent studies
provide evidence indicating that as long-term PM2.5 concentrations decrease there is an increase
in life expectancy.
A large body of recent studies examining both short- and long-term PM2.5 exposure and
cardiovascular effects supports and extends the evidence base evaluated in the 2009 PM ISA.
Some of the strongest evidence from both experimental and epidemiologic studies examining
short-term PM2.5 exposures are for ischemic heart disease (IHD) and heart failure. The evidence
for cardiovascular effects is coherent across studies of short-term PM2.5 exposure that have
observed associations with a continuum of effects ranging from subtle changes in indicators of
cardiovascular health to serious clinical events, such as increased emergency department visits
and hospital admissions due to cardiovascular disease and cardiovascular mortality. For long-
term PM2.5 exposure, there is strong and consistent epidemiologic evidence of a relationship with
cardiovascular mortality. This evidence is supported by epidemiologic and animal toxicological
studies demonstrating a range of cardiovascular effects including coronary heart disease, stroke,
impaired heart function, and subclinical markers (e.g., coronary artery calcification,
atherosclerotic plaque progression), which collectively provide coherence and biological
plausibility.
Recent studies continue to provide evidence of a relationship between both short- and long-
term PM2.5 exposure and respiratory effects. Epidemiologic and animal toxicological studies
examining short-term PM2.5 exposure provide consistent evidence of asthma and COPD
7-4
-------
exacerbations, in children and adults, respectively. This evidence is supported by epidemiologic
studies examining asthma and COPD emergency department visits and hospital admissions, as
well as, respiratory mortality. However, there is inconsistent evidence of respiratory effects,
specifically lung function declines and pulmonary inflammation, in controlled human exposure
studies. Epidemiologic studies conducted in the U.S. and abroad provide evidence of a
relationship between long-term PM2.5 exposure and respiratory effects, including consistent
changes in lung function and lung function growth rate, increased asthma incidence, asthma
prevalence, and wheeze in children; acceleration of lung function decline in adults; and
respiratory mortality. The epidemiologic evidence is supported by animal toxicological studies,
which provide coherence and biological plausibility for a range of effects including impaired
lung development, decrements in lung function growth, and asthma development.
Since the 2009 PM ISA, a growing body of scientific evidence examined the relationship
between long-term PM2.5 exposure and nervous system effects, resulting for the first time in a
causality determination for this health effects category. The strongest evidence for effects on the
nervous system come from epidemiologic studies that consistently report cognitive decrements
and reductions in brain volume in adults. The effects observed in epidemiologic studies are
supported by animal toxicological studies demonstrating effects on the brain of adult animals
including inflammation, morphologic changes, and neurodegeneration of specific regions of the
brain. There is more limited evidence for neurodevelopmental effects in children with some
studies reporting positive associations with autism spectrum disorder (ASD) and others
providing limited evidence of an association with cognitive function. While there is some
evidence from animal toxicological studies indicating effects on the brain (i.e., inflammatory and
morphological changes) to support a biologically plausible pathway, epidemiologic studies of
neurodevelopmental effects are limited due to their lack of control for potential confounding by
copollutants, the small number of studies conducted, and uncertainty regarding critical exposure
windows.
Building off the decades of research demonstrating mutagenicity, DNA damage, and
endpoints related to genotoxicity due to whole PM exposures, recent experimental and
epidemiologic studies focusing specifically on PM2.5 provide evidence of a relationship between
long-term PM2.5 exposure and cancer. Epidemiologic studies examining long-term PM2.5
exposure and lung cancer incidence and mortality provide evidence of generally positive
associations in cohort studies spanning different populations, locations, and exposure assignment
techniques. Additionally, there is evidence of positive associations in analyses limited to never
smokers. The epidemiologic evidence is supported by both experimental and epidemiologic
evidence of genotoxicity, epigenetic effects, carcinogenic potential, and that PM2.5 exhibits
several characteristics of carcinogens, which collectively provides biological plausibility for
cancer development.
For the additional health effects categories evaluated for PM2.5 in the 2019 PM ISA,
experimental and epidemiologic studies provide limited and/or inconsistent evidence of a
relationship with PM2.5 exposure. As a result, the 2019 PM ISA concluded that the evidence is
"suggestive of, but not sufficient to infer a causal relationship" for short-term PM2.5 exposure
and metabolic effects and nervous system effects, and long-term PM2.5 exposures and metabolic
effects as well as reproductive and developmental effects.
7-5
-------
In addition to evaluating the health effects attributed to short- and long-term exposure to
PM2.5, the 2019 PM ISA also conducted an extensive evaluation as to whether specific
components or sources of PM2.5 are more strongly related with health effects than PM2.5 mass.
An evaluation of those studies resulted in the 2019 PM ISA concluding that "many PM2.5
components and sources are associated with many health effects, and the evidence does not
indicate that any one source or component is consistently more strongly related to health effects
than PM2.5 mass."8
For both PM10-2.5 and UFPs, for all health effects categories evaluated, the 2019 PM ISA
concluded that the evidence was "suggestive of, but not sufficient to infer, a causal relationship"
or "inadequate to determine the presence or absence of a causal relationship." For PM10-2.5,
although a Federal Reference Method (FRM) was instituted in 2011 to measure PM10-2.5
concentrations nationally, the causality determinations reflect that the same uncertainty identified
in the 2009 PM ISA with respect to the method used to estimate PM10-2.5 concentrations in
epidemiologic studies persists. Specifically, across epidemiologic studies, different approaches
are used to estimate PM10-2.5 concentrations (e.g., direct measurement of PM10-2.5, difference
between PM10 and PM2.5 concentrations), and it remains unclear how well correlated PM10-2.5
concentrations are both spatially and temporally across the different methods used.
For UFPs, the uncertainty in the evidence for the health effect categories evaluated across
experimental and epidemiologic studies reflects the inconsistency in the exposure metric used
(i.e., particle number concentration, surface area concentration, mass concentration) as well as
the size fractions examined. In epidemiologic studies the size fraction can vary depending on the
monitor used and exposure metric, with some studies examining number count over the entire
particle size range, while experimental studies that use a particle concentrator often examine
particles up to 0.3 |im. Additionally, due to the lack of a monitoring network, there is limited
information on the spatial and temporal variability of UFPs within the U.S., as well as population
exposures to UFPs, which adds uncertainty to epidemiologic study results.
The 2019 PM ISA cites extensive evidence indicating that "both the general population as
well as specific populations and lifestages are at risk for PM2.5-related health effects"8'10 For
example, in support of its "causal" and "likely to be causal" determinations, the ISA cites
substantial evidence for (1) PM-related mortality and cardiovascular effects in older adults; (2)
PM-related cardiovascular effects in people with pre-existing cardiovascular disease; (3) PM-
related respiratory effects in people with pre-existing respiratory disease, particularly asthma
exacerbations in children; and (4) PM-related impairments in lung function growth and asthma
development in children. The ISA additionally notes that stratified analyses (i.e., analyses that
directly compare PM-related health effects across groups) provide strong evidence for racial and
ethnic differences in PM2.5 exposures and in the risk of PM2.5-related health effects, specifically
within Hispanic and non-Hispanic black populations. Additionally, evidence spanning
epidemiologic studies that conducted stratified analyses, experimental studies focusing on animal
models of disease or individuals with pre-existing disease, dosimetry studies, as well as studies
focusing on differential exposure suggest that populations with pre-existing cardiovascular or
respiratory disease, populations that are overweight or obese, populations that have particular
genetic variants, populations that are of low socioeconomic status, and current/former smokers
could be at increased risk for adverse PM2.5-related health effects.
7-6
-------
7.1.2.2 Ozone
This section provides a summary of the health effects associated with exposure to ambient
concentrations of ozone.12 The information in this section is based on the information and
conclusions in the April 2020 Integrated Science Assessment for Ozone (Ozone ISA).13 The
Ozone ISA concludes that human exposures to ambient concentrations of ozone are associated
with a number of adverse health effects and characterizes the weight of evidence for these health
effects.14 The discussion below highlights the Ozone ISA's conclusions pertaining to health
effects associated with both short-term and long-term periods of exposure to ozone.
For short-term exposure to ozone, the Ozone ISA concludes that respiratory effects, including
lung function decrements, pulmonary inflammation, exacerbation of asthma, respiratory-related
hospital admissions, and mortality, are causally associated with ozone exposure. It also
concludes that metabolic effects, including metabolic syndrome (i.e., changes in insulin or
glucose levels, cholesterol levels, obesity and blood pressure) and complications due to diabetes
are likely to be causally associated with short-term exposure to ozone and that evidence is
suggestive of a causal relationship between cardiovascular effects, central nervous system effects
and total mortality and short-term exposure to ozone.
For long-term exposure to ozone, the Ozone ISA concludes that respiratory effects, including
new onset asthma, pulmonary inflammation and injury, are likely to be causally related with
ozone exposure. The Ozone ISA characterizes the evidence as suggestive of a causal relationship
for associations between long-term ozone exposure and cardiovascular effects, metabolic effects,
reproductive and developmental effects, central nervous system effects and total mortality. The
evidence is inadequate to infer a causal relationship between chronic ozone exposure and
increased risk of cancer.
This section provides a summary of the health effects associated with exposure to ambient
concentrations of ozone.c The information in this section is based on the information and
conclusions in the April 2020 Integrated Science Assessment for Ozone (Ozone ISA).15 The
Ozone ISA concludes that human exposures to ambient concentrations of ozone are associated
with a number of adverse health effects and characterizes the weight of evidence for these health
effects.d The discussion below highlights the Ozone ISA's conclusions pertaining to health
effects associated with both short-term and long-term periods of exposure to ozone.
For short-term exposure to ozone, the Ozone ISA concludes that respiratory effects, including
lung function decrements, pulmonary inflammation, exacerbation of asthma, respiratory-related
hospital admissions, and mortality, are causally associated with ozone exposure. It also
concludes that metabolic effects, including metabolic syndrome (i.e., changes in insulin or
glucose levels, cholesterol levels, obesity and blood pressure) and complications due to diabetes
c Human exposure to ozone varies over time due to changes in ambient ozone concentration and because people
move between locations which have notably different ozone concentrations. Also, the amount of ozone delivered to
the lung is influenced not only by the ambient concentrations but also by the breathing route and rate.
d The ISA evaluates evidence and draws conclusions on the causal relationship between relevant pollutant exposures
and health effects, assigning one of five "weight of evidence" determinations: causal relationship, likely to be a
causal relationship, suggestive of a causal relationship, inadequate to infer a causal relationship, and not likely to be
a causal relationship. For more information on these levels of evidence, please refer to Table II in the Preamble of
the ISA.
7-7
-------
are likely to be causally associated with short-term exposure to ozone and that evidence is
suggestive of a causal relationship between cardiovascular effects, central nervous system effects
and total mortality and short-term exposure to ozone.
For long-term exposure to ozone, the Ozone ISA concludes that respiratory effects, including
new onset asthma, pulmonary inflammation and injury, are likely to be causally related with
ozone exposure. The Ozone ISA characterizes the evidence as suggestive of a causal relationship
for associations between long-term ozone exposure and cardiovascular effects, metabolic effects,
reproductive and developmental effects, central nervous system effects and total mortality. The
evidence is inadequate to infer a causal relationship between chronic ozone exposure and
increased risk of cancer.
Finally, interindividual variation in human responses to ozone exposure can result in some
groups being at increased risk for detrimental effects in response to exposure. In addition, some
groups are at increased risk of exposure due to their activities, such as outdoor workers and
children. The Ozone ISA identified several groups that are at increased risk for ozone-related
health effects. These groups are people with asthma, children and older adults, individuals with
reduced intake of certain nutrients (i.e., Vitamins C and E), outdoor workers, and individuals
having certain genetic variants related to oxidative metabolism or inflammation. Ozone exposure
during childhood can have lasting effects through adulthood. Such effects include altered
function of the respiratory and immune systems. Children absorb higher doses (normalized to
lung surface area) of ambient ozone, compared to adults, due to their increased time spent
outdoors, higher ventilation rates relative to body size, and a tendency to breathe a greater
fraction of air through the mouth. Children also have a higher asthma prevalence compared to
adults. Recent epidemiologic studies provide generally consistent evidence that long-term ozone
exposure is associated with the development of asthma in children. Studies comparing age
groups reported higher magnitude associations for short-term ozone exposure and respiratory
hospital admissions and emergency room visits among children than for adults. Panel studies
also provide support for experimental studies with consistent associations between short-term
ozone exposure and lung function and pulmonary inflammation in healthy children.
7.1.2.3 Nitrogen Oxides
The most recent review of the health effects of oxides of nitrogen completed by EPA can be
found in the 2016 Integrated Science Assessment for Oxides of Nitrogen - Health Criteria
(Oxides of Nitrogen ISA).16 The primary source of NO2 is motor vehicle emissions, and ambient
NO2 concentrations tend to be highly correlated with other traffic-related pollutants. Thus, a key
issue in characterizing the causality of N02-health effect relationships consists of evaluating the
extent to which studies supported an effect of NO2 that is independent of other traffic-related
pollutants. EPA concluded that the findings for asthma exacerbation integrated from
epidemiologic and controlled human exposure studies provided evidence that is sufficient to
infer a causal relationship between respiratory effects and short-term NO2 exposure. The
strongest evidence supporting an independent effect of NO2 exposure comes from controlled
human exposure studies demonstrating increased airway responsiveness in individuals with
asthma following ambient-relevant NO2 exposures. The coherence of this evidence with
epidemiologic findings for asthma hospital admissions and ED visits as well as lung function
decrements and increased pulmonary inflammation in children with asthma describe a plausible
pathway by which NO2 exposure can cause an asthma exacerbation. The 2016 ISA for Oxides of
7-8
-------
Nitrogen also concluded that there is likely to be a causal relationship between long-term NO2
exposure and respiratory effects. This conclusion is based on new epidemiologic evidence for
associations of NO2 with asthma development in children combined with biological plausibility
from experimental studies.
In evaluating a broader range of health effects, the 2016 ISA for Oxides of Nitrogen
concluded that evidence is "suggestive of, but not sufficient to infer, a causal relationship"
between short-term NO2 exposure and cardiovascular effects and mortality and between long-
term NO2 exposure and cardiovascular effects and diabetes, birth outcomes, and cancer. In
addition, the scientific evidence is inadequate (insufficient consistency of epidemiologic and
toxicological evidence) to infer a causal relationship for long-term NO2 exposure with fertility,
reproduction, and pregnancy, as well as with postnatal development. A key uncertainty in
understanding the relationship between these non-respiratory health effects and short- or long-
term exposure to NO2 is co-pollutant confounding, particularly by other roadway pollutants. The
available evidence for non-respiratory health effects does not adequately address whether NO2
has an independent effect or whether it primarily represents effects related to other or a mixture
of traffic-related pollutants.
The 2016 ISA for Oxides of Nitrogen concluded that people with asthma, children, and older
adults are at increased risk for N02-related health effects. In these groups and lifestages, NO2 is
consistently related to larger effects on outcomes related to asthma exacerbation, for which there
is confidence in the relationship with NO2 exposure.
7.1.2.4 Sulfur Oxides
This section provides an overview of the health effects associated with SO2. Additional
information on the health effects of SO2 can be found in the 2017 Integrated Science Assessment
for Sulfur Oxides - Health Criteria (SOx ISA).17 Following an extensive evaluation of health
evidence from animal toxicological, controlled human exposure, and epidemiologic studies, the
EPA has concluded that there is a causal relationship between respiratory health effects and
short-term exposure to SO2. The immediate effect of SO2 on the respiratory system in humans is
bronchoconstriction. People with asthma are more sensitive to the effects of SO2, likely resulting
from preexisting inflammation associated with this disease. In addition to those with asthma
(both children and adults), there is suggestive evidence that all children and older adults may be
at increased risk of S02-related health effects. In free-breathing laboratory studies involving
controlled human exposures to SO2, respiratory effects have consistently been observed
following 5-10 min exposures at SO2 concentrations > 400 ppb in people with asthma engaged in
moderate to heavy levels of exercise, with respiratory effects occurring at concentrations as low
as 200 ppb in some individuals with asthma. A clear concentration-response relationship has
been demonstrated in these studies following exposures to SO2 at concentrations between 200
and 1000 ppb, both in terms of increasing severity of respiratory symptoms and decrements in
lung function, as well as the percentage of individuals with asthma adversely affected.
Epidemiologic studies have reported positive associations between short-term ambient SO2
concentrations and hospital admissions and emergency department visits for asthma and for all
respiratory causes, particularly among children and older adults (> 65 years). The studies provide
supportive evidence for the causal relationship.
For long-term SO2 exposure and respiratory effects, the EPA has concluded that the evidence
is suggestive of a causal relationship. This conclusion is based on new epidemiologic evidence
7-9
-------
for positive associations between long-term SO2 exposure and increases in asthma incidence
among children, together with animal toxicological evidence that provides a pathophysiologic
basis for the development of asthma. However, uncertainty remains regarding the influence of
other pollutants on the observed associations with SO2 because these epidemiologic studies have
not examined the potential for co-pollutant confounding.
Consistent associations between short-term exposure to SO2 and mortality have been observed
in epidemiologic studies, with larger effect estimates reported for respiratory mortality than for
cardiovascular mortality. While this finding is consistent with the demonstrated effects of SO2 on
respiratory morbidity, uncertainty remains with respect to the interpretation of these observed
mortality associations due to potential confounding by various co-pollutants. Therefore, the EPA
has concluded that the overall evidence is suggestive of a causal relationship between short-term
exposure to SO2 and mortality.
7.1.2.5 Carbon Monoxide
Information on the health effects of carbon monoxide (CO) can be found in the January 2010
Integrated Science Assessment for Carbon Monoxide (CO ISA).18 The CO ISA presents
conclusions regarding the presence of causal relationships between CO exposure and categories
of adverse health effects.19 This section provides a summary of the health effects associated with
exposure to ambient concentrations of CO, along with the CO ISA conclusions.20
Controlled human exposure studies of subjects with coronary artery disease show a decrease
in the time to onset of exercise-induced angina (chest pain) and electrocardiogram changes
following CO exposure. In addition, epidemiologic studies observed associations between short-
term CO exposure and cardiovascular morbidity, particularly increased emergency room visits
and hospital admissions for coronary heart disease (including ischemic heart disease, myocardial
infarction, and angina). Some epidemiologic evidence is also available for increased hospital
admissions and emergency room visits for congestive heart failure and cardiovascular disease as
a whole. The CO ISA concludes that a causal relationship is likely to exist between short-term
exposures to CO and cardiovascular morbidity. It also concludes that available data are
inadequate to conclude that a causal relationship exists between long-term exposures to CO and
cardiovascular morbidity.
Animal studies show various neurological effects with in-utero CO exposure. Controlled
human exposure studies report central nervous system and behavioral effects following low-level
CO exposures, although the findings have not been consistent across all studies. The CO ISA
concludes that the evidence is suggestive of a causal relationship with both short- and long-term
exposure to CO and central nervous system effects.
A number of studies cited in the CO ISA have evaluated the role of CO exposure in birth
outcomes such as preterm birth or cardiac birth defects. There is limited epidemiologic evidence
of a CO-induced effect on preterm births and birth defects, with weak evidence for a decrease in
birth weight. Animal toxicological studies have found perinatal CO exposure to affect birth
weight, as well as other developmental outcomes. The CO ISA concludes that the evidence is
suggestive of a causal relationship between long-term exposures to CO and developmental
effects and birth outcomes.
7-10
-------
Epidemiologic studies provide evidence of associations between short-term CO
concentrations and respiratory morbidity such as changes in pulmonary function, respiratory
symptoms, and hospital admissions. A limited number of epidemiologic studies considered
copollutants such as ozone, SO2, and PM in two-pollutant models and found that CO risk
estimates were generally robust, although this limited evidence makes it difficult to disentangle
effects attributed to CO itself from those of the larger complex air pollution mixture. Controlled
human exposure studies have not extensively evaluated the effect of CO on respiratory
morbidity. Animal studies at levels of 50-100 ppm CO show preliminary evidence of altered
pulmonary vascular remodeling and oxidative injury. The CO ISA concludes that the evidence is
suggestive of a causal relationship between short-term CO exposure and respiratory morbidity,
and inadequate to conclude that a causal relationship exists between long-term exposure and
respiratory morbidity.
Finally, the CO ISA concludes that the epidemiologic evidence is suggestive of a causal
relationship between short-term concentrations of CO and mortality. Epidemiologic evidence
suggests an association exists between short-term exposure to CO and mortality, but limited
evidence is available to evaluate cause-specific mortality outcomes associated with CO exposure.
In addition, the attenuation of CO risk estimates which was often observed in co-pollutant
models contributes to the uncertainty as to whether CO is acting alone or as an indicator for other
combustion-related pollutants. The CO ISA also concludes that there is not likely to be a causal
relationship between relevant long-term exposures to CO and mortality.
7.1.2.6 Air Toxics
7.1.2.6.1 Health Effects Associated with Exposure to Benzene
EPA's Integrated Risk Information System (IRIS) database lists benzene as a known human
carcinogen (causing leukemia) by all routes of exposure, and concludes that exposure is
associated with additional health effects, including genetic changes in both humans and animals
and increased proliferation of bone marrow cells in mice.21'22'23 EPA states in its IRIS database
that data indicate a causal relationship between benzene exposure and acute lymphocytic
leukemia and suggest a relationship between benzene exposure and chronic non-lymphocytic
leukemia and chronic lymphocytic leukemia. EPA's IRIS documentation for benzene also lists a
range of 2.2 x 10-6 to 7.8 x 10-6 per |ig/m3 as the unit risk estimate (URE) for benzene.24'25 The
International Agency for Research on Cancer (IARC) has determined that benzene is a human
carcinogen, and the U.S. Department of Health and Human Services (DHHS) has characterized
benzene as a known human carcinogen.26'27
A number of adverse noncancer health effects, including blood disorders such as preleukemia
and aplastic anemia, have also been associated with long-term exposure to benzene.28'29 The
most sensitive noncancer effect observed in humans, based on current data, is the depression of
the absolute lymphocyte count in blood.30'31 EPA's inhalation reference concentration (RfC) for
benzene is 30 |ig/m3. The RfC is based on suppressed absolute lymphocyte counts seen in
humans under occupational exposure conditions. In addition, studies sponsored by the Health
Effects Institute (HEI) provide evidence that biochemical responses occur at lower levels of
benzene exposure than previously known.32'33'34'35 EPA's IRIS program has not yet evaluated
these new data. EPA does not currently have an acute reference concentration for benzene. The
Agency for Toxic Substances and Disease Registry (ATSDR) Minimal Risk Level (MRL) for
acute exposure to benzene is 29 |ig/m3 for 1-14 days exposure.36'37
7-11
-------
7.1.2.6.2 Health Effects Associated with Exposure to Formaldehyde
In 1991, EPA concluded that formaldehyde is a Class B1 probable human carcinogen based
on limited evidence in humans and sufficient evidence in animals.38 An Inhalation URE for
cancer and a Reference Dose for oral noncancer effects were developed by EPA and posted on
the IRIS database. Since that time, the NTP and IARC have concluded that formaldehyde is a
known human carcinogen.39'40'41
The conclusions by IARC and NTP reflect the results of epidemiologic research published
since 1991 in combination with previous animal, human and mechanistic evidence. Research
conducted by the National Cancer Institute reported an increased risk of nasopharyngeal cancer
and specific lymphohematopoietic malignancies among workers exposed to formaldehyde.42'43'44
A National Institute of Occupational Safety and Health study of garment workers also reported
increased risk of death due to leukemia among workers exposed to formaldehyde.45 Extended
follow-up of a cohort of British chemical workers did not report evidence of an increase in
nasopharyngeal or lymphohematopoietic cancers, but a continuing statistically significant excess
in lung cancers was reported.46 Finally, a study of embalmers reported formaldehyde exposures
to be associated with an increased risk of myeloid leukemia but not brain cancer.47
Health effects of formaldehyde in addition to cancer were reviewed by the Agency for Toxics
Substances and Disease Registry in 1999, supplemented in 2010, and by the World Health
Organization. 48>49>50 These organizations reviewed the scientific literature concerning health
effects linked to formaldehyde exposure to evaluate hazards and dose response relationships and
defined exposure concentrations for minimal risk levels (MRLs). The health endpoints reviewed
included sensory irritation of eyes and respiratory tract, reduced pulmonary function, nasal
histopathology, and immune system effects. In addition, research on reproductive and
developmental effects and neurological effects were discussed along with several studies that
suggest that formaldehyde may increase the risk of asthma - particularly in the young.
In June 2010, EPA released a draft Toxicological Review of Formaldehyde - Inhalation
Assessment through the IRIS program for peer review by the National Research Council (NRC)
and public comment.51 That draft assessment reviewed more recent research from animal and
human studies on cancer and other health effects. The NRC released their review report in April
2011.52	EPA's draft assessment, which addresses NRC recommendations, was suspended in
2018.53	The draft assessment was unsuspended in March 2021.
7.1.2.6.3 Health Effects Associated with Exposure to Acetaldehyde
Acetaldehyde is classified in EPA's IRIS database as a probable human carcinogen, based on
nasal tumors in rats, and is considered toxic by the inhalation, oral, and intravenous routes.54
The URE in IRIS for acetaldehyde is 2.2 x 10-6 per |ig/m3.55 Acetaldehyde is reasonably
anticipated to be a human carcinogen by the NTP in the 14th Report on Carcinogens and is
classified as possibly carcinogenic to humans (Group 2B) by the IARC.56'57
The primary noncancer effects of exposure to acetaldehyde vapors include irritation of the
eyes, skin, and respiratory tract.58 In short-term (4 week) rat studies, degeneration of olfactory
epithelium was observed at various concentration levels of acetaldehyde exposure.59'60 Data from
these studies were used by EPA to develop an inhalation reference concentration of 9 |ig/m3.
Some asthmatics have been shown to be a sensitive subpopulation to decrements in functional
7-12
-------
expiratory volume (FEV1 test) and bronchoconstriction upon acetaldehyde inhalation.61
Children, especially those with diagnosed asthma, may be more likely to show impaired
pulmonary function and symptoms of asthma than are adults following exposure to
acetaldehyde.62
7.1.2.6.4	Health Effects Associated with Exposure to Naphthalene
Naphthalene is found in small quantities in gasoline and diesel fuels. Naphthalene emissions
have been measured in larger quantities in both gasoline and diesel exhaust compared with
evaporative emissions from mobile sources, indicating it is primarily a product of combustion.
Acute (short-term) exposure of humans to naphthalene by inhalation, ingestion, or dermal
contact is associated with hemolytic anemia and damage to the liver and the nervous system.63
Chronic (long term) exposure of workers and rodents to naphthalene has been reported to cause
cataracts and retinal damage.64 EPA released an external review draft of a reassessment of the
inhalation carcinogenicity of naphthalene based on a number of recent animal carcinogenicity
studies.65 The draft reassessment completed external peer review.66 Based on external peer
review comments received, EPA was developing a revised draft assessment that considers all
routes of exposure, as well as cancer and noncancer effects; this reassessment was suspended in
2018.67 The external review draft does not represent official agency opinion and was released
solely for the purposes of external peer review and public comment. The NTP listed naphthalene
as "reasonably anticipated to be a human carcinogen" in 2004 on the basis of bioassays reporting
clear evidence of carcinogenicity in rats and some evidence of carcinogenicity in mice.68
California EPA has released a new risk assessment for naphthalene, and the IARC has
reevaluated naphthalene and re-classified it as Group 2B: possibly carcinogenic to humans.69
Naphthalene also causes a number of chronic non-cancer effects in animals, including
abnormal cell changes and growth in respiratory and nasal tissues.70 The current EPA IRIS
assessment includes noncancer data on hyperplasia and metaplasia in nasal tissue that form the
basis of the inhalation RfC of 3 |ig/m3.71 The ATSDR MRL for acute exposure to naphthalene is
0.6 mg/kg/day.
7.1.2.6.5	Health Effects Associated with Exposure to 1,3-Butadiene
EPA has characterized 1,3-butadiene as carcinogenic to humans by inhalation.72'73 The IARC
has determined that 1,3-butadiene is a human carcinogen and the U.S. DHHS has characterized
1,3-butadiene as a known human carcinogen.74'75'76' 77 There are numerous studies consistently
demonstrating that 1,3-butadiene is metabolized into genotoxic metabolites by experimental
animals and humans. The specific mechanisms of 1,3-butadiene-induced carcinogenesis are
unknown; however, the scientific evidence strongly suggests that the carcinogenic effects are
mediated by genotoxic metabolites. Animal data suggest that females may be more sensitive than
males for cancer effects associated with 1,3-butadiene exposure; there are insufficient data in
humans from which to draw conclusions about sensitive subpopulations. The URE for 1,3-
butadiene is 3 x 10-5 per |ig/m3.78 1,3-butadiene also causes a variety of reproductive and
developmental effects in mice; no human data on these effects are available. The most sensitive
effect was ovarian atrophy observed in a lifetime bioassay of female mice.79 Based on this
critical effect and the benchmark concentration methodology, an RfC for chronic health effects
was calculated at 0.9 ppb (approximately 2 |ig/m3).
7-13
-------
7.1.2.6.6 Health effects Associated with exposure to other air toxics
In addition to the compounds described above, other compounds found in gaseous
hydrocarbon and PM emissions from engines will be affected by this rulemaking. Mobile source
air toxic compounds that will potentially be affected include acrolein, ethylbenzene,
propionaldehyde, toluene, and xylene. Information regarding the health effects of these
compounds can be found in EPA's IRIS database.80
7.1.2.7 Exposure and Health Effects Associated with Traffic
Locations in close proximity to major roadways generally have elevated concentrations of
many air pollutants emitted from motor vehicles. Hundreds of such studies have been published
in peer-reviewed journals, concluding that concentrations of CO, CO2, NO, NO2, benzene,
aldehydes, particulate matter, black carbon, and many other compounds are elevated in ambient
air within approximately 300-600 meters (about 1,000-2,000 feet) of major roadways. The
highest concentrations of most pollutants emitted directly by motor vehicles are found at
locations within 50 meters (about 165 feet) of the edge of a roadway's traffic lanes.
A large-scale review of air quality measurements in the vicinity of major roadways between
1978 and 2008 concluded that the pollutants with the steepest concentration gradients in
vicinities of roadways were CO, ultrafine particles, metals, elemental carbon (EC), NO, NOx,
and several VOCs.81 These pollutants showed a large reduction in concentrations within 100
meters downwind of the roadway. Pollutants that showed more gradual reductions with distance
from roadways included benzene, NO2, PM2.5, and PM10. In the review article, results varied
based on the method of statistical analysis used to determine the gradient in concentration. More
recent studies continue to show significant concentration gradients of traffic-related air pollution
around major roads 82>83>84>85>86;87-88-89 There is evidence that EPA's regulations for vehicles have
lowered the near-road concentrations and gradients.90
For pollutants with relatively high background concentrations relative to near-road
concentrations, detecting concentration gradients can be difficult. For example, many aldehydes
have high background concentrations as a result of photochemical breakdown of precursors from
many different organic compounds. However, several studies have measured aldehydes in
multiple weather conditions and found higher concentrations of many carbonyls downwind of
roadways.91'92 These findings suggest a substantial roadway source of these carbonyls.
In the past 20 years, many studies have been published with results reporting that populations
who live, work, or go to school near high-traffic roadways experience higher rates of numerous
adverse health effects, compared to populations far away from major roads.93 In addition,
numerous studies have found adverse health effects associated with spending time in traffic, such
as commuting or walking along high-traffic roadways.94'95'96'97 The health outcomes with the
strongest evidence linking them with traffic-associated air pollutants are respiratory effects,
particularly in asthmatic children, and cardiovascular effects.
Numerous reviews of this body of health literature have been published as well. In 2010, an
expert panel of the Health Effects Institute (HEI) published a review of hundreds of exposure,
epidemiology, and toxicology studies.98 The panel rated how the evidence for each type of
health outcome supported a conclusion of a causal association with traffic-associated air
pollution as either "sufficient," "suggestive but not sufficient," or "inadequate and insufficient."
7-14
-------
The panel categorized evidence of a causal association for exacerbation of childhood asthma as
"sufficient." The panel categorized evidence of a causal association for new onset asthma as
between "sufficient" and "suggestive but not sufficient." "Suggestive of a causal association"
was how the panel categorized evidence linking traffic-associated air pollutants with
exacerbation of adult respiratory symptoms and lung function decrement. It categorized as
"inadequate and insufficient" evidence of a causal relationship between traffic-related air
pollution and health care utilization for respiratory problems, new onset adult asthma, chronic
obstructive pulmonary disease (COPD), nonasthmatic respiratory allergy, and cancer in adults
and children. Currently, HEI is conducting another expert review of health studies associated
with traffic-related air pollution published after the studies included in their 2010 review." Other
literature reviews have been published with conclusions generally similar to the 2010 HEI
panel's.100,101'102'103 However, in 2014, researchers from the U.S. Centers for Disease Control
and Prevention (CDC) published a systematic review and meta-analysis of studies evaluating the
risk of childhood leukemia associated with traffic exposure and reported positive associations
between "postnatal" proximity to traffic and leukemia risks, but no such association for
"prenatal" exposures.104 The U.S. Department of Health and Human Services' National
Toxicology Program (NTP) recently published a monograph including a systematic review of
traffic-related air pollution (TRAP) and its impacts on hypertensive disorders of pregnancy. NTP
concluded that exposure to TRAP is "presumed to be a hazard to pregnant women" for
developing hypertensive disorders of pregnancy.105
Health outcomes with few publications suggest the possibility of other effects still lacking
sufficient evidence to draw definitive conclusions. Among these outcomes with a small number
of positive studies are neurological impacts (e.g., autism and reduced cognitive function) and
reproductive outcomes (e.g., preterm birth, low birth weight).I06-107-108-109
In addition to health outcomes, particularly cardiopulmonary effects, conclusions of numerous
studies suggest mechanisms by which traffic-related air pollution affects health. Numerous
studies indicate that near-roadway exposures may increase systemic inflammation, affecting
organ systems, including blood vessels and lungs 110.in.112.113 Long-term exposures in near-road
environments have been associated with inflammation-associated conditions, such as
atherosclerosis and asthma.114'115'116
Several studies suggest that some factors may increase susceptibility to the effects of traffic-
associated air pollution. Several studies have found stronger respiratory associations in children
experiencing chronic social stress, such as in violent neighborhoods or in homes with high
family stress.117'118'119
The risks associated with residence, workplace, or schools near major roads are of potentially
high public health significance due to the large population in such locations. Every two years
from 1997 to 2009 and in 2011, the U.S. Census Bureau's American Housing Survey (AHS)
conducted a survey that includes whether housing units are within 300 feet of an "airport,
railroad, or highway with four or more lanes."120 The 2013 AHS was the last AHS that included
that question. The 2013 survey reports that 17.3 million housing units, or 13 percent of all
housing units in the U.S., were in such areas. Assuming that populations and housing units are in
the same locations, this corresponds to a population of more than 41 million U.S. residents in
close proximity to high-traffic roadways or other transportation sources. According to the
Central Intelligence Agency's World Factbook, based on data collected between 2012-2014, the
7-15
-------
United States had 6,586,610 km of roadways, 293,564 km of railways, and 13,513 airports. As
such, highways represent the overwhelming majority of transportation facilities described by this
factor in the AHS.
On average, populations near major roads have higher fractions of minority residents and
lower socioeconomic status (see Chapter 8.3) 12U22>123>124>125 Furthermore, on average,
Americans spend more than an hour traveling each day, bringing nearly all residents into a high-
exposure microenvironment for part of the day.126
7.1.3 Environmental Effects Associated with Exposure to Non-GHG Pollutants
7.1.3.1 Visibility
Visibility can be defined as the degree to which the atmosphere is transparent to visible
light.127 Visibility impairment is caused by light scattering and absorption by suspended
particles and gases. It is dominated by contributions from suspended particles except under
pristine conditions. Visibility is important because it has direct significance to people's
enjoyment of daily activities in all parts of the country. Individuals value good visibility for the
well-being it provides them directly, where they live and work, and in places where they enjoy
recreational opportunities. Visibility is also highly valued in significant natural areas, such as
national parks and wilderness areas, and special emphasis is given to protecting visibility in these
areas. For more information on visibility see the final 2019 PM ISA.8
EPA is working to address visibility impairment. Reductions in air pollution from
implementation of various programs associated with the Clean Air Act Amendments of 1990
(CAAA) provisions have resulted in substantial improvements in visibility and will continue to
do so in the future. Because trends in haze are closely associated with trends in particulate sulfate
and nitrate due to the relationship between their concentration and light extinction, visibility
trends have improved as emissions of SO2 and NOx have decreased over time due to air
pollution regulations such as the Acid Rain Program.8
In the Clean Air Act Amendments of 1977, Congress recognized visibility's value to society
by establishing a national goal to protect national parks and wilderness areas from visibility
impairment caused by manmade pollution.128 In 1999, EPA finalized the regional haze program
to protect the visibility in Mandatory Class I Federal areas.129 There are 156 national parks,
forests and wilderness areas categorized as Mandatory Class I Federal areas.130 These areas are
defined in CAA Section 162 as those national parks exceeding 6,000 acres, wilderness areas and
memorial parks exceeding 5,000 acres, and all international parks which were in existence on
August 7, 1977.
EPA has also concluded that PM2.5 causes adverse effects on visibility in other areas that are
not targeted by the Regional Haze Rule, such as urban areas, depending on PM2.5 concentrations
and other factors such as dry chemical composition and relative humidity (i.e., an indicator of the
water composition of the particles). EPA revised the PM2.5 NAAQS in 2012, retained it in 2020,
and established a target level of protection that is expected to be met through attainment of the
existing secondary standards for PM2.5.131
7-16
-------
7.1.3.2 Ozone Effects on Ecosystems
The welfare effects of ozone include effects on ecosystems, which can be observed across a
variety of scales, i.e., subcellular, cellular, leaf, whole plant, population, and ecosystem. Ozone
effects that begin at small spatial scales, such as the leaf of an individual plant, when they occur
at sufficient magnitudes (or to a sufficient degree) can result in effects being propagated along a
continuum to higher and higher levels of biological organization. For example, effects at the
individual plant level, such as altered rates of leaf gas exchange, growth and reproduction, can,
when widespread, result in broad changes in ecosystems, such as productivity, carbon storage,
water cycling, nutrient cycling, and community composition.
Ozone can produce both acute and chronic injury in sensitive plant species depending on the
concentration level and the duration of the exposure.132 In those sensitive species,133 effects from
repeated exposure to ozone throughout the growing season of the plant can tend to accumulate,
so that even relatively low concentrations experienced for a longer duration have the potential to
create chronic stress on vegetation.134'135 Ozone damage to sensitive plant species includes
impaired photosynthesis and visible injury to leaves. The impairment of photosynthesis, the
process by which the plant makes carbohydrates (its source of energy and food), can lead to
reduced crop yields, timber production, and plant productivity and growth. Impaired
photosynthesis can also lead to a reduction in root growth and carbohydrate storage below
ground, resulting in other, more subtle plant and ecosystems impacts.136 These latter impacts
include increased susceptibility of plants to insect attack, disease, harsh weather, interspecies
competition and overall decreased plant vigor. The adverse effects of ozone on areas with
sensitive species could potentially lead to species shifts and loss from the affected ecosystems,137
resulting in a loss or reduction in associated ecosystem goods and services. Additionally, visible
ozone injury to leaves can result in a loss of aesthetic value in areas of special scenic significance
like national parks and wilderness areas and reduced use of sensitive ornamentals in
landscaping.138 In addition to ozone effects on vegetation, newer evidence suggests that ozone
affects interactions between plants and insects by altering chemical signals (e.g., floral scents)
that plants use to communicate to other community members, such as attraction of pollinators.
The Ozone ISA presents more detailed information on how ozone affects vegetation and
ecosystems.13 The Ozone ISA reports causal and likely causal relationships between ozone
exposure and a number of welfare effects and characterizes the weight of evidence for different
effects associated with ozone.6 The ISA concludes that visible foliar injury effects on vegetation,
reduced vegetation growth, reduced plant reproduction, reduced productivity in terrestrial
ecosystems, reduced yield and quality of agricultural crops, alteration of below-ground
biogeochemical cycles, and altered terrestrial community composition are causally associated
with exposure to ozone. It also concludes that increased tree mortality, altered herbivore growth
and reproduction, altered plant-insect signaling, reduced carbon sequestration in terrestrial
ecosystems, and alteration of terrestrial ecosystem water cycling are likely to be causally
associated with exposure to ozone.
e The Ozone ISA evaluates the evidence associated with different ozone related health and welfare effects, assigning
one of five "weight of evidence" determinations: causal relationship, likely to be a causal relationship, suggestive of
a causal relationship, inadequate to infer a causal relationship, and not likely to be a causal relationship. For more
information on these levels of evidence, please refer to Table II of the ISA.
7-17
-------
7.1.3.3 Deposition
The Integrated Science Assessment for Oxides of Nitrogen, Oxides of Sulfur, and Particulate
Matter - Ecological Criteria documents the ecological effects of the deposition of these criteria
air pollutants.139 It is clear from the body of evidence that oxides of nitrogen, oxides of sulfur,
and particulate matter contribute to total nitrogen (N) and sulfur (S) deposition. In turn, N and S
deposition cause either nutrient enrichment or acidification depending on the sensitivity of the
landscape or the species in question. Both enrichment and acidification are characterized by an
alteration of the biogeochemistry and the physiology of organisms, resulting in harmful declines
in biodiversity in terrestrial, freshwater, wetland, and estuarine ecosystems in the U.S. Decreases
in biodiversity mean that some species become relatively less abundant and may be locally
extirpated. In addition to the loss of unique living species, the decline in total biodiversity can be
harmful because biodiversity is an important determinant of the stability of ecosystems and their
ability to provide socially valuable ecosystem services.
Terrestrial, wetland, freshwater, and estuarine ecosystems in the U.S. are affected by nitrogen
enrichment/eutrophication caused by nitrogen deposition. These effects have been consistently
documented across the U.S. for hundreds of species. In aquatic systems increased nitrogen can
alter species assemblages and cause eutrophication. In terrestrial systems nitrogen loading can
lead to loss of nitrogen-sensitive lichen species, decreased biodiversity of grasslands, meadows
and other sensitive habitats, and increased potential for invasive species.
The sensitivity of terrestrial and aquatic ecosystems to acidification from nitrogen and sulfur
deposition is predominantly governed by geology. Prolonged exposure to excess nitrogen and
sulfur deposition in sensitive areas acidifies lakes, rivers and soils. Increased acidity in surface
waters creates inhospitable conditions for biota and affects the abundance and biodiversity of
fishes, zooplankton and macroinvertebrates and ecosystem function. Over time, acidifying
deposition also removes essential nutrients from forest soils, depleting the capacity of soils to
neutralize future acid loadings and negatively affecting forest sustainability. Major effects in
forests include a decline in sensitive tree species, such as red spruce (Picea rubens) and sugar
maple (Acer saccharum).
Building materials including metals, stones, cements, and paints undergo natural weathering
processes from exposure to environmental elements (e.g., wind, moisture, temperature
fluctuations, sunlight, etc.). Pollution can worsen and accelerate these effects. Deposition of PM
is associated with both physical damage (materials damage effects) and impaired aesthetic
qualities (soiling effects). Wet and dry deposition of PM can physically affect materials, adding
to the effects of natural weathering processes, by potentially promoting or accelerating the
corrosion of metals, by degrading paints and by deteriorating building materials such as stone,
concrete and marble.140 The effects of PM are exacerbated by the presence of acidic gases and
can be additive or synergistic due to the complex mixture of pollutants in the air and surface
characteristics of the material. Acidic deposition has been shown to have an effect on materials
including zinc/galvanized steel and other metal, carbonate stone (as monuments and building
facings), and surface coatings (paints).141 The effects on historic buildings and outdoor works of
art are of particular concern because of the uniqueness and irreplaceability of many of these
objects. In addition to aesthetic and functional effects on metals, stone and glass, altered energy
efficiency of photovoltaic panels by PM deposition is also becoming an important consideration
for impacts of air pollutants on materials.
7-18
-------
7.1.3.4 Environmental Effects of Air Toxics
Emissions from producing, transporting and combusting fuel contribute to ambient levels of
pollutants that contribute to adverse effects on vegetation. Volatile organic compounds (VOCs),
some of which are considered air toxics, have long been suspected to play a role in vegetation
damage.142 In laboratory experiments, a wide range of tolerance to VOCs has been observed.143
Decreases in harvested seed pod weight have been reported for the more sensitive plants, and
some studies have reported effects on seed germination, flowering and fruit ripening. Effects of
individual VOCs or their role in conjunction with other stressors (e.g., acidification, drought,
temperature extremes) have not been well studied. In a recent study of a mixture of VOCs
including ethanol and toluene on herbaceous plants, significant effects on seed production, leaf
water content and photosynthetic efficiency were reported for some plant species.144
Research suggests an adverse impact of vehicle exhaust on plants, which has in some cases
been attributed to aromatic compounds and in other cases to nitrogen oxides.145'146,147 The
impacts of VOCs on plant reproduction may have long-term implications for biodiversity and
survival of native species near major roadways. Most of the studies of the impacts of VOCs on
vegetation have focused on short-term exposure and few studies have focused on long-term
effects of VOCs on vegetation and the potential for metabolites of these compounds to affect
herbivores or insects.
7.2 Non-GHG Monetized Health Benefits
It is important to quantify the health and environmental impacts associated with the revised
program because a failure to adequately consider ancillary impacts could lead to an incorrect
assessment of a program's costs and benefits. Moreover, the health and other impacts of
exposure to criteria air pollutants and airborne toxics tend to occur in the near term, while most
effects from reduced climate change are likely to occur only over a time frame of several decades
or longer. Ideally, human health benefits would be estimated based on changes in ambient PM2.5
and ozone as determined by full-scale air quality modeling. However, the projected non-GHG
emissions impacts associated with the final rule will be expected to contribute to only very small
changes in ambient air quality (see Preamble Section V.C for more detail). EPA intends to
develop a future rule to control emissions of GHGs, criteria pollutants, and air toxic pollutants
from light-duty vehicles for model years beyond 2026. We are considering how to project air
quality impacts, and associated health benefits, from the changes in non-GHG emissions for that
future rulemaking.
In lieu of air quality modeling, we use a reduced-form benefit-per-ton (BPT) approach to
inform our assessment of health impacts, which is conceptually consistent with EPA's use of
BPT estimates in several previous RIAs.148'149 In this approach, the PM2.5-related BPT values are
the total monetized human health benefits (the sum of the economic value of the reduced risk of
premature death and illness) that are expected from reducing one ton of directly-emitted PM2.5 or
PM2.5 precursor such as NOx or SO2. We note, however, that the complex, non-linear
photochemical processes that govern ozone formation prevent us from developing reduced-form
ozone BPT values for mobile sources. This is an important limitation to recognize when using
the BPT approach.
7-19
-------
For tailpipe emissions, we apply national PIVfo.s-related BPT values that were recently derived
for the "Onroad Light-duty Vehicle" sector. 150'f The onroad light-duty vehicle BPT values were
derived using detailed mobile sector source-apportionment air quality modeling, and apply
EPA's existing method for using reduced-form tools to estimate PM2.5 -related benefits.151'152
To monetize the PM2.5-related impacts of upstream emissions, we apply BPT values that were
developed for the refinery and electric generating unit (EGU) sectors.153 While upstream
emissions also include petroleum extraction, storage and transport sources, as well as sources
upstream from the refinery, the modeling tool used to support this analysis only provides
estimates of upstream emissions impacts aggregated across refinery and EGU sources. We
believe for purposes of this rule the separate accounting of refinery and EGU impacts adequately
monetizes upstream PM-related health impacts.
EPA bases its benefits analyses on peer-reviewed studies of air quality and health effects and
peer-reviewed studies of the monetary values of public health and welfare improvements. Very
recently, EPA updated its approach to estimating the benefits of changes in PM2.5 and
ozone.154'155 These updates were based on information drawn from the recent 2019 PM2.5 and
2020 Ozone Integrated Science Assessments (ISAs), which were reviewed by the Clean Air
Science Advisory Committee (CASAC) and the public.156'157 As part of the update, EPA
identified PM2.5-related long-term premature mortality risk estimates from two studies deemed
most appropriate to inform a benefits analysis: a retrospective analysis of Medicare beneficiaries
(Medicare) and the American Cancer Society Cancer Prevention II study (ACS CPS-II). l58J59"g
EPA has not updated its mobile source BPT estimates to reflect these updates in time for this
analysis. Instead, we use PM2.5 BPT estimates that are based on the review of the 2009 PM
ISA160 and 2012 PM ISA Provisional Assessment161 and include a mortality risk estimate
derived from the Krewski et al. (2009)162 analysis of the ACS CPS-II cohort and nonfatal
illnesses consistent with benefits analyses performed for the analysis of the final Tier 3 Vehicle
Rule,163 the final 2012 PM NAAQS Revision,164 and the final 2017-2025 Light-duty Vehicle
GHG Rule.165 We expect this lag in updating our BPT estimates to have only a small impact on
total PM benefits, since the underlying mortality risk estimate based on the Krewski study is
identical to the updated PM2.5 mortality risk estimate derived from an expanded analysis of the
same ACS CPS-II cohort.166 The Agency is currently working to update its mobile source BPT
estimates to reflect these recent updates for use in future rulemaking analyses.
Table 7-1 and Table 7-2 displays the health effects associated with human exposure to
ambient concentrations of PM2.5 and ozone, respectively, including the quantified PM2.5-related
benefits included in the BPT estimates used in this analysis and the unquantified PM2.5 and
ozone health effects the BPT estimates do not capture. Table 7-3 also displays additional criteria
pollutant-related health and environmental effects not captured in the BPT estimates.
-p
Available for download here: https://www.epa.gov/benmap/mobile-sector-source-apportionment-air-quality-and-
benefits-ton.
g The Harvard Six Cities Study (Lepeule et al., 2012), which had been identified for use in estimating mortality
impacts in previous PM benefits analyses, was not identified as most appropriate for the benefits update due to
geographic limitations.
7-20
-------
Table 7-1: Health Effects of Ambient 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 25-99 or age
30-99)
/
/
2009/2012 PM
ISA167168
Infant mortality (age <1)
/
/
2009/2012 PM
ISA

Adult premature mortality from long-term exposure (age 65-99)
—
—
169
2019 PM ISA

Non-fatal heart attacks (age >18)
/
/
2009/2012 PM
ISA

Hospital admissions - respiratory (all ages)
/
/
2009/2012 PM
ISA

Hospital admissions - cardiovascular (age > 20)
/
/
2009/2012 PM
ISA

Emergency department visits—respiratory (all ages)
/
/
2009/2012 PM
ISA

Acute bronchitis (age 8-12)
/
/
2009/2012 PM
ISA

Lower respiratory symptoms (age 7-14)
/
/
2009/2012 PM
ISA

Upper respiratory symptoms (asthmatics age 9-11)
/
/
2009/2012 PM
ISA

Asthma exacerbation (asthmatics age 6-18)
/
/
2009/2012 PM
ISA

Lost work days (age 18-65)
/
/
2009/2012 PM
ISA
Nonfatal morbidity
Minor restricted-activity days (age 18-65)
/
/
2009/2012 PM
ISA
Hospital admissions—cardiovascular (ages 65-99)
—
—
2019 PM ISA
from exposure to PM2.5
Emergency department visits— cardiovascular (age 0-99)
—
—
2019 PM ISA

Hospital admissions—respiratory (ages 0-18 and 65-99)
—
—
2019 PM ISA

Cardiac arrest (ages 0-99; excludes initial hospital and/or emergency
department visits)
—
—
2019 PM ISA

Stroke (ages 65-99)
—
—
2019 PM ISA

Asthma onset (ages 0-17)
—
—
2019 PM ISA

Asthma symptoms/exacerbation (6-17)
—
—
2019 PM ISA

Lung cancer (ages 30-99)
—
—
2019 PM ISA

Allergic rhinitis (hay fever) symptoms (ages 3-17)
—
—
2019 PM ISA

Hospital admissions—Alzheimer's disease (ages 65-99)
—
—
2019 PM ISA

Hospital admissions—Parkinson's disease (ages 65-99)
—
—
2019 PM ISA

Other cardiovascular effects (e.g., other ages)
—
—
2019 PM ISA

Other respiratory effects (e.g., pulmonary function, non-asthma ER
visits, non-bronchitis chronic diseases, other ages and populations)
—
—
2019 PM ISA

Other nervous system effects (e.g., autism, cognitive decline,
dementia)
—
—
2019 PM ISA

Metabolic effects (e.g., diabetes)
—
—
2019 PM ISA

Reproductive and developmental effects (e.g., low birth weight, pre-
term births, etc.)
—
—
2019 PM ISA

Cancer, mutagenicity, and genotoxicity effects
—
—
2019 PM ISA
7-21
-------
Table 7-2: Health Effects of Ambient Ozone
Mortality from exposure to
ozone
Premature respiratory mortality from short-term
exposure (0-99)
—
—
2020 Ozone
ISA170
Premature respiratory mortality from long-term
exposure (age 30-99)
—
—
2020 Ozone
ISA
Nonfatal morbidity from
exposure to ozone
Hospital admissions—respiratory (ages 65-99)
—
—
2020 Ozone
ISA
Emergency department visits—respiratory (ages 0-99)
—
—
2020 Ozone
ISA
Asthma onset (0-17)
—
—
2020 Ozone
ISA
Asthma symptoms/exacerbation (asthmatics age 5-17)
—
—
2020 Ozone
ISA
Allergic rhinitis (hay fever) symptoms (ages 3-17)
—
—
2020 Ozone
ISA
Minor restricted-activity days (age 18-65)
—
—
2020 Ozone
ISA
School absence days (age 5-17)
—
—
2020 Ozone
ISA
Decreased outdoor worker productivity (age 18-65)
—
—
2020 Ozone
ISA
Metabolic effects (e.g., diabetes)
—
—
2020 Ozone
ISA
Other respiratory effects (e.g., premature aging of
lungs)
—
—
2020 Ozone
ISA
Cardiovascular and nervous system effects
—
—
2020 Ozone
ISA
Reproductive and developmental effects
—
—
2020 Ozone
ISA
7-22
-------
Table 7-3: Additional Unquantified Health and Welfare Benefits Categories
Category
Effect
Effect
Quantified
Effect
Monetized
More
Information
Improved Human Health



Reduced incidence of
morbidity from exposure
to N02
Asthma hospital admissions
—
—
2016 N02
ISA171
Chronic lung disease hospital admissions
—
—
2016NO2 ISA
Respiratory emergency department visits
—
—
2016 N02 ISA
Asthma exacerbation
—
—
2016 N02 ISA
Acute respiratory symptoms
—
—
2016 N02 ISA
Premature mortality
—
—
2016 N02 ISA
Other respiratory effects (e.g., airway hyperresponsiveness
and inflammation, lung function, other ages and
populations)
—
—
2016 N02 ISA
Improved Environment



Reduced visibility
impairment
Visibility in Class 1 areas
—
—
2019 PM ISA
Visibility in residential areas
—
—
2019 PM ISA
Reduced effects on
materials
Household soiling
—
—
2019 PM ISA
Materials damage (e.g., corrosion, increased wear)
—
—
2019 PM ISA
Reduced effects from
PM deposition (metals
and organics)
Effects on Individual organisms and ecosystems
—
—
2019 PM ISA
Reduced vegetation and
ecosystem effects from
exposure to ozone
Visible foliar injury on vegetation
—
—
2020 Ozone ISA
Reduced vegetation growth and reproduction
—
—
2020 Ozone ISA
Yield and quality of commercial forest products and crops
—
—
2020 Ozone ISA
Damage to urban ornamental plants
—
—
2020 Ozone ISA
Carbon sequestration in terrestrial ecosystems
—
—
2020 Ozone ISA
Recreational demand associated with forest aesthetics
—
—
2020 Ozone ISA
Other non-use effects


2020 Ozone ISA
Ecosystem functions (e.g., water cycling, biogeochemical
cycles, net primary productivity, leaf-gas exchange,
community composition)
—
—
2020 Ozone ISA
Reduced effects from
acid deposition
Recreational fishing
—
—
2008 NOx SOx
172
ISA
Tree mortality and decline
—
—
2008 NOx SOx
ISA
Commercial fishing and forestry effects
—
—
2008 NOx SOx
ISA
Recreational demand in terrestrial and aquatic ecosystems
—
—
2008 NOx SOx
ISA
Other non-use effects


2008 NOx SOx
ISA
Ecosystem functions (e.g., biogeochemical cycles)
—
—
2008 NOx SOx
ISA
Reduced effects from
nutrient enrichment
Species composition and biodiversity in terrestrial and
estuarine ecosystems
—
—
2008 NOx SOx
ISA
Coastal eutrophication
—
—
2008 NOx SOx
ISA
Recreational demand in terrestrial and estuarine
ecosystems
—
—
2008 NOx SOx
ISA
Other non-use effects


2008 NOx SOx
ISA
Ecosystem functions (e.g., biogeochemical cycles, fire
regulation)
—
—
2008 NOx SOx
ISA
Reduced vegetation
effects from ambient
exposure to S02 and
NOx
Injury to vegetation from S02 exposure
—
—
2008 NOx SOx
ISA
Injury to vegetation from NOx exposure
—
—
2008 NOx SOx
ISA
7-23
-------
In addition to omitting ozone-related impacts from this analysis, there are other impacts
associated with reductions in exposure to NO2, ecosystem benefits, and visibility improvement
that EPA is unable to quantify due to data, resource, and methodological limitations. Chapter 7.1
provides a qualitative description of both the health and environmental effects of the criteria
pollutants controlled by the revised program.
There would also be impacts associated with reductions in air toxic pollutant emissions that
result from the final program (see Chapters 5.1 and 7.1), but the Agency does not attempt to
monetize those impacts. This is because currently available tools and methods to assess air toxics
risk from mobile sources at the national scale are not adequate for extrapolation to incidence
estimation or benefits assessment. While EPA has worked to improve these tools, there remain
critical limitations for estimating incidence and assessing benefits of reducing mobile source air
toxics.
The PM-related BPT estimates used in this analysis are provided in Table 7-4. We multiply
these BPT values by national changes in projected NOx, SO2 and directly-emitted PM2.5, in tons,
to estimate the total PIVfo.s-related monetized human health benefits associated with the final
program. As the table indicates, these values differ among pollutants and depend on their original
source, because emissions from different sources can result in different degrees of population
exposure and resulting health impacts. The BPT values for emissions of non-GHG pollutants
from both onroad light-duty vehicle use and upstream sources such as fuel refineries will
increase over time. These projected increases reflect rising income levels, which increase
affected individuals' willingness to pay for reduced exposure to health threats from air pollution.
The BPT values also reflect future population growth and increased life expectancy, which
expands the size of the population exposed to air pollution in both urban and rural areas,
especially among older age groups with the highest mortality risk.173
Table 7-5 through Table 7-7 display the total undiscounted stream of PIVfo.s-related benefits
and the present value of those benefits for the final rule and two alternatives. Using PM2.5-related
BPT estimates to monetize the non-GHG impacts of the final standards omits ozone-related
impacts as well as other impacts associated with reductions in exposure to air toxics, ecosystem
benefits, and visibility improvement. RIA Chapter 7.1 provides a qualitative description of both
the health and environmental effects of the non-GHG pollutants impacted by the final program.
7-24
-------
Table 7-4: PM-related Benefit-per-ton Values (2018$)a
Year
Onroad Light-duty Vehicles'3
Upstream Sources - Refineries0
Upstream Sources - EGUs0

Direct PM2.5
S02
NOx
Direct PM2.5
S02
NOx
Direct PM2.5
S02
NOx
Estimated Using a 3 Percent Discount Rate



2020
$600,000
$150,000
$6,400
$380,000
$81,000
$8,100
$160,000
$44,000
$6,600
2025
$660,000
$170,000
$6,900
$420,000
$90,000
$8,800
$180,000
$49,000
$7,100
2030
$740,000
$190,000
$7,600
$450,000
$98,000
$9,600
$190,000
$52,000
$7,600
2035
$830,000
$210,000
$8,400
-
-
-
-
-
-
2040
$920,000
$230,000
$9,000
-
-
-
-
-
-
2045
$1,000,000
$250,000
$9,600
-
-
-
-
-
-
Estimated Using a 7 Percent Discount Rate



2020
$540,000
$140,000
$5,800
$350,000
$74,000
$7,300
$150,000
$40,000
$5,900
2025
$600,000
$150,000
$6,200
$380,000
$80,000
$7,900
$160,000
$43,000
$6,400
2030
$660,000
$170,000
$6,800
$410,000
$88,000
$8,600
$170,000
$48,000
$6,900
2035
$750,000
$190,000
$7,500
-
-
-
-
-
-
2040
$830,000
$210,000
$8,200
-
-
-
-
-
-
2045
$900,000
$230,000
$8,600
-
-
-
-
-
-
Notes:









a The benefit-per-ton estimates presented in this table are based on estimates derived from the American Cancer Society
cohort study (Krewski et al., 2009). They also assume either a 3 percent or 7 percent discount rate in the valuation of
premature mortality to account for a twenty-year segmented premature mortality cessation lag.
b Benefit-per-ton values for onroad light-duty vehicles were estimated for the years 2020, 2025, 2030, 2035, 2040, and 2045.
We hold values constant for intervening years (e.g., the 2020 values are assumed to apply to years 2021-2024; 2025 values
for years 2026-2029; and 2045 values for years 2046 and beyond).
0 Benefit-per-ton values for upstream sources were estimated only for the years 2020, 2025 and 2030. We hold values
constant for intervening years and 2030 values are applied to years 2031 and beyond.



7-25
-------
Table 7-5: Undiscounted Stream, Present and Annualized Value of PIVh.s-related Benefits from 2023 through
2050 for the Final Rule (Discounted at 3 percent and 7 percent; SBillions of 2018$)a
Calendar Year
Tailpipe
Upstream
Total

3% DR
7% DR
3% DR
7% DR
3% DR
7% DR
2023
-$0.0034
-$0.0031
$0.02
$0,018
$0,016
$0,015
2024
-$0.0012
-$0.0011
$0,026
$0,024
$0,025
$0,023
2025
$0.0057
$0.0051
$0,054
$0.05
$0.06
$0,055
2026
$0,018
$0,016
$0,097
$0,088
$0.11
$0.1
2027
$0,036
$0,032
$0.18
$0.16
$0.21
$0.19
2028
$0,063
$0,057
$0.27
$0.25
$0.34
$0.3
2029
$0,095
$0,086
$0.34
$0.31
$0.44
$0.4
2030
$0.15
$0.13
$0.45
$0.41
$0.6
$0.54
2031
$0.19
$0.17
$0.52
$0.47
$0.71
$0.64
2032
$0.23
$0.21
$0.59
$0.54
$0.82
$0.74
2033
$0.27
$0.24
$0.65
$0.59
$0.92
$0.83
2034
$0.31
$0.28
$0.72
$0.66
$1
$0.93
2035
$0.44
$0.4
$0.79
$0.72
$1.2
$1.1
2036
$0.48
$0.44
$0.85
$0.77
$1.3
$1.2
2037
$0.52
$0.47
$0.9
$0.82
$1.4
$1.3
2038
$0.56
$0.5
$0.95
$0.86
$1.5
$1.4
2039
$0.59
$0.53
$1
$0.91
$1.6
$1.4
2040
$0.68
$0.62
$1
$0.95
$1.7
$1.6
2041
$0.71
$0.64
$1.1
$0.99
$1.8
$1.6
2042
$0.73
$0.66
$1.1
$1
$1.9
$1.7
2043
$0.75
$0.68
$1.2
$1.1
$1.9
$1.7
2044
$0.77
$0.69
$1.2
$1.1
$2
$1.8
2045
$0.85
$0.77
$1.2
$1.1
$2.1
$1.9
2046
$0.86
$0.78
$1.3
$1.2
$2.1
$1.9
2047
$0.87
$0.79
$1.3
$1.2
$2.2
$2
2048
$0.88
$0.79
$1.3
$1.2
$2.2
$2
2049
$0.88
$0.8
$1.4
$1.2
$2.3
$2
2050
$0.89
$0.8
$1.4
$1.3
$2.3
$2.1
PV
$6.7
$2.8
$12
$5.3
$19
$8.1
Annualized
$0.34
$0.22
$0.61
$0.43
$0.96
$0.65
Notes:






a Note that the non-GHG impacts associated with the standards presented here do not include the full complement
of health and environmental effects that, if quantified and monetized, would change the total monetized estimate
of rule-related impacts. Instead, the non-GHG benefits are based on benefit-per-ton values that reflect only
human health impacts associated with reductions in PM2 5 exposure.



b Calendar year non-GHG benefits presented in this table assume either a 3 percent or 7 percent discount rate in
the valuation of PM-related premature mortality to account for a twenty-year segmented cessation lag. Note that
annual benefits estimated using a 3 percent discount rate were used to calculate the present and annualized values
using a 3 percent discount rate and the annual benefits estimated using a 7 percent discount rate were used to
calculate the present and annualized values using a 7 percent discount rate.


7-26
-------
Table 7-6: Undiscounted Stream, Present and Annualized Value of PIVh.s-related Benefits from 2023 through
2050 for the Proposal (Discounted at 3 percent and 7 percent; SBillions of 2018$)a
Calendar Year
Tailpipe
Upstream
Total
3% DR
7% DR
3% DR
7% DR
3% DR
7% DR
2023
-$0.0023
-$0.0021
$0.0017
$0.0016
-$0.00061
-$0.00053
2024
-$0.0014
-$0.0013
$0.00021
$0.00019
-$0.0012
-$0.0011
2025
$0.0059
$0.0053
$0,015
$0,015
$0,021
$0.02
2026
$0,012
$0,011
$0,048
$0,045
$0,061
$0,056
2027
$0,029
$0,026
$0.11
$0,096
$0.13
$0.12
2028
$0,047
$0,043
$0.17
$0.16
$0.22
$0.2
2029
$0,069
$0,063
$0.22
$0.2
$0.29
$0.26
2030
$0.1
$0,094
$0.29
$0.27
$0.4
$0.36
2031
$0.13
$0.12
$0.34
$0.31
$0.47
$0.43
2032
$0.15
$0.14
$0.39
$0.35
$0.54
$0.49
2033
$0.18
$0.16
$0.43
$0.39
$0.61
$0.55
2034
$0.2
$0.18
$0.47
$0.43
$0.68
$0.61
2035
$0.29
$0.26
$0.52
$0.47
$0.81
$0.73
2036
$0.31
$0.28
$0.56
$0.51
$0.87
$0.79
2037
$0.34
$0.3
$0.59
$0.54
$0.93
$0.84
2038
$0.36
$0.32
$0.62
$0.56
$0.98
$0.88
2039
$0.37
$0.34
$0.65
$0.59
$1
$0.93
2040
$0.43
$0.39
$0.68
$0.62
$1.1
$1
2041
$0.45
$0.4
$0.71
$0.64
$1.2
$1
2042
$0.46
$0.42
$0.73
$0.66
$1.2
$1.1
2043
$0.47
$0.43
$0.75
$0.68
$1.2
$1.1
2044
$0.48
$0.44
$0.78
$0.7
$1.3
$1.1
2045
$0.54
$0.48
$0.8
$0.72
$1.3
$1.2
2046
$0.54
$0.49
$0.82
$0.74
$1.4
$1.2
2047
$0.55
$0.5
$0.84
$0.76
$1.4
$1.3
2048
$0.55
$0.5
$0.85
$0.77
$1.4
$1.3
2049
$0.56
$0.5
$0.88
$0.8
$1.4
$1.3
2050
$0.56
$0.51
$0.91
$0.82
$1.5
$1.3
PV
$4.3
$1.8
$7.7
$3.4
$12
$5.2
Annualized
$0.22
$0.14
$0.39
$0.27
$0.61
$0.42
Notes:
a Note that the non-GHG impacts associated with the standards presented here do not include the full
complement of health and environmental effects that, if quantified and monetized, would change the total
monetized estimate of rule-related impacts. Instead, the non-GHG benefits are based on benefit-per-ton values
that reflect only human health impacts associated with reductions in PM2 5 exposure.
b Calendar year non-GHG benefits presented in this table assume either a 3 percent or 7 percent discount rate in
the valuation of PM-related premature mortality to account for a twenty-year segmented cessation lag. Note
that annual benefits estimated using a 3 percent discount rate were used to calculate the present and annualized
values using a 3 percent discount rate and the annual benefits estimated using a 7 percent discount rate were
used to calculate the present and annualized values using a 7 percent discount rate.
7-27
-------
Table 7-7: Undiscounted Stream, Present and Annualized Value of PIVh.s-related Benefits from 2023 through
2050 for Alternative 2 minus 10 (Discounted at 3 percent and 7 percent; SBillions of 2018$)a
Calendar Year
Tail
pipe
Upstream
Total

3% DR
7% DR
3% DR
7% DR
3% DR
7% DR
2023
-$0.0076
-$0.0068
$0,031
$0,028
$0,023
$0,021
2024
-$0.0083
-$0.0075
$0.05
$0,046
$0,041
$0,038
2025
-$0.00018
-$0.00016
$0,095
$0,087
$0,095
$0,087
2026
$0,018
$0,016
$0.15
$0.13
$0.17
$0.15
2027
$0,037
$0,034
$0.23
$0.21
$0.27
$0.24
2028
$0,069
$0,062
$0.33
$0.29
$0.39
$0.36
2029
$0.1
$0,092
$0.39
$0.35
$0.49
$0.45
2030
$0.16
$0.14
$0.5
$0.45
$0.66
$0.6
2031
$0.2
$0.18
$0.56
$0.51
$0.77
$0.69
2032
$0.24
$0.22
$0.63
$0.57
$0.87
$0.79
2033
$0.28
$0.26
$0.68
$0.62
$0.97
$0.88
2034
$0.32
$0.29
$0.75
$0.68
$1.1
$0.97
2035
$0.46
$0.42
$0.81
$0.74
$1.3
$1.2
2036
$0.51
$0.46
$0.87
$0.79
$1.4
$1.2
2037
$0.54
$0.49
$0.92
$0.83
$1.5
$1.3
2038
$0.58
$0.52
$0.96
$0.87
$1.5
$1.4
2039
$0.61
$0.55
$1
$0.91
$1.6
$1.5
2040
$0.71
$0.64
$1.1
$0.95
$1.8
$1.6
2041
$0.73
$0.66
$1.1
$0.99
$1.8
$1.7
2042
$0.76
$0.68
$1.1
$1
$1.9
$1.7
2043
$0.78
$0.7
$1.2
$1.1
$1.9
$1.8
2044
$0.79
$0.71
$1.2
$1.1
$2
$1.8
2045
$0.88
$0.79
$1.2
$1.1
$2.1
$1.9
2046
$0.89
$0.8
$1.3
$1.1
$2.2
$1.9
2047
$0.9
$0.81
$1.3
$1.2
$2.2
$2
2048
$0.91
$0.82
$1.3
$1.2
$2.2
$2
2049
$0.92
$0.83
$1.4
$1.2
$2.3
$2.1
2050
$0.93
$0.84
$1.4
$1.3
$2.3
$2.1
PV
$7
$2.9
$12
$5.5
$19
$8.4
Annualized
$0.36
$0.23
$0.63
$0.45
$0.99
$0.68
Notes:






" Note that the non-GHG impacts associated with the standards presented here do not include the full complement
of health and environmental effects that, if quantified and monetized, would change the total monetized estimate
of rule-related impacts. Instead, the non-GHG benefits are based on benefit-per-ton values that reflect only
human health impacts associated with reductions in PM2 5 exposure.



b Calendar year non-GHG benefits presented in this table assume either a 3 percent or 7 percent discount rate in
the valuation of PM-related premature mortality to account for a twenty-year segmented cessation lag. Note that
annual benefits estimated using a 3 percent discount rate were used to calculate the present and annualized values
using a 3 percent discount rate and the annual benefits estimated using a 7 percent discount rate were used to
calculate the present and annualized values using a 7 percent discount rate.


7.2.1 Uncertainty
Uncertainties and limitations exist at each stage of the emissions-to-health benefit analysis
pathway (e.g., projected emissions inventories, air quality modeling, health impact assessment,
economic valuation). The BPT approach to monetizing benefits relies on many assumptions;
7-28
-------
when uncertainties associated with these assumptions are compounded, even small uncertainties
can greatly influence the size of the total quantified benefits. Some key assumptions associated
with PM2.5-related health benefits and uncertainties associated with the BPT approach are
described below.
We assume that all fine particles, regardless of their chemical composition, are equally potent
in causing premature mortality. Support for this assumption comes from the 2019 PM ISA,
which concluded that "many PM2.5 components and sources are associated with many health
effects and that the evidence does not indicate that any one source or component is consistently
more strongly related with health effects than PM2.5 mass."174
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 with projected annual mean concentrations that are
above the level of the fine particle standard and areas with projected concentrations below the
level of the standard.
We also 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 Science Advisory Board Health
Effect Subcommittee,175 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. Applying BPT values to estimates of changes in policy-related
emissions precludes us from assessing the distribution of risk as it relates to the associated
distribution of baseline concentrations of PM2.5.
Another limitation of using the BPT approach is an inability to provide estimates of the health
benefits associated with exposure to ozone, ambient NOx, and air toxics. Furthermore, the air
quality modeling that underlies the PM2.5 BPT value did not provide estimates of the PM2.5-
related benefits associated with reducing VOC emissions, but these unquantified benefits are
generally small compared to benefits associated with other PM2.5 precursors.176
National-average BPT values reflect the geographic distribution of the underlying modeled
emissions used in their calculation, which may not exactly match the geographic distribution of
the emission reductions that would occur due to a specific rulemaking. Similarly, BPT estimates
may not reflect local variability in population density, meteorology, exposure, baseline health
incidence rates, or other local factors for any specific location. For instance, even though we
assume that all fine particles have equivalent health effects, the BPT estimates vary across
precursors depending on the location and magnitude of their impact on PM2.5 levels, which
drives population exposure. The emissions and photochemically-modeled PM2.5 concentrations
7-29
-------
used to derive the BPT values may not match the changes in air quality that would result from
the final rule.
Finally, as mentioned earlier in this chapter, EPA recently updated its approach to estimating
the benefits of changes in PM2.5 and ozone. EPA has not had an opportunity to update its mobile
source BPT estimates to reflect these updates in time for this analysis. The Agency is currently
working to update its BPT estimates to reflect these changes for use in future rulemaking
analyses.
7-30
-------
References for Chapter 7
1	U.S. EPA. Policy Assessment (PA) for the Review of the National Ambient Air Quality Standards for Particulate
Matter (Final Report, 2020). U.S. Environmental Protection Agency, Washington, DC, EPA/452/R-20/002, 2020.
2	U.S. EPA. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, 2019). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-19/188, 2019. Table 2-1.
3	U.S. EPA. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, 2019). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-19/188, 2019. Table 2-1.
4	See https://www.epa.gov/air-trends/particulate-matter-pm25-trends and https://www.epa.gov/air-trends/particulate-
matter-pm25-trends#pmnat for more information.
5	U.S. EPA, (2010). Integrated Science Assessment for Carbon Monoxide (Final Report). U.S. Environmental Protection Agency, Washington,
DC, EPA/600/R-09/019F, 2010. http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=218686. See Section 2.1.
6	U.S. EPA (2018) Technical Support Document EPA's 2014 National Air Toxics Assessment.
https://www.epa.gov/national-air-toxics-assessment/2014-nata-assessment-results
7	U.S. EPA (2018) 2014 NATA Summary of Results, https://www.epa.gov/sites/production/files/2020-
07/documents/nata_2014_summary _of_results.pdf
8	U.S. EPA. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, 2019). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-19/188, 2019.
9	U.S. EPA. (2019). Integrated Science Assessment for Particulate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-19/188, Section P. 3.2.3
10	U.S. EPA. Policy Assessment (PA) for the Review of the National Ambient Air Quality Standards for Particulate
Matter (Final Report, 2020). U.S. Environmental Protection Agency, Washington, DC, EPA/452/R-20/002, 2020.
11	U.S. EPA. (2009). Integrated Science Assessment for Particulate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F.
12	Human exposure to ozone varies over time due to changes in ambient ozone concentration and because people
move between locations which have notably different ozone concentrations. Also, the amount of ozone delivered to
the lung is influenced not only by the ambient concentrations but also by the breathing route and rate.
13	U.S. EPA. 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.
14	The ISA evaluates evidence and draws conclusions on the causal relationship between relevant pollutant
exposures and health effects, assigning one of five "weight of evidence" determinations: causal relationship, likely
to be a causal relationship, suggestive of a causal relationship, inadequate to infer a causal relationship, and not
likely to be a causal relationship. For more information on these levels of evidence, please refer to Table II in the
Preamble of the ISA.
15	U.S. EPA. 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.
16	U.S. EPA. Integrated Science Assessment for Oxides of Nitrogen - Health Criteria (2016 Final Report). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-15/068, 2016.
17	U.S. EPA. Integrated Science Assessment (ISA) for Sulfur Oxides - Health Criteria (Final Report, Dec 2017).
U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-17/451, 2017.
18	U.S. EPA, (2010). Integrated Science Assessment for Carbon Monoxide (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-09/019F, 2010.
http://cfpub.epa.gov/ncea/cfm/recordisplay. cfm?deid=218686.
19	The ISA evaluates the health evidence associated with different health effects, assigning one of five "weight of
evidence" determinations: causal relationship, likely to be a causal relationship, suggestive of a causal relationship,
inadequate to infer a causal relationship, and not likely to be a causal relationship. For definitions of these levels of
evidence, please refer to Section 1.6 of the ISA.
7-31
-------
20
Personal exposure includes contributions from many sources, and in many different environments. Total personal
exposure to CO includes both ambient and non-ambient components; and both components may contribute to
adverse health effects.
21	U.S. EPA. (2000). Integrated Risk Information System File for Benzene. This material is available electronically
at: https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm7substance_nmbF276.
99
International Agency for Research on Cancer. (1982). I ARC monographs on the evaluation of carcinogenic risk of
chemicals to humans, Volume 29, Some industrial chemicals and dyestuffs, International Agency for Research on
Cancer, World Health Organization, Lyon, France 1982.
23	Irons, R.D.; Stillman, W.S.; Colagiovanni, D.B.; Henry, V.A. (1992). Synergistic action of the benzene metabolite
hydroquinone on myelopoietic stimulating activity of granulocyte/macrophage colony-stimulating factor in vitro,
Proc. Natl. Acad. Sci. 89:3691-3695.
24
A unit risk estimate is defined as the increase in the lifetime risk of an individual who is exposed for a lifetime to
1 |ig/m3 benzene in air.
25
U.S. EPA. (2000). Integrated Risk Information System File for Benzene. This material is available electronically
at: https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm7substance_nmbF276.
26	International Agency for Research on Cancer (IARC, 2018. Monographs on the evaluation of carcinogenic risks
to humans, volume 120. World Health Organization - Lyon, France. http://publications.iarc.fr/Book-And-Report-
Series/Iarc-Monographs-On-The-Identification-Of-Carcinogenic-Hazards-To-Humans/Benzene-2018.
27
NTP (National Toxicology Program). 2016. Report on Carcinogens, Fourteenth Edition.; Research Triangle Park,
NC: U.S. Department of Health and Human Services, Public Health Service, https://ntp.niehs.nih.gov/go/rocl4
28
Aksoy, M. (1989). Hematotoxicity and carcinogenicity of benzene. Environ. Health Perspect. 82: 193-197. EPA-
HQ-OAR-2011-0135.
29
Goldstein, B.D. (1988). Benzene toxicity. Occupational medicine. State of the Art Reviews. 3: 541-554.
30	Rothman, N., G.L. Li, M. Dosemeci, W.E. Bechtold, G.E. Marti, Y.Z. Wang, M. Linet, L.Q. Xi, W. Lu, M.T.
Smith, N. Titenko-Holland, L.P. Zhang, W. Blot, S.N. Yin, andR.B. Hayes. (1996). Hematotoxicity among Chinese
workers heavily exposed to benzene. Am. J. Ind. Med. 29: 236-246.
o 1
U.S. EPA (2002). Toxicological Review of Benzene (Noncancer Effects). Environmental Protection Agency,
Integrated Risk Information System (IRIS), Research and Development, National Center for Environmental
Assessment, Washington DC. This material is available electronically at
https://cfpub.epa.gov/ncea/iris/iris_documents/documents/toxreviews/0276tr.pdf.
32	Qu, O.; Shore, R.; Li, G.; Jin, X.; Chen, C.L.; Cohen, B.; Melikian, A.; Eastmond, D.; Rappaport, S.; Li, H.;
Rupa, D.; Suramaya, R.; Songnian, W.; Huifant, Y.; Meng, M.; Winnik, M.; Kwok, E.; Li, Y.; Mu, R.; Xu, B.;
Zhang, X.; Li, K. (2003). HEI Report 115, Validation & Evaluation of Biomarkers in Workers Exposed to Benzene
in China.
33	Qu, Q., R. Shore, G. Li, X. Jin, L.C. Chen, B. Cohen, et al. (2002). Hematological changes among Chinese
workers with a broad range of benzene exposures. Am. J. Industr. Med. 42: 275-285.
34	Lan, Qing, Zhang, L., Li, G., Vermeulen, R., et al. (2004). Hematotoxically in Workers Exposed to Low Levels of
Benzene. Science 306: 1774-1776.
35	Turtletaub, K.W. and Mani, C. (2003). Benzene metabolism in rodents at doses relevant to human exposure from
Urban Air. Research Reports Health Effect Inst. Report No.113.
36	U.S. Agency for Toxic Substances and Disease Registry (ATSDR). (2007). Toxicological profile for benzene.
Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service.
http: //www. atsdr. cdc. gov/T oxProfiles/tp3. pdf.
37	A minimal risk level (MRL) is defined as an estimate of the daily human exposure to a hazardous substance that
is likely to be without appreciable risk of adverse noncancer health effects over a specified duration of exposure.
7-32
-------
38 EPA. Integrated Risk Information System. Formaldehyde (CASRN 50-00-0)
https://cfpub.epa.gov/ncea/iris2/chemicalLanding. cfm?substance_nmbr=419.
OQ
NTP (National Toxicology Program). 2016. Report on Carcinogens, Fourteenth Edition.; Research Triangle Park,
NC: U.S. Department of Health and Human Services, Public Health Service, https://ntp.niehs.nih.gov/go/rocl4.
40	IARC Monographs on the Evaluation of Carcinogenic Risks to Humans Volume 88 (2006): Formaldehyde, 2-
Butoxyethanol and l-tert-Butoxypropan-2-ol.
41	IARC Monographs on the Evaluation of Carcinogenic Risks to Humans Volume 100F (2012): Formaldehyde.
42	Hauptmann, M.; Lubin, J. H.; Stewart, P. A.; Hayes, R. B.; Blair, A. 2003. Mortality from lymphohematopoetic
malignancies among workers in formaldehyde industries. Journal of the National Cancer Institute 95: 1615-1623.
43	Hauptmann, M.; Lubin, J. H.; Stewart, P. A.; Hayes, R. B.; Blair, A. 2004. Mortality from solid cancers among
workers in formaldehyde industries. American Journal of Epidemiology 159: 1117-1130.
44	Beane Freeman, L. E.; Blair, A.; Lubin, J. H.; Stewart, P. A.; Hayes, R. B.; Hoover, R. N.; Hauptmann, M. 2009.
Mortality from lymphohematopoietic malignancies among workers in formaldehyde industries: The National Cancer
Institute cohort. J. National Cancer Inst. 101: 751-761.
45	Pinkerton, L. E. 2004. Mortality among a cohort of garment workers exposed to formaldehyde: an update. Occup.
Environ. Med. 61: 193-200.
46	Coggon, D, EC Harris, J Poole, KT Palmer. 2003. Extended follow-up of a cohort of British chemical workers
exposed to formaldehyde. J National Cancer Inst. 95:1608-1615.
47
Hauptmann, M,; Stewart P. A.; Lubin J. H.; Beane Freeman, L. E.; Hornung, R. W.; Herrick, R. F.; Hoover, R.
N.; Fraumeni, J. F.; Hayes, R. B. 2009. Mortality from lymphohematopoietic malignancies and brain cancer among
embalmers exposed to formaldehyde. Journal of the National Cancer Institute 101:1696-1708.
48	ATSDR. 1999. Toxicological Profile for Formaldehyde, U.S. Department of Health and Human Services (HHS),
July 1999.
49	ATSDR. 2010. Addendum to the Toxicological Profile for Formaldehyde. U.S. Department of Health and Human
Services (HHS), October 2010.
50	IPCS. 2002. Concise International Chemical Assessment Document 40. Formaldehyde. World Health
Organization.
51	EPA (U.S. Environmental Protection Agency). 2010. Toxicological Review of Formaldehyde (CAS No. 50-00-0)
- Inhalation Assessment: In Support of Summary Information on the Integrated Risk Information System (IRIS).
External Review Draft. EPA/635/R-10/002A. U.S. Environmental Protection Agency, Washington DC [online].
Available: http ://cfpub .epa.gov/ncea/iris_drafts/recordisplay .cfm?deid=223614.
NRC (National Research Council). 2011. Review of the Environmental Protection Agency's Draft IRIS
Assessment of Formaldehyde. Washington DC: National Academies Press.
http://books.nap.edu/openbook.php?record_id= 13142.
53	U.S. EPA (2018). See https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm?substance_nmbr=419 .
54	U.S. EPA (1991). Integrated Risk Information System File of Acetaldehyde. Research and Development, National
Center for Environmental Assessment, Washington, DC. This material is available electronically at
https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm?substance_nmbr=290.
55	U.S. EPA (1991). Integrated Risk Information System File of Acetaldehyde. This material is available
electronically at https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm?substance_nmbr=290.
56	NTP (National Toxicology Program). 2016. Report on Carcinogens, Fourteenth Edition.; Research Triangle Park,
NC: U.S. Department of Health and Human Services, Public Health Service, https://ntp.niehs.nih.gov/go/rocl4.
57	International Agency for Research on Cancer (IARC). (1999). Re-evaluation of some organic chemicals,
hydrazine, and hydrogen peroxide. IARC Monographs on the Evaluation of Carcinogenic Risk of Chemical to
Humans, Vol 71. Lyon, France.
7-33
-------
CO
U.S. EPA (1991). Integrated Risk Information System File of Acetaldehyde. This material is available
electronically at https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm?substance_nmbr=290.
59	U.S. EPA. (2003). Integrated Risk Information System File of Acrolein. Research and Development, National
Center for Environmental Assessment, Washington, DC. This material is available electronically at
https://cfpub.epa.gov/ncea/iris2/chemicalLanding. cfm?substance_nmbr=364.
60	Appleman, L.M., R.A. Woutersen, and V.J. Feron. (1982). Inhalation toxicity of acetaldehyde in rats. I. Acute and
subacute studies. Toxicology. 23: 293-297.
61	Myou, S.; Fujimura, M.; Nishi K.; Ohka, T.; and Matsuda, T. (1993). Aerosolized acetaldehyde induces
histamine-mediatedbronchoconstriction in asthmatics. Am. Rev. Respir.Dis. 148(4 Pt 1): 940-943.
62	California OEHHA, 2014. TSD forNoncancerRELs: Appendix D. Individual, Acute, 8-Hour, and Chronic
Reference Exposure Level Summaries. December 2008 (updated July 2014).
https://oehha.ca.gov/media/downloads/crnr/appendixdlfinal.pdf
63	U. S. EPA. 1998. Toxicological Review of Naphthalene (Reassessment of the Inhalation Cancer Risk),
Environmental Protection Agency, Integrated Risk Information System, Research and Development, National
Center for Environmental Assessment, Washington, DC. This material is available electronically at
https://cfpub.epa.gov/ncea/iris_drafts/recordisplay.cfm?deid=56434.
64	U. S. EPA. 1998. Toxicological Review of Naphthalene (Reassessment of the Inhalation Cancer Risk),
Environmental Protection Agency, Integrated Risk Information System, Research and Development, National
Center for Environmental Assessment, Washington, DC. This material is available electronically at
https://cfpub.epa.gov/ncea/iris_drafts/recordisplay.cfm?deid=56434.
65	U. S. EPA. (1998). Toxicological Review of Naphthalene (Reassessment of the Inhalation Cancer Risk),
Environmental Protection Agency, Integrated Risk Information System, Research and Development, National
Center for Environmental Assessment, Washington, DC. This material is available electronically at
https://cfpub.epa.gov/ncea/iris_drafts/recordisplay.cfm?deid=56434.
66	Oak Ridge Institute for Science and Education. (2004). External Peer Review for the IRIS Reassessment of the
Inhalation Carcinogenicity of Naphthalene. August 2004.
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=84403.
67	U.S. EPA. (2018) See: https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm?substance_nmbr=436.
68	NTP (National Toxicology Program). 2016. Report on Carcinogens, Fourteenth Edition.; Research Triangle Park,
NC: U.S. Department of Health and Human Services, Public Health Service, https://ntp.niehs.nih.gov/go/rocl4.
69	International Agency for Research on Cancer (IARC). (2002). Monographs on the Evaluation of the Carcinogenic
Risk of Chemicals for Humans. Vol. 82. Lyon, France.
70
U. S. EPA. (1998). Toxicological Review of Naphthalene, Environmental Protection Agency, Integrated Risk
Information System, Research and Development, National Center for Environmental Assessment, Washington, DC.
This material is available electronically at https://cfpub.epa.gov/ncea/iris_drafts/recordisplay.cfm?deid=56434.
71
U.S. EPA. (1998). Toxicological Review of Naphthalene. Environmental Protection Agency, Integrated Risk
Information System (IRIS), Research and Development, National Center for Environmental Assessment,
Washington, DC https://cfpub.epa.gov/ncea/iris_drafts/recordisplay.cfm?deid=56434.
72	U.S. EPA. (2002). Health Assessment of 1,3-Butadiene. Office of Research and Development, National Center
for Environmental Assessment, Washington Office, Washington, DC. Report No. EPA600-P-98-001F. This
document is available electronically at https://cfpub.epa.gov/ncea/iris_drafts/recordisplay.cfm?deid=54499.
73	U.S. EPA. (2002) "Full IRIS Summary for 1,3-butadiene (CASRN 106-99-0)" Environmental Protection Agency,
Integrated Risk Information System (IRIS), Research and Development, National Center for Environmental
Assessment, Washington, DC https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm?substance_nmbr=139.
74	International Agency for Research on Cancer (IARC). (1999). Monographs on the evaluation of carcinogenic risk
of chemicals to humans, Volume 71, Re-evaluation of some organic chemicals, hydrazine and hydrogen peroxide ,
World Health Organization, Lyon, France.
7-34
-------
75	International Agency for Research on Cancer (IARC). (2008). Monographs on the evaluation of carcinogenic risk
of chemicals to humans, 1,3-Butadiene, Ethylene Oxide and Vinyl Halides (Vinyl Fluoride, Vinyl Chloride and
Vinyl Bromide) Volume 97, World Health Organization, Lyon, France.
76	NTP (National Toxicology Program). 2016. Report on Carcinogens, Fourteenth Edition.; Research Triangle Park,
NC: U.S. Department of Health and Human Services, Public Health Service, https://ntp.niehs.nih.gov/go/rocl4.
77	International Agency for Research on Cancer (IARC). (2012). Monographs on the evaluation of carcinogenic risk
of chemicals to humans, Volume 100F chemical agents and related occupations, World Health Organization, Lyon,
France.
78	U.S. EPA. (2002). "Full IRIS Summary for 1,3-butadiene (CASRN 106-99-0)" Environmental Protection
Agency, Integrated Risk Information System (IRIS), Research and Development, National Center for Environmental
Assessment, Washington, DC https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm?substance_nmbr=139.
79
Bevan, C.; Stadler, J.C.; Elliot, G.S.; et al. (1996). Subchronic toxicity of 4-vinylcyclohexene in rats and mice by
inhalation. Fundam. Appl. Toxicol. 32:1-10.
80	U.S. EPA Integrated Risk Information System (IRIS) database is available at: www.epa.gov/iris.
o 1
Karner, A.A.; Eisinger, D.S.; Niemeier, D.A. (2010). Near-roadway air quality: synthesizing the findings from
real-world data. Environ Sci Technol 44: 5334-5344.
09
McDonald, B.C.; McBride, Z.C.; Martin, E.W.; Harley, R.A. (2014) High-resolution mapping of motor vehicle
carbon dioxide emissions. J. Geophys. Res.Atmos.,119, 5283-5298, doi:10.1002/2013JD021219.
83	Kimbrough, S.; Baldauf, R.W.; Hagler, G.S.W.; Shores, R.C.; Mitchell, W.; Whitaker, D.A.; Croghan, C.W.;
Vallero, D.A. (2013) Long-term continuous measurement of near-road air pollution in Las Vegas: seasonal
variability in traffic emissions impact on air quality. Air Qual Atmos Health 6: 295-305. DOI 10.1007/sl 1869-012-
0171-x
84	Kimbrough, S.; Palma, T.; Baldauf, R.W. (2014) Analysis of mobile source air toxics (MSATs)—Near-road VOC
and carbonyl concentrations. Journal of the Air &Waste Management Association, 64:3, 349-359, DOI:
10.1080/10962247.2013.863814
oc
Kimbrough, S.; Owen, R.C.; Snyder, M.; Richmond-Bryant, J. (2017) NO to NO2 Conversion Rate Analysis and
Implications for Dispersion Model Chemistry Methods using Las Vegas, Nevada Near-Road Field Measurements.
Atmos Environ 165: 23-24.
86 Hilker, N.; Wang, J.W.; Jong, C-H.; Healy, R.M.; Sofowote, U.; Debosz, J.; Su, Y.; Noble, M.; Munoz, A.;
Doerkson, G.; White, L.; Audette, C.; Herod, D.; Brook, J.R.; Evans, G.J. (2019) Traffic-related air pollution near
roadways: discerning local impacts from background. Atmos. Meas. Tech., 12, 5247-5261.
https://doi.org/10.5194/amt-12-5247-2019.
on
Grivas, G.; Stavroulas, I.; Liakakou, E.; Kaskaoutis, D.G.; Bougiatioti, A.; Paraskevopoulou, D.; Gerasopoulos,
E.; Mihalopoulos, N. (2019) Measuring the spatial variability of black carbon in Athens during wintertime. Air
Quality, Atmosphere & Health (2019) 12:1405-1417. https://doi.org/10.1007/sll869-019-00756-y
88	Apte, J.S.; Messier, K.P.; Gani, S.; Brauer, M.; Kirchstetter, T.W.; Lunden, M.M.; Marshall, J.D.; Portier, C.J.;
Vermeulen, R.C.H.; Hamburg, S.P. (2017) High-Resolution Air Pollution Mapping with Google Street View Cars:
Exploiting Big Data. Environ Sci Technol 51: 6999-7008. https://doi.org/10.1021/acs.est.7b00891.
89	Dabek-Zlotorzynska, E.; Celo, V.; Ding, L.; Herod, D.; Jeong, C-H.; Evans, G.; Hilker, N. (2019) Characteristics
and sources of PM2 5 and reactive gases near roadways in two metropolitan areas in Canada. Atmos Environ 218:
116980. https://doi.Org/10.1016/j.atmosenv.2019.116980
90	Sarnat, J.A.; Russell, A.; Liang, D.; Moutinho, J.L; Golan, R.; Weber, R.; Gao, D.; Sarnat, S.; Chang, H.H.;
Greenwald, R.; Yu, T. (2018) Developing Multipollutant Exposure Indicators of Traffic Pollution: The Dorm Room
Inhalation to Vehicle Emissions (DRIVE) Study. Health Effects Institute Research Report Number 196. [Online at:
https://www.healtheffects.org/publication/developing-multipollutant-exposure-indicators-traffic-pollution-dorm-
room-inhalation]
7-35
-------
91	Liu, W.; Zhang, J.; Kwon, J.l; et 1. (2006). Concentrations and source characteristics of airborne carbonyl comlbs
measured outside urban residences. J Air Waste Manage Assoc 56: 1196-1204.
92
Cahill, T.M.; Charles, M.J.; Seaman, V.Y. (2010). Development and application of a sensitive method to
determine concentrations of acrolein and other carbonyls in ambient air. Health Effects Institute Research Report
149.Available at http://dx.doi.org.
Q-5
In the widely-used PubMed database of health publications, between January 1, 1990 and August 18, 2011, 605
publications contained the keywords "traffic, pollution, epidemiology," with approximately half the studies
published after 2007.
94	Laden, F.; Hart, J.E.; Smith, T.J.; Davis, M.E.; Garshick, E. (2007) Cause-specific mortality in the unionized U.S.
trucking industry. Environmental Health Perspect 115:1192-1196.
95	Peters, A.; vonKlot, S.; Heier, M.; Trentinaglia, I.; Hormann, A.; Wichmann, H.E.; Lowel, H. (2004) Exposure to
traffic and the onset of myocardial infarction. New England J Med 351: 1721-1730.
96	Zanobetti, A.; Stone, P.H.; Spelzer, F.E.; Schwartz, J.D.; Coull, B.A.; Suh, H.H.; Nearling, B.D.; Mittleman,
M. A.; Verrier, R.L.; Gold, D.R. (2009) T-wave alternans, air pollution and traffic in high-risk subjects. Am J
Cardiol 104: 665-670.
97	Adar, S.; Adamkiewicz, G.; Gold, D.R.; Schwartz, J.; Coull, B.A.; Suh, H. (2007) Ambient and
microenvironmental particles and exhaled nitric oxide before and after a group bus trip. Environ Health Perspect
115: 507-512.
98	Health Effects Institute Panel on the Health Effects of Traffic-Related Air Pollution. (2010). Traffic-related air
pollution: a critical review of the literature on emissions, exposure, and health effects. HEI Special Report 17.
Available at http://www.healtheffects.org.
99	Health Effects Institute. (2019) Protocol for a Systematic Review and Meta-Analysis of Selected Health Effects
of Long-Term Exposure to Traffic-Related Air Pollution. PROSPERO 2019 CRD42019150642 Available from:
https://www.crd.york.ac.uk/prospero/display _record.php?ID=CRD42019150642
100	Boothe, V.L.; Shendell, D.G. (2008). Potential health effects associated with residential proximity to freeways
and primary roads: review of scientific literature, 1999-2006. J Environ Health 70: 33-41.
101	Salam, M.T.; Islam, T.; Gilliland, F.D. (2008). Recent evidence for adverse effects of residential proximity to
traffic sources on asthma. Curr Opin Pulm Med 14: 3-8.
102	Sun, X.; Zhang, S.; Ma, X. (2014) No association between traffic density and risk of childhood leukemia: a
meta-analysis. AsiaPac J Cancer Prev 15: 5229-5232.
103
Raaschou-Nielsen, O.; Reynolds, P. (2006). Air pollution and childhood cancer: a review of the epidemiological
literature. Int J Cancer 118: 2920-9.
104	Boothe, VL.; Boehmer, T.K.; Wendel, A.M.; Yip, F.Y. (2014) Residential traffic exposure and childhood
leukemia: a systematic review and meta-analysis. Am J Prev Med 46: 413-422.
105	National Toxicology Program (2019) NTP Monograph n the Systematic Review of Traffic-related Air Pollution
and Hypertensive Disorders of Pregnancy. NTP Monograph 7.
https://ntp.niehs.nih.gov/ntp/ohat/trap/mgraph/trap_final_508.pdf
106	Volk, H.E.; Hertz-Picciotto, I.; Delwiche, L.; et al. (2011). Residential proximity to freeways and autism in the
CHARGE study. Environ Health Perspect 119: 873-877.
107
Franco-Suglia, S.; Gryparis, A.; Wright, R.O.; et al. (2007). Association of black carbon with cognition among
children in a prospective birth cohort study. Am J Epidemiol, doi: 10.1093/aje/kwm308. [Online at
http://dx.doi.org],
108	Power, M.C.; Weisskopf, M.G.; Alexeef, S.E.; et al. (2011). Traffic-related air pollution and cognitive function
in a cohort of older men. Environ Health Perspect 2011: 682-687.
109	Wu, J.; Wilhelm, M.; Chung, J.; et al. (2011). Comparing exposure assessment methods for traffic-related air
pollution in and adverse pregnancy outcome study. Environ Res 111: 685-6692.
7-36
-------
110	Riediker, M. (2007). Cardiovascular effects of fine particulate matter components in highway patrol officers.
Inhal Toxicol 19: 99-105. doi: 10.1080/08958370701495238 Available at http://dx.doi.org.
111	Alexeef, S.E.; Coull, B.A.; Gryparis, A.; et al. (2011). Medium-term exposure to traffic-related air pollution and
markers of inflammation and endothelial function. Environ Health Perspect 119: 481-486. doi: 10.1289/ehp. 1002560
Available at http://dx.doi.org.
112	Eckel. S.P.; Berhane, K.; Salam, M.T.; et al. (2011). Traffic-related pollution exposure and exhaled nitric oxide
in the Children's Health Study. Environ Health Perspect (IN PRESS), doi: 10.1289/ehp. 1103516. Available at
http://dx.doi.org.
113	Zhang, J.; McCreanor, J.E.; Cullinan, P.; et al. (2009). Health effects of real-world exposure diesel exhaust in
persons with asthma. Res Rep Health Effects Inst 138. [Online at http://www.healtheffects.org],
114	Adar, S.D.; Klein, R.; Klein, E.K.; et al. (2010). Air pollution and the microvasculatory: a cross-sectional
assessment of in vivo retinal images in the population-based Multi-Ethnic Study of Atherosclerosis. PLoS Med
7(11): E1000372. doi:10.1371/journal.pmed.l000372. Available at http://dx.doi.org.
115	Kan, H.; Heiss, G.; Rose, K.M.; et al. (2008). Prospective analysis of traffic exposure as a risk factor for incident
coronary heart disease: the Atherosclerosis Risk in Communities (ARIC) study. Environ Health Perspect 116: 1463-
1468. doi: 10.1289/ehp. 11290. Available at http://dx.doi.org.
116	McConnell, R.; Islam, T.; Shankardass, K.; et al. (2010). Childhood incident asthma and traffic-related air
pollution at home and school. Environ Health Perspect 1021-1026.
117	Islam, T.; Urban, R.; Gauderman, W.J.; et al. (2011). Parental stress increases the detrimental effect of traffic
exposure on children's lung function. Am J Respir Crit Care Med (In press).
110
Clougherty, J.E.; Levy, J.I.; Kubzansky, L.D.; et al. (2007). Synergistic effects of traffic-related air pollution and
exposure to violence on urban asthma etiology. Environ Health Perspect 115: 1140-1146.
119	Chen, E.; Schrier, H.M.; Strunk, R.C.; et al. (2008). Chronic traffic-related air pollution and stress interact to
predict biologic and clinical outcomes in asthma. Environ Health Perspect 116: 970-5.
120	The variable was known as "ETRANS" in the questions about the neighborhood.
121	Rowangould, G.M. (2013) A census of the near-roadway population: public health and environmental justice
considerations. Trans Res D 25: 59-67. http://dx.doi.Org/10.1016/j.trd.2013.08.003
111
Marshall, J.D., Swor, K.R.; Nguyen, N.P (2014) Prioritizing environmental justice and equality: diesel
emissions in Southern California. Environ Sci Technol 48: 4063-4068. https://doi.org/10.1021/es405167f
123
Marshall, J.D. (2008) Environmental inequality: air pollution exposures in California's South Coast Air Basin.
Atmos Environ21: 5499-5503. https://doi.Org/10.1016/j.atmosenv.2008.02.005
114
Tian, N.; Xue, J.; Barzyk. T.M. (2013) Evaluating socioeconomic and racial differences in traffic-related metrics
in the United States using a GIS approach. J Exposure Sci Environ Epidemiol 23: 215-222.
125Boehmer, T.K.; Foster, S.L.; Henry, J.R.; Woghiren-Akinnifesi, E.L.; Yip, F.Y. (2013) Residential proximity to
major highways - United States, 2010. Morbidity and Mortality Weekly Report 62(3): 46-50.
126	US EPA, 2011. Exposure Factors Handbook: 2011 Edition, Chapter 16. [Online at
https://www.epa.gov/sites/production/files/2015-09/documents/efh-chapterl6.pdf]
127	National Research Council, (1993). Protecting Visibility in National Parks and Wilderness Areas. National
Academy of Sciences Committee on Haze in National Parks and Wilderness Areas. National Academy Press,
Washington, DC. This book can be viewed on the National Academy Press Website at
http://www.nap.edu/books/0309048443/html/.
128	Section 169(a) of the Clean Air Act.
129	64 FR 35714, July 1, 1999.
130	62 FR 38680-38681, July 18, 1997.
131
https://www.epa.gov/pm-pollution/national-ambient-air-quality-standards-naaqs-pm
132	73 FR 16486, March 27, 2008.
7-37
-------
1 33
73 FR 16491, March 27, 2008. Only a small percentage of all the plant species growing within the U.S. (over
43,000 species have been catalogued in the USD A PLANTS database) have been studied with respect to ozone
sensitivity.
134 U.S. EPA. 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.
1 -5 C
The concentration at which ozone levels overwhelm a plant's ability to detoxify or compensate for oxidant
exposure varies. Thus, whether a plant is classified as sensitive or tolerant depends in part on the exposure levels
being considered.
136	73 FR 16492, March 27, 2008.
137
73 FR 16493-16494, March 27, 2008, Ozone impacts could be occurring in areas where plant species sensitive to
ozone have not yet been studied or identified.
138	73 FR 16490-16497, March 27, 2008.
139	U.S. EPA. Integrated Science Assessment (ISA) for Oxides of Nitrogen, Oxides of Sulfur and Particulate Matter
Ecological Criteria (Final Report). U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-20/278,
2020.
140	U.S. EPA. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, 2019). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-19/188, 2019.
141	Irving, P.M., e.d. 1991. Acid Deposition: State of Science and Technology, Volume III, Terrestrial, Materials,
Health, and Visibility Effects, The U.S. National Acid Precipitation Assessment Program, Chapter 24, page 24-76.
142	U.S. EPA. (1991). Effects of organic chemicals in the atmosphere on terrestrial plants. EPA/600/3-91/001.
143	Cape JN, ID Leith, J Binnie, J Content, M Donkin, M Skewes, DN Price AR Brown, AD Sharpe. (2003). Effects
of VOCs on herbaceous plants in an open-top chamber experiment. Environ. Pollut. 124:341-343.
144	Cape JN, ID Leith, J Binnie, J Content, M Donkin, M Skewes, DN Price AR Brown, AD Sharpe. (2003). Effects
of VOCs on herbaceous plants in an open-top chamber experiment. Environ. Pollut. 124:341-343.
145	Viskari E-L. (2000). Epicuticular wax of Norway spruce needles as indicator of traffic pollutant deposition.
Water, Air, and Soil Pollut. 121:327-337.
146	Ugrekhelidze D, F Korte, G Kvesitadze. (1997). Uptake and transformation of benzene and toluene by plant
leaves. Ecotox. Environ. Safety 37:24-29.
147
Kammerbauer H, H Selinger, R Rommelt, A Ziegler-Jons, D Knoppik, B Hock. (1987). Toxic components of
motor vehicle emissions for the spruce Picea abies. Environ. Pollut. 48:235-243.
148	U.S. Environmental Protection Agency (U.S. EPA). 2015. Regulatory Impact Analysis for the Final Revisions
to the National Ambient Air Quality Standards for Ground-Level Ozone. EPA452/R-15-007. Office of Air Quality
Planning and Standards, Health and Environmental Impacts Division, Research Triangle Park, NC. December.
Available at: http://www.epa.gov/ttnecasl/regdata/RIAs/finalria.pdf.
149	U.S. Environmental Protection Agency (U.S. EPA). (2012). Regulatory Impact Analysis: Final Rulemaking for
2017-2025 Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average Fuel Economy
Standards, Assessment and Standards Division, Office of Transportation and Air Quality, EPA-420-R-12-016,
August 2012. Available on the Internet at: http://www3.epa.gov/otaq/climate/documents/420rl2016.pdf.
150	Wolfe, P.; Davidson, K.; Fulcher, C.; Fann, N.; Zawacki, M.; Baker, K. R. 2019. Monetized Health Benefits
Attributable to Mobile Source Emission Reductions across the United States in 2025. Sci. Total Environ. 650, 2490-
2498. https://doi.Org/10.1016/J.SCITOTENV.2018.09.273.
151	Zawacki, M.; Baker, K. R.; Phillips, S.; Davidson, K.; Wolfe, P. 2018. Mobile Source Contributions to Ambient
Ozone and Particulate Matter in 2025. Atmos. Environ. 188, 129-141.
https://doi.Org/10.1016/J.ATMOSENV.2018.04.057.
1 52
Fann, N.; Fulcher, C. M.; Baker, K. 2013. The Recent and Future Health Burden of Air Pollution Apportioned
across U.S. Sectors. Environ. Sci. Technol. 47 (8), 3580-3589. https://doi.org/10.1021/es304831q.
7-38
-------
1 53
U.S. Environmental Protection Agency (U.S. EPA). 2018. Technical Support Document: Estimating the Benefit
per Ton of Reducing PM2 5 Precursors from 17 Sectors. 2018. Office of Air Quality Planning and Standards.
Research Triangle Park, NC.
154	U.S. Environmental Protection Agency (U.S. EPA). 2021a. Regulatory Impact Analysis for the Final Revised
Cross-State Air Pollution Rule (CSAPR) Update for the 2008 Ozone NAAQS. EPA-452/R-21-002. March.
155	U.S. Environmental Protection Agency (U.S. EPA). 2021b. Estimating PM2 5- and Ozone-Attributable Health
Benefits. Technical Support Document (TSD) for the Final Revised Cross-State Air Pollution Rule Update for the
2008 Ozone Season NAAQS. EPA-HQ-OAR-2020-0272. March.
156	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.
157	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.
158	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.
159	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.
160	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: .
161	U.S. Environmental Protection Agency (U.S. EPA). 2012. Provisional Assessment of Recent Studies on Health
Effect of Particulate Matter Exposure. EPA/600/R-12/056F. National Center for Environmental Assessment - RTP
Division, Research Triangle Park, NC. December. Available at:
https ://cfpub .epa.gov/ncea/isa/recordisplay. cfm?deid=247132.
162	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.
163	U.S. Environmental Protection Agency. (2014). Control of Air Pollution from Motor Vehicles: Tier 3 Motor
Vehicle Emission and Fuel Standards Final Rule: Regulatory Impact Analysis, Assessment and Standards Division,
Office of Transportation and Air Quality, EPA-420-R-14-005, March 2014. Available on the internet:
http://www3.epa.gOv/otaq/documents/tier3/420rl4005.pdf.
164	U.S. Environmental Protection Agency. (2012). Regulatory Impact Analysis for the Final Revisions to the
National Ambient Air Quality Standards for Particulate Matter, Health and Environmental Impacts Division, Office
of Air Quality Planning and Standards, EPA-452-R-12-005, December 2012. Available on the internet:
http://www3.epa.gov/ttnecasl/regdata/RIAs/finalria.pdf.
165	U.S. Environmental Protection Agency (U.S. EPA). (2012). Regulatory Impact Analysis: Final Rulemaking for
2017-2025 Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average Fuel Economy
Standards, Assessment and Standards Division, Office of Transportation and Air Quality, EPA-420-R-12-016,
August 2012. Available on the Internet at: http://www3.epa.gov/otaq/climate/documents/420rl2016.pdf.
166	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.
167	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: .
7-39
-------
168	U.S. Environmental Protection Agency (U.S. EPA). 2012. Provisional Assessment of Recent Studies on Health
Effect of Particulate Matter Exposure. EPA/600/R-12/056F. National Center for Environmental Assessment - RTP
Division, Research Triangle Park, NC. December. Available at:
https ://cfpub .epa.gov/ncea/isa/recordisplay. cfm?deid=247132.
169	U.S. Environmental Protection Agency (U.S. EPA). 2019. Integrated Science Assessment (ISA) for Particulate
Matter (Final Report, 2019). U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-19/188, 2019.
170
U.S. Environmental Protection Agency (U.S. EPA). 2020. 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.
171	U.S. Environmental Protection Agency (U.S. EPA). 2016. 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>.
1 72
U.S. Environmental Protection Agency (U.S. EPA). 2008. 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:
.
173
For more information about income growth adjustment factors and EPA's population projections, please refer to
the following: https://www.epa.gov/sites/production/files/2015-04/documents/benmap-
ce_user_manual_march_2015 .pdf.
174	U.S. Environmental Protection Agency (U.S. EPA). 2019. Integrated Science Assessment (ISA) for Particulate
Matter (Final Report, 2019). U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-19/188, 2019.
1 75
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:
.
176 U.S. EPA. 2012. Regulatory Impact Analysis for the Proposed Revisions to the National Ambient Air Quality
Standards for Particulate Matter.
7-40
-------
Chapter 8: Vehicle Sales, Employment, Environmental Justice, and
Affordability and Equity Impacts
8.1 Sales Impacts
8.1.1 Conceptual Framework
A significant question in vehicle GHG rules has been why there have appeared to be existing
technologies that, if adopted, would reduce fuel consumption enough to pay for themselves in
short periods, but which were not widely adopted. If the benefits to vehicle buyers outweighed
the costs to those buyers of the new technologies, economic principles suggest that automakers
would provide them, and people would buy them. Yet engineering analyses have identified a
number of technologies, such as downsized-turbocharged engines, gasoline direct injection, and
improved aerodynamics, with short payback periods that were not widely adopted before the
standards, but which were adopted rapidly afterwards.1 Why did markets fail, on their own, to
adopt these technologies?
This question, termed the "energy paradox" or "energy efficiency gap,"2 has received a great
deal of discussion in previous rulemakings.3 The gap exists if the estimates of net benefits of
these new technologies are correct, and if there are no major adverse effects associated with the
technologies (hidden costs) that provide clear disincentives to adopt the technologies. A separate
question is to explain why the gap exists.
8.1.1.1 Existence of the Energy Efficiency Gap
EPA has previously explored the existence of the paradox, including in the Midterm
Evaluation.4 In terms of the costs and effectiveness of the fuel-saving technologies, EPA has
relied on published research, highly-regarded teardown studies,5 and extensive testing to ensure
the best available estimates for its analyses. In the MTE's TAR and Proposed Determination
TSD, EPA undertook retrospective analysis of its cost and effectiveness estimates and generally
confirmed the previous estimates. See Chapter 4 of this RIA for more discussion of the
technology and cost estimates for this rule.
The 2021 National Academies of Science (NAS) report6 (p. 11-348) raises the issue of
tradeoffs between improved performance and fuel economy and recommends that "agencies
should collect further evidence on the influence of vehicle performance trade-offs on automaker
compliance strategies and consumers, and reassess whether forgone performance improvements
should be included in benefit-cost analysis of the standards. The agencies should assess how new
technologies penetrating the market will affect the trade-offs among greenhouse gas (GHG)
emissions rates, performance, and other attributes."
EPA has considered evidence related to potential adverse effects on other vehicle attributes.
First, EPA sponsored research to evaluate how auto reviewers — professionals expected to be
especially sensitive to vehicle performance and attributes — evaluated MY 2014 and 2015
vehicles with fuel-saving technologies (Helfand et al. 2016, Huang et al. 2018).7 These studies
found that all technologies were evaluated positively more often than they were evaluated
negatively, suggesting that it is possible to implement these technologies without imposing
hidden costs. In addition, they looked for correlations between evaluations of each technology
8-1
-------
and a range of operational characteristics (handling, acceleration, noise, etc.). Both papers found
few correlations between the existence of a technology and negative rating of operational
characteristics; and even fewer of those were consistently correlated in both model years. In
addition, Huang et al. (2018) found that overall evaluation of a vehicle's quality is more
associated with operational characteristics than with the technologies themselves; as noted in the
previous paragraph, there is little association between operational characteristics and the
technologies.
Additional research (Huang et al. 2018a)8 explored the use of results from a consumer
satisfaction survey conducted by Strategic Vision to look at how vehicle buyers responded to the
presence of fuel-saving technologies. Preliminary results were developed using a subset of the
data, due to incomplete matching of technology information with survey information. Overall,
people were highly satisfied with their newly purchased vehicles; less than 3 percent of owners
expressed dissatisfaction. This result is not surprising; people are unlikely to buy new vehicles
that they find unsatisfactory. Further, comparing negative satisfaction ratings from before and
after the presence of fuel saving technology, results show little correlation between the presence
of a technology and a change in satisfaction ratings for overall experience, power and pickup,
driving performance, noise/vibration/harshness, or fuel economy. EPA continues to explore these
data.
In these three studies, a limitation is that it is not possible to demonstrate causally that the
presence or absence of a technology affects people's perceptions of vehicle quality. For instance,
it may be that some of the fuel-saving technologies are used primarily in market segments that
would, regardless of the presence of the technology, not be considered as high-quality as vehicles
in other market segments; an association between the presence of the technology and an
evaluation of quality may be based, not on a causal effect of the technology on quality, but rather
a correlation due to its use in that market segment. Nevertheless, the research indicates that it has
been possible for these technologies to be adopted without observed adverse impacts on
assessments by expert reviewers or on consumer satisfaction.
Some research9 has argued that reducing fuel consumption must come at the expense of either
vehicle acceleration or vehicle weight. The concept is based in the principle that energy is
required to move the vehicle, so that heavier and faster vehicles will require more energy use.
While this statement is true when all else (e.g., powertrain, body style) is held constant, it is
uncommon that, in fact, all else is held constant. The Midterm Evaluation Proposed
Determination TSD, Chapter 4.1.2, discussed some concerns and limitations with the existing
literature. One issue is that the papers typically assume that the tradeoff between power and fuel
economy, or between weight and fuel economy, does not vary over time. MacKenzie and
Hey wood (2015) further point out that these studies are based on horsepower and weight, two
measures that may not accurately reflect the characteristics sought by vehicle buyers. If, as they
have found, the relationship between acceleration (measured as 0-to-60 speed) and horsepower
divided by weight has changed over time, then studies holding constant the relationship between
horsepower-to-weight and fuel consumption are not accurately measuring the tradeoff of concern
to vehicle buyers.
Recent work by Moskalik et al. (2018)10 suggests that using historic data to estimate tradeoffs
may miss changes in the relationship between acceleration and CO2 emissions with new
technologies. Moskalik et al. (2018) shows results using the ALPHA model for trade-off curves
8-2
-------
between CO2 emissions and 0-to-60 acceleration time for five different engine types covering a
range of production years: carbureted, port fuel injection (PFI), gasoline direct injection (GDI),
Atkinson, and turbo-downsized (TDS) engines. These engines have different operating efficiency
characteristics, and thus different tradeoff curves. Most notably, the newest TDS engines have
much flatter tradeoffs than earlier carbureted, GDI and PFI; in fact, the "future" TDS engine
shows almost no change in CO2 emissions over a wide range of acceleration times. Thus, the
assumption in the previous research that the tradeoffs among acceleration, fuel economy, and
weight are constant does not appear to accurately represent the new technologies, and in fact may
substantially overestimate the magnitude of the performance-fuel economy tradeoff.
Watten et al. (2021)11 develops a theoretical framework that incorporates producer decisions
on technological adoption and attribute production, taking into account consumer preferences
toward performance and fuel economy. This paper distinguishes between technologies that
improve, or do not adversely affect, both performance and fuel economy, and reducing engine
displacement, which does trade off improved fuel economy for performance. Following
Moskalik et al. (2018), it observes that the "marginal rate of attribute substitution" between
power and fuel economy has changed substantially over time. In particular, it has become
relatively more costly to improve efficiency by reducing power, and relatively less costly to add
technologies that improve efficiency. These technology improvements do not reduce power and
in some cases may increase it. It supports the concept that automakers take consumer preferences
into account in identifying where to add technology.
Regarding the NAS's recommendation to evaluate whether forgone performance
improvements should be included in the benefit-cost analysis, the agencies have typically
included the costs of holding performance constant in the rulemakings. As discussed in previous
paragraphs, ways exist to enhance performance without adverse effects on fuel economy. In that
case, Helfand and Dorsey-Palmateer (2015)12 argue that the only additional cost due to the
standards associated with additional performance would occur if adding that performance is
more expensive for a vehicle with higher fuel economy. In addition, it would be important to add
benefits associated with improvements in other attributes due to fuel-saving technologies. As
EPA has found in the past, many fuel-saving technologies can enhance performance, handling, or
other attributes. Whitefoot et al. (2017)13 find that allowing for tradeoffs between performance
and fuel economy reduces the costs of the standards, by allowing an additional way to achieve
compliance. EPA has not included either these potential changes in costs or increased benefits in
its analysis, due to lack of sufficient data to estimate these effects. Recent technology trends and
other evidence suggest that tradeoffs need not lead to forgone performance attributes; and, if
manufacturers lower costs by reducing other attributes, EPA's constant-performance assumption
results in a conservative estimate of the rule's overall compliance costs.
This discussion does not reject the observation that the energy efficiency gap has existed for
light-duty vehicles. Cost and effectiveness values for the technologies have not been shown to
have significant errors. Helfand and Dorsey-Palmateer (2015) conclude that, in response to the
standards, automakers have improved fuel economy without adversely affecting other vehicle
attributes, and any remaining tradeoffs are likely to be included in the costs of the technologies.
Thus, it appears that markets on their own have not led to adoption of a number of technologies
with short payback periods in the absence of the standards.
8-3
-------
8.1.1.2 Potential Explanations for the Existence of the Energy Efficiency Gap
While EPA has documented the existence of the energy efficiency gap for provision of fuel-
saving technologies, various hypotheses have been raised for the causes of that gap. Previous
rules have discussed a number of these hypotheses for this apparent market failure.14 Here we
summarize a number of theories that have been suggested. Researchers and some commenters
use different names and organizational principles in defining these theories. As a result, this list
may not seem complete to all.
On the consumer side, the 2021 NAS Report (p. 11-355) observes that "the literature has not
settled on a single explanation for potential consumer undervaluation of fuel cost savings."
Hypotheses include:
•	Consumers might lack the information necessary to estimate the value of future fuel
savings, not have a full understanding of this information even when it is presented, or
not trust the presented information
•	Consumers might be "myopic" and hence undervalue future fuel savings in their
purchasing decisions
•	Consumers may be accounting for uncertainty in future fuel savings when comparing
upfront cost to future returns
•	Consumers may consider fuel economy after other vehicle attributes and, as such, not
optimize the level of this attribute (instead "satisficing" - that is, selecting a vehicle
that is acceptable rather than optimal — or selecting vehicles that have some sufficient
amount of fuel economy)
•	Consumers might be especially averse to the short-term losses associated with the
higher prices of energy efficient products relative to the long-term gains of future fuel
savings (the behavioral phenomenon of "loss aversion")
•	Consumers might associate higher fuel economy with inexpensive, less well designed
vehicles
•	When buying vehicles, consumers may focus on visible attributes that convey status,
such as size, and pay less attention to attributes such as fuel economy that typically do
not visibly convey status
•	Even if consumers have relevant knowledge, selecting a vehicle is a highly complex
undertaking, involving many vehicle characteristics. In the face of such a complicated
choice, consumers may use simplified decision rules
•	Because consumers differ in how much they drive, they may already sort themselves
into vehicles with different, but individually appropriate, levels of fuel economy in
ways that an analysis based on an average driver does not identify
EPA has explored the evidence on how consumers evaluate fuel economy in their vehicle
purchase decisions.15 Some research finds that vehicle buyers consider close to all fuel
consumption over a vehicle's lifetime in the purchase decision.16 Others find that vehicle buyers
consider only a small share of that future consumption in the purchase decision.17 Sallee (2014)
argues that it might be rational for consumers not to expend too much effort to calculate fuel
savings, because increased precision might not have much effect on their purchase decisions.18
The variation in estimates of willingness to pay is very large, even after outliers are removed;
Greene et al. (2018) estimated a mean willingness to pay for a $0.01 reduction in fuel cost per
8-4
-------
mile among published estimates of $1880, a median of $990, and a standard deviation of $6880,
compared to a reference value of $1150 for the value of reducing fuel costs by $0.01/mile over
the lifetime of an average vehicle.21 The estimates vary based on the type of study (revealed
preference, stated preference, or market sales), and the form of statistical model used to analyze
the data. These observations provide little guidance on whether consumers overvalue or
undervalue fuel economy, or get the estimates approximately right. Thus, it is not clear whether
consumer behavior is responsible for the energy efficiency gap, much less which hypotheses
might explain it.
For possible explanations on the producer side, two major themes arise: the role of market
structure and business strategy, and the nature of technological invention and innovation.
•	Light-duty vehicle production involves significant fixed costs, and automakers strive
to differentiate their products from each other. These observations suggest that
automakers can act strategically in how they design and market products. In this
context, the fuel economy of a vehicle can become a factor in product differentiation
rather than a decision based solely on cost-effectiveness of a fuel-saving technology.19
Product differentiation carves out corners of the market for different automobile
brands and models. For instance, automakers may emphasize luxury characteristics in
some vehicles to attract people with preferences for those characteristics, and they
may emphasize cost and fuel economy for people attracted to frugality. By separating
products into different market segments, producers both provide consumers with
goods targeted for their tastes, and may reduce competition among vehicle models,
creating the possibility of greater profits. From the producer perspective, fuel economy
is not necessarily closely related to the cost-effectiveness of the technologies to
consumers, but rather is one of many factors that manufacturers use to market their
models to different consumer groups. As Fischer (2005) points out, this strategy can
lead to inefficiencies in the market: an under-supply of fuel economy relative to what
is cost-effective to consumers in some segments, and an over-supply of fuel economy
in other sectors.20 The structure of the automobile industry may inefficiently allocate
car attributes—fuel economy among them—and help to explain the existence of an
energy efficiency gap.
•	Innovation - the first commercialization of a new product - occurs on a continuum
between two extremes: "major" innovation where product characteristics change, and
"incremental" innovation13 which exploits relatively minor changes to the existing
product.21 In the absence of standards, automakers have seemed willing to invest in
small improvements upon existing technologies ("incremental" technologies) that can
be used to improve fuel economy or other vehicle attributes (Helfand and Dorsey-
Palmateer 2015). However, they may be more hesitant to invest in "major"
innovations in the absence of standards, for several reasons, including being the first
(or one of the first) to invest in a new technology.
a It also provides a reference value of $1150 for the value of reducing fuel costs by $0.01/mile over the lifetime of
an average vehicle.
b Abernathy and Utterback use "major" and "incremental" Henderson and Clark, with a two-dimensional framework,
use "radical" and "incremental."
8-5
-------
•	There may be first-mover disadvantages to investing in new technologies. Many
manufacturers prefer to observe the market and follow other manufacturers rather than
be the first to market with a specific technology. The "first-mover disadvantage" has
been recognized in other research where the "first-mover" pays a higher proportion of
the costs of developing technology, but loses the long-term advantage when other
businesses follow quickly.22 This effect may be even more significant when the
benefits of new technologies provide public goods: because the general public benefits
more from public goods than does the individual producer, producers do not receive
appropriate incentives to provide those technologies.
•	There could be "dynamic increasing returns" to adopting new technologies, wherein
the value of a new technology may depend on how many other companies have
adopted the technology — for instance, creating multiple suppliers for a technology
should increase competition, improve quality, and reduce price. This could be due to
network effects or learning-by-doing. In a network effects situation, the usefulness of
the technology depends on others' adoption of the technology: e.g., a telephone is only
useful if other people also have telephones. Learning by doing is the concept that the
costs (benefits) of using a particular technology decrease (increase) with use. Both of
these incentivize firms to pursue a "wait and see" strategy when it comes to adopting
new technologies.23
•	There can be synergies when companies work on the same technologies at the same
time.24 Research among multiple parties can be a synergistic process: ideas by one
researcher may stimulate new ideas by others, and more and better results occur than if
the one researcher operated in isolation.c Standards can promote research into low-
CO2 technologies that would not take place in the absence of the standards. Because
all companies (both auto firms and auto suppliers) have incentives to find better, less
expensive ways of meeting the standards, the possibilities for synergistic interactions
may increase. Thus, the standards, by focusing all companies on finding more efficient
ways of achieving the standards, may lead to better outcomes than if any one company
operated on its own.
Much less research has been conducted to evaluate the producer side of the market. The 2015
NAS report (cited in the 2021 NAS report) observes that automakers "perceive that typical
consumers would pay upfront for only one to four years of fuel savings" (p. 9-10),25 a range of
values within that identified in Greene et al. (2018) for consumer response, but well below the
median or mean. It may be possible, though puzzling, that automakers operate under a
misperception of consumer willingness to pay for additional fuel economy. The 2021 NAS
Report (p. 11-356) observes that the auto industry is concentrated, "in part owing to the large
capital investments necessary to enter the automotive market," and raises the "first-mover
disadvantage" argument. In addition, it discusses the challenges associated with a "disruptive"
technology such as the transition to electrification (p. 11-358). Thus, it supports the concept that
c Powell and Giannella (2010) discuss how a "collective momentum" has led uncoordinated research efforts among
a diverse set of players to develop advances in a number of technologies (such as electricity and telephones). They
contrast this view of technological innovation with that of proprietary research in corporate laboratories, where the
research is part of a corporate strategy. Such momentum may result in part from alignment of economic, social,
political, and other goals.
8-6
-------
there potentially are barriers to adoption of new technologies on the part of automakers, though it
also does not provide conclusive evidence.
Some theories involve the interaction between producers and consumers. For instance,
• "Split incentives" refers to a situation where a person authorized to make a decision
for someone else — an agent — may have different objectives than the person who will
live with the decision — the principal. For instance, the purchasing agent for a fleet
may focus on minimizing purchase costs, in order to buy as many vehicles as possible,
though such practice may lead to higher operating and maintenance costs for the fleet
managers. This effect may also appear within auto makers, if those influencing the
fuel economy of new vehicles have reason to be more focused on up-front costs than
in the total cost of ownership that vehicle buyers will face. Split incentives might lead
to under-provision of charging facilities for rental properties and workplaces.
In sum, it continues to be an open question which combination of theories may best explain
why there was limited adoption of cost-effective fuel-saving technologies before the
implementation of more stringent standards, that were adopted without serious disruption to the
vehicle market after the standards became effective. Nevertheless, it appears to have happened.
Some combination of market failures must explain why markets have not provided all fuel-
saving technologies that would save money. Regulation appears to help correct such market
failures, while also addressing other externalities like pollution.
8.1.2 How Sales Impacts were Modeled
As discussed in Chapter 4, EPA is using the CAFE Compliance and Effects Modeling System
(CCEMS) model for this analysis. The FRIA for the SAFE rule (starting p. 871) describes the
approach to vehicle sales impacts used in the model. First, it projects future new vehicle sales in
the reference case based on projections of macroeconomic variables. Second, it applies a demand
elasticity (that is, the percent change in the quantity sold resulting from a one percent increase in
price) to the change in net price, where net price is the difference in technology costs less an
estimate of the change in fuel costs over 2.5 years. This approach assumes that vehicle buyers
and automakers take into consideration the fuel savings that consumers expect to accrue over the
first 2.5 years of vehicle ownership — an assumption that warrants further evaluation as
discussed below. This assumption applies to both the without-program and with-program
calculations. It does not allow for different perceptions of the value of fuel economy to buyers on
the part of automakers, in providing fuel-saving technologies, and those buyers.
As discussed in Chapter 8.1, there does not yet appear to be consensus around the role of fuel
consumption in people's vehicle purchase decisions, and the assumption that 2.5 years of fuel
consumption is the right number deserves further evaluation. As noted there, this assumption is
consistent with automakers' statements of their perceptions of consumer behavior. Also as noted
there, Greene et al. (2018) provides a reference value of $1,150 for the value of reducing fuel
costs by $0.01/mile over the lifetime of an average vehicle; for comparison, 2.5 years of fuel
8-7
-------
savings is about 30 percent of that value, or about $333.d This value is within the large standard
deviation in Greene et al. (2018) for the willingness to pay to reduce fuel costs, but it is lower
than both the mean of $1,880 (160 percent of the reference value) and the median of $990 (85
percent of the reference value) per one cent per mile in the paper. EPA estimates that the present
value of 85 percent of fuel consumption is about 10 years of fuel consumption, using Greene et
al.'s assumptions.6 It appears possible that automakers may operate under a different perception
of consumer willingness to pay for additional fuel economy than how consumers actually
behave. CCEMS does not allow automaker perception to differ from consumer behavior.
In the NPRM, EPA used an elasticity of demand at -1 as its central case, and an elasticity of -
0.4 for sensitivity analysis. The value of -1 was based on literature more than 25 years old, and
was based on studies that focus on the short run, a period typically considered to be less than one
year.26 For durable goods, such as vehicles, people are expected to have more flexibility about
when they purchase new vehicles than whether they purchase new vehicles; thus, their behavior
is more inflexible (less elastic) in the long run than in the short run. For this reason, estimates for
long-term elasticities for durable goods are expected to be smaller (in absolute value) than short-
run elasticities. At a market level, short-run responses typically focus entirely on the new-vehicle
market; longer time spans allow for adjustments between the new and used vehicle markets, and
even adjustments outside those markets, such as with public transit. Because this rule has effects
over time, and could have effects related to the used vehicle market, long-run elasticities that
account for effects in the used vehicle market are more appropriate for estimating the impacts of
standards in the new vehicle market than short-run elasticities.
EPA commissioned work with RTI International and its subject matter expert, Dr. Mark
Jacobsen of the University of California at San Diego, to review more recent estimates of the
elasticity of demand for new vehicles.27 RTI found that all but one study of short-run elasticities
since 1997 have estimated elasticities to be between -0.37 and -0.78; in addition, two studies that
account for changes in the used vehicle market provide estimates of -0.18 and -0.36. The RTI
report also developed an approach based on economic principles to estimate how changes in the
new vehicle market relate to the used vehicle market. Using available parameters from published
research, RTI calculated that the "policy elasticity," the value that takes into account effects in
the used vehicle market, is much smaller than the short-run demand elasticity. Table 8-1 presents
its calculations of the effects of multiple combinations of key parameters and mostly finds
elasticity values between -0.14 and -0.27; the one exception, -0.39, is based on the high (in
absolute value) new-vehicle demand elasticity of-1.27 and a high estimate for substitution out of
the auto market. Thus, this new research suggests that a new-vehicle demand elasticity
d Greene et al. (2018) does not provide enough detail to replicate their analysis perfectly. The 30 percent estimate is
calculated by assuming, following assumptions in Greene et al. (2018), that a vehicle is driven 15,000 miles per year
for 13.5 years, 10 percent discount rate. Those figures produce a "present value of miles" of 108,600; thus, a
$0.01/mile change in the cost of driving would be worth $1086. In contrast, saving $0.01/mile for 2.5 years is worth
about $318, or 29 percent of the value over 13.5 years. Here we use 29 percent of Greene et al.'s estimate
($1150*0.29 = $333).
e With a 10 percent discount rate, the present value of 15,000 miles per year at age 10 is 85 percent of the present
value of 15,000 miles peryear at age 13.5. For comparison, a 5 percent discount rate achieves 85 percent of the 13.5
years of present value at roughly age 11.
8-8
-------
appropriate for these standards — one that recognizes effects in the used vehicle market — is
between -0.14 and -0.39, and most likely between -0.14 and -0.3.
Table 8-1: Policy Elasticities Corresponding to Selected Demand and Scrappage Elasticities.


Scenario

Effect of 1% Increase in Generalized Cost of






New Vehicles


Vehicle Demand Elasticities
Quantities (% changes)

New-








Vehicle
Cross-Price
Outside
Scrappage
New


Average

Demand
New/Used
Option
Elasticity
(Policy)
Used
All
age
A
-0.40
0.03
0
-0.70
-0.14
0.01
0.00
0.09
B
-0.40
0.03
-0.05
-0.70
-0.17
-0.04
-0.05
0.08
C
-0.40
0.03
-0.14
-0.70
-0.23
-0.10
-0.11
0.08
D
-0.80
0.05
-0.05
-0.70
-0.25
-0.04
-0.05
0.15
E
-1.27
0.09
-0.14
-0.70
-0.39
-0.12
-0.14
0.21
F
-0.40
0.03
-0.05
-0.20
-0.14
-0.06
-0.06
0.07
G
-0.40
0.03
-0.05
-1.20
-0.19
-0.03
-0.04
0.08
H
-0.80
0.05
-0.05
-0.20
-0.19
-0.07
-0.08
0.12
I
-0.80
0.05
-0.05
-1.20
-0.27
-0.03
-0.04
0.15
Note: The policy elasticities, italicized, are the effects in the new vehicle market, taking into account interactions
with the used vehicle market and scrappage
This table is Table 7-2 from U.S. Environmental Protection Agency (2021), "The Effects of New-Vehicle Price
Changes on New- and Used-Vehicle Markets and Scrappage," EPA-420-R-21-019,
A report submitted in comments from the New York University Institute for Policy
Integrity28 as well as comments provided by the Center for Biological Diversity et al.29
summarize studies by whether they are short-run or long-run, and when they were conducted.
These assessments also point to using a smaller elasticity (in absolute value) than even the -0.4
used as a sensitivity case in the NPRM.
The elasticity does not affect whether the sales are projected to increase or decrease, but it
does affect the magnitude of those increases: a 1 percent change in sales for a 1 percent change
in net price is a larger effect than a 0.4 percent change for a 1 percent change in net price. Based
on the RTI review and analysis, as well as summaries of the literature provided by commenters,
there appears to be agreement that newer estimates that account for long-term adjustments
indicate an elasticity much smaller (in absolute value) than -1, and likely smaller than -0.4. For
this final rule, EPA is using as its primary estimate for the new-vehicle demand elasticity a value
of -0.4, to facilitate comparison with the NPRM. We recognize that these assessments point to a
more inelastic value than even -0.4, such that a value of -0.4 leads to conservatively large
estimates of sales effect, and thus we also conduct a sensitivity analysis of -0.15, to encompass
what appears to be the plausible range. For further comparison with the NPRM, we also include
a sensitivity using the NPRM central-case elasticity of-1; as discussed above, though, use of this
value does not seem to be supported in recent literature.
CCEMS also makes use of a dynamic fleet share model (FRIA p. 877) that estimates,
separately, the shares of passenger cars and light trucks based on vehicle characteristics, and then
adjusts them so that the market shares sum to one; see RIA Chapter 4.1.4.3. The model also
includes the effects of the standards on vehicle scrappage based on a statistical analysis (FRIA
starting p. 926). The model looks for associations between age, change in new vehicle prices,
fuel prices, cost per mile of driving, and macroeconomic measures and the scrappage rate, with
8-9
-------
different equations for cars, SUVs/vans, and pickups. Because the scrappage model was revised
from the version in the SAFE proposal due to public comments received, the current version has
not been subject to review/EPA's project to review new vehicle demand elasticities is also
reviewing the literature on the relationship between new and used vehicle markets and
scrappage.
With the exception of the demand elasticity, as discussed above, EPA is maintaining the
NPRM assumptions for its modeling.
8.1.3 Sales Impacts
With the modeling assumption, described in Chapter 8.1.2, that vehicle buyers consider 2.5
years of future fuel consumption in the purchase decision and a new-vehicle demand elasticity of
-0.4, the sales impacts projected by the model are in Table 8-2. Vehicle sales decrease by
roughly 1 percent compared to sales in the baseline SAFE rule.
Table 8-2: Sales Impacts, 2.5 Years of Fuel Savings in Net Price, Demand Elasticity -0.4
Year
SAFE (Pre-existing)
Standards
Final Rule
Difference
Percent Change
2022
16,498,879
16,451,086
-47,793
-0.3%
2023
17,272,407
17,193,809
-78,598
-0.5%
2024
17,117,138
17,008,480
-108,658
-0.6%
2025
16,765,071
16,637,174
-127,897
-0.8%
2026
16,338,831
16,182,656
-156,175
-1.0%
2027
15,994,684
15,828,358
-166,326
-1.0%
2028
15,761,125
15,596,046
-165,079
-1.0%
2029
15,610,927
15,463,951
-146,976
-0.9%
2030
15,688,194
15,547,979
-140,215
-0.9%
2031
15,920,720
15,779,435
-141,285
-0.9%
2032
16,163,535
16,025,087
-138,448
-0.9%
2033
16,348,003
16,211,481
-136,522
-0.8%
2034
16,486,793
16,353,057
-133,736
-0.8%
2035
16,501,910
16,371,417
-130,493
-0.8%
Table 8-3 examines the impact of using an elasticity of -0.15 on sales. As expected, the smaller
(in absolute value) elasticity produces smaller sales impacts. Sales under both the SAFE (pre-
existing) standards and the final standards will be higher with the smaller elasticity.
-F
Details on the changes made to the scrappage model can be found in the SAFE FRIA.
8-10
-------
Table 8-3: Sales Impacts, 2.5 Years of Fuel Savings in Net Price, Demand Elasticity -0.15
Year
SAFE (Pre-existing)
Standards
Final Rule
Difference
Percent Change
2022
16,528,000
16,519,000
-10,000
-0.1%
2023
17,302,000
17,284,000
-18,000
-0.1%
2024
17,144,000
17,116,000
-28,000
-0.2%
2025
16,794,000
16,755,000
-39,000
-0.2%
2026
16,363,000
16,314,000
-49,000
-0.3%
2027
16,016,000
15,963,000
-53,000
-0.3%
2028
15,781,000
15,728,000
-53,000
-0.3%
2029
15,630,000
15,583,000
-47,000
-0.3%
2030
15,707,000
15,662,000
-45,000
-0.3%
2031
15,939,000
15,894,000
-45,000
-0.3%
2032
16,182,000
16,138,000
-44,000
-0.3%
2033
16,367,000
16,324,000
-44,000
-0.3%
2034
16,506,000
16,463,000
-43,000
-0.3%
2035
16,521,000
16,479,000
-42,000
-0.3%
Finally, Table 8-4 provides estimates using -1 as the demand elasticity, for purposes of
continuity with the NPRM. Sales impacts are larger, as expected. As discussed above, this
elasticity, and thus results based on it, are not supported by current literature.
Table 8-4: Sales Impacts, 2.5 Years of Fuel Savings in Net Price, Demand Elasticity -1
Year
SAFE (Pre-existing)
Standards
Final Rule
Difference
Percent Change
2022
16,387,000
16,323,000
-64,000
-0.4%
2023
17,157,000
17,037,000
-120,000
-0.7%
2024
16,997,000
16,813,000
-184,000
-1.1%
2025
16,631,000
16,373,000
-258,000
-1.6%
2026
16,198,000
15,874,000
-324,000
-2.0%
2027
15,855,000
15,501,000
-354,000
-2.2%
2028
15,630,000
15,279,000
-351,000
-2.2%
2029
15,492,000
15,179,000
-314,000
-2.0%
2030
15,575,000
15,276,000
-299,000
-1.9%
2031
15,807,000
15,506,000
-301,000
-1.9%
2032
16,053,000
15,758,000
-295,000
-1.8%
2033
16,236,000
15,945,000
-291,000
-1.8%
2034
16,377,000
16,092,000
-285,000
-1.7%
2035
16,394,000
16,116,000
-278,000
-1.7%
As discussed above, the use of 2.5 years by consumers for consideration of future fuel
consumption is smaller than the mean or median estimates in the Greene et al. (2018) meta-
analysis for consumer valuation of fuel savings, though it appears to reflect automakers'
perception of that value. In addition, it is possible that automakers and vehicle buyers may differ
in their practices relating to consumers' willingness to pay for fuel economy. If automakers
underestimate consumers' valuation of fuel economy, then sales may increase relative to the
baseline under the standards. EPA will continue to evaluate the sales impacts of the standards,
including the assumption on consumer valuation of future fuel savings.
8-11
-------
How easily new vehicle buyers will be willing to substitute EVs for ICEVs is a matter of
some uncertainty. With up-front costs dropping, the total cost of ownership for EVs is also
dropping and becoming more competitive with ICEVs. As shown in Table 4-26, our analysis
suggests that EV penetration under these standards is projected to increase from about 7 percent
in MY 2023 to about 17 percent in MY 2026. The transition to zero emission vehicles is
important for achieving climate goals; in addition, as discussed in Chapter 8.2.3, domestic
production of EVs is important for future competitiveness of the U.S. auto industry as other
markets also make this transition.
8.2 Employment Impacts
8.2.1 Conceptual Framework
Economic theory of labor demand indicates that employers affected by environmental
regulation may increase their demand for some types of labor, decrease demand for other types
of labor, or for still other types, not change it at all. A variety of conditions can affect
employment impacts of environmental regulation, including baseline labor market conditions
and employer and worker characteristics such as industry, region, and skill level.
A growing literature has investigated employment effects of environmental regulation.
Morgenstern et al. (2002)30 decompose the labor consequences in a regulated industry facing
increased abatement costs into three separate components. First, there is a demand effect caused
by higher production costs raising market prices. Higher prices reduce consumption (and
production) reducing demand for labor within the regulated industry. Second, there is a cost
effect where, as production costs increase, plants use more of all inputs, including labor, to
produce the same level of output. Third, there is a factor-shift effect where post-regulation
production technologies may have different labor intensities. These three effects outlined by
Morgenstern et al. (2002) provides the theoretical foundation for EPA's analysis of the impacts
of the regulation on labor throughout Chapter 8.2.29
Additional papers approach employment effects through similar frameworks. Berman and Bui
(2001) model two components that drive changes in firm-level labor demand: output effects and
substitution effects.31'8 If regulation causes marginal cost to increase, it will place upward
pressure on output prices, leading to a decrease in the quantity demanded, and resulting in a
decrease in production that they term the output effect. The substitution effect describes how,
holding output constant, regulation affects labor intensity of production. Deschenes describes
environmental regulations as requiring additional capital equipment for pollution abatement that
does not increase labor productivity.32 These higher production costs induce regulated firms to
reduce output and decrease labor demand (an output effect) while simultaneously shifting away
from the use of more expensive capital towards increased labor demand (a substitution effect).
At the industry level, labor demand is more likely to be responsive to regulatory costs if: (1) the
elasticity of labor demand is high relative to the elasticity of labor supply, and (2) labor costs are
a large share of total production costs.33 Labor demand might also respond to regulation if
compliance activities change labor intensity in production.
g Berman and Bui (2001) also discuss a third component, the impact of regulation on factor prices, but conclude that
this effect is unlikely to be important for large competitive factor markets, such as labor and capital.
8-12
-------
To study labor demand impacts empirically, researchers have compared employment levels at
facilities subject to an environmental regulation to employment levels at similar facilities not
subject to that environmental regulation; some studies find no employment effects, and others
find statistically significant, usually small differences. For example, see Berman and Bui,
Greenstone (2002), Ferris et al. (2014), Walker (2013), and Curtis (2018, 2020).34
Workers affected by changes in labor demand due to regulation may experience a variety of
impacts including job gains or involuntary job loss and unemployment. Localized reductions in
employment may adversely impact individuals and communities just as localized increases may
have positive impacts. Workforce adjustments in response to decreases in labor demand can be
costly to firms as well as workers, so employers may choose to adjust their workforce over time
through natural attrition or reduced hiring, rather than incur costs associated with job separations
(see, for instance, Curtis (2018) and Hafstead and Williams (2018).35
In addition to impacts on labor demand in directly regulated industries, impacts on related
industries are possible too. Industries operating upstream or downstream from the regulated
industries may experience changes in labor demand. For example, as described elsewhere in this
RIA, we expect the rule to cause a small decline in extracting, refining, transporting, and storing
of petroleum fuels, and a small increase in electricity generation which may have consequences
for labor demand in those upstream industries. Or lower per-mile fuel costs could lead to
increases in demand for ride-sharing or ride-hailing services and cause increases in demand for
drivers in those jobs. Firms producing substitutes or complements to the goods produced by the
regulated industry may also experience changes in demand for labor. For example, the expected
decline in gas station visits may lead to reduced demand for labor in that sector. The magnitude
of these impacts depends on a variety of factors including the labor intensities of the related
sectors as well as the nature of the linkages (which can be reflected in measures of elasticity)
between them and the regulated firms.
As suggested in this discussion, the overall employment effects of environmental regulation
are difficult to estimate. Estimation is difficult due to the multitude of small changes in different
sectors related to the regulated industry, and because employment impacts are hard to
disentangle from other economic changes and business decisions that affect employment, over
time and across regions and industries. If the U.S. economy is at full employment, even a large-
scale environmental regulation is unlikely to have a noticeable impact on aggregate net
employment.11 Instead, labor is likely primarily to be reallocated from one productive use to
another, and net national employment effects from environmental regulation will be small and
transitory (e.g., as workers move from one job to another).36 However, localized reductions in
employment may adversely impact individuals and communities just as localized increases may
have positive impacts. If the economy is operating at less than full employment, economic theory
does not clearly indicate the direction or magnitude of the net impact of environmental regulation
on employment; it could cause either a short-run net increase or short-run net decrease.37
Because it is challenging to know the state of the macroeconomy when these standards
become effective, and also because of the difficulties of modeling impacts on employment in a
h Full employment is a conceptual target for the economy where everyone who wants to work and is available to do
so at prevailing wages is actively employed. The unemployment rate at Ml employment is not zero.
8-13
-------
complex national economy, we focus our analysis on the direct impacts in closely affected
sectors, as described in the next section.
8.2.2	How Employment Impacts were Modeled
The SAFE FRIA (starting p. 1067) describes the calculation of employment impacts for three
sets of affected sectors: automotive dealers, final assembly labor and parts production, and fuel-
saving (or GHG-reducing) technology labor. The first two of these (automobile dealers and final
assembly) are examples of demand-effect employment, while the third (technology labor)
reflects cost-effect employment. For automotive dealers, the model estimates the hours involved
in each new vehicle sale. Estimating the labor involved in final assembly used average labor
hours per vehicle at a sample of U.S. assembly plants, adjusted by the ratio of vehicle assembly
manufacturing employment to employment for total vehicle and equipment manufacturing for
new vehicles. Finally, for fuel-saving technology labor, the analysis calculated the average
revenue per job-year for automakers, and used the revised revenue estimates for calculation of
the change in job-years. As with the NPRM, these estimates are still in use for this final rule.
8.2.3	Employment Impacts
Table 8-5 below provides the results of these calculations, combined for these three sectors. It
indicates a very small effect on employment, becoming increasingly positive over time, when
using the estimate that both automakers and vehicle buyers take 2.5 years of fuel savings into
consideration in the purchase decision. Employment increases by roughly 2.5 percent: even
though sales decrease slightly, positive cost effect due to increased technology costs outweighs
the negative demand effect due to decreased sales.
Table 8-5: Employment Impacts, Based on Sales Estimates in Table 8-2 (Demand Elasticity -0.4)
Year
SAFE (Pre-existing)
Standards
Final Rule
Difference
Percent Difference
2022
1,156,000
1,161,000
6,000
0.5%
2023
1,210,000
1,220,000
10,000
0.8%
2024
1,200,000
1,216,000
16,000
1.3%
2025
1,177,000
1,195,000
18,000
1.5%
2026
1,147,000
1,170,000
23,000
2.0%
2027
1,122,000
1,148,000
26,000
2.3%
2028
1,105,000
1,131,000
26,000
2.4%
2029
1,093,000
1,119,000
25,000
2.3%
2030
1,097,000
1,123,000
26,000
2.3%
2031
1,113,000
1,139,000
26,000
2.3%
2032
1,129,000
1,155,000
26,000
2.3%
2033
1,141,000
1,168,000
27,000
2.3%
2034
1,150,000
1,177,000
27,000
2.3%
2035
1,151,000
1,177,000
26,000
2.3%
Table 8-6 shows the effects of using the sales estimates based on an elasticity of -0.15, as
shown in Table 8-3. As with the sales impacts, employment under both the SAFE program and
this rule are higher with the smaller (in absolute value) elasticity. The effects on employment due
to the standards, with this lower elasticity, are positive in all years, and almost the same as those
using the elasticity of -0.4. As with that analysis, the positive cost effect outweighs the negative
demand effect across all analyzed years.
8-14
-------
Table 8-6: Employment Impacts, Based on Sales Estimates in Table 8-3 (Demand Elasticity -0.15)
Year
SAFE (Pre-existing)
Standards
Final Rule
Difference
Percent Difference
2022
1,161,000
1,164,000
4,000
0.3%
2023
1,217,000
1,224,000
7,000
0.6%
2024
1,208,000
1,221,000
13,000
1.1%
2025
1,185,000
1,203,000
18,000
1.5%
2026
1,157,000
1,181,000
24,000
2.1%
2027
1,133,000
1,160,000
27,000
2.4%
2028
1,115,000
1,142,000
27,000
2.4%
2029
1,104,000
1,129,000
25,000
2.3%
2030
1,108,000
1,133,000
25,000
2.3%
2031
1,123,000
1,149,000
26,000
2.3%
2032
1,139,000
1,165,000
26,000
2.3%
2033
1,152,000
1,178,000
26,000
2.2%
2034
1,161,000
1,187,000
26,000
2.2%
2035
1,161,000
1,187,000
25,000
2.2%
Finally, Table 8-7 results using a demand elasticity of-1. The results show impacts ranging
from 0 to 0.7 percent. As discussed in the context of vehicle sales analysis, while these estimates
are presented for continuity with the NPRM analysis, our current assessment does not support
the use of results based on this elasticity.
Table 8-7: Employment Impacts, Based on Sales Estimates in Table 8-4 (Demand Elasticity -1)
Year
SAFE (Pre-existing)
Standards
Final Rule
Difference
Percent Difference
2022
1,151,000
1,151,000
0
0.0%
2023
1,207,000
1,208,000
0
0.0%
2024
1,198,000
1,200,000
2,000
0.2%
2025
1,174,000
1,176,000
2,000
0.2%
2026
1,145,000
1,150,000
5,000
0.4%
2027
1,121,000
1,127,000
5,000
0.5%
2028
1,105,000
1,110,000
6,000
0.5%
2029
1,094,000
1,100,000
6,000
0.6%
2030
1,098,000
1,106,000
7,000
0.7%
2031
1,114,000
1,121,000
7,000
0.7%
2032
1,130,000
1,138,000
8,000
0.7%
2033
1,143,000
1,151,000
8,000
0.7%
2034
1,152,000
1,160,000
8,000
0.7%
2035
1,153,000
1,161,000
8,000
0.7%
If automakers underestimate consumers' valuation of fuel economy, as noted in Chapter 8.2.3,
then demand-effect employment is likely to be higher, and employment impacts are likely to be
more positive.
As mentioned, we are only providing partial estimates of employment impacts in the directly
regulated sector, plus the impacts for automotive dealers. These do not include economy-wide
labor impacts. As discussed in Chapter 8.2.1, economy-wide impacts on employment are
generally driven by broad macroeconomic effects. It also does not reflect employment effects
due to impacts on related sectors other than car dealerships (those that are upstream or
downstream, or producing substitutes and complements). For example, we have not estimated
8-15
-------
the impacts of reduced spending on fuel consumption. Those changes may lead to some
reductions in employment in gas stations, and some increases in other sectors to which people
reallocate those expenditures.
Electrification of the vehicle fleet is likely to affect both the number and the nature of
employment in the auto and parts sectors and related sectors, such as providers of charging
infrastructure. The kinds of jobs in auto manufacturing are expected to change: for instance,
there will be no need for exhaust system assembly for EVs, while wiring will become more
complex. The effect on total employment for auto manufacturing is uncertain: some suggest that
fewer workers will be needed because BEVs have fewer moving parts,38 while others estimate
that the labor-hours involved in BEVs is almost identical to that for ICE vehicles.39 Effects in the
supply chain, as SAFE and the Alliance noted, depend on where goods in the supply chain are
developed. Blue-Green Alliance, BICEP, Ceres, and SAFE all argue that developing EVs in the
U.S. is critical for domestic employment and for the global competitiveness of the U.S. in the
future auto industry: as other countries are moving rapidly to develop EVs, the U.S. auto industry
risks falling behind. EPA will continue to assess changes in employment as electrification of the
auto industry proceeds.
8.3 Environmental Justice
Executive Order 12898 (59 FR 7629, February 16, 1994) establishes federal executive policy
on environmental justice. It directs federal agencies, to the greatest extent practicable and
permitted by law, to make achieving environmental justice part of their mission by identifying
and addressing, as appropriate, disproportionately high and adverse human health or
environmental effects of their programs, policies, and activities on minority populations and low-
income populations in the United States. EPA defines environmental justice as the fair treatment
and meaningful involvement of all people regardless of race, color, national origin, or income
with respect to the development, implementation, and enforcement of environmental laws,
regulations, and policies.1
Executive Order 14008 (86 FR 7619, February 1, 2021) also calls on federal agencies to make
achieving environmental justice part of their respective missions "by developing programs,
policies, and activities to address the disproportionately high and adverse human health,
environmental, climate-related and other cumulative impacts on disadvantaged communities, as
well as the accompanying economic challenges of such impacts." It declares a policy "to secure
environmental justice and spur economic opportunity for disadvantaged communities that have
1 Fair treatment means that "no group of people should bear a disproportionate burden of environmental harms and
risks, including those resulting from the negative environmental consequences of industrial, governmental and
commercial operations or programs and policies." Meaningful involvement occurs when "1) potentially affected
populations have an appropriate opportunity to participate in decisions about a proposed activity [e.g., rulemaking]
that will affect their environment and/or health; 2) the public's contribution can influence [the EPA's rulemaking]
decision; 3) the concerns of all participants involved will be considered in the decision-making process; and 4) [the
EPA will] seek out and facilitate the involvement of those potentially affected" A potential EJ concern is defined as
"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." See "Guidance on Considering Environmental Justice During the Development of an
Action." Environmental Protection Agency, www.epa.gov/environmentaljustice/guidanceconsidering-
environmental-justice-duringdevelopment-action. See also https://www.epa.gov/environmentaljustice.
8-16
-------
been historically marginalized and overburdened by pollution and under-investment in housing,
transportation, water and wastewater infrastructure and health care."
Under Executive Order 13563, federal agencies may consider equity, human dignity, fairness,
and distributional considerations in their regulatory analyses, where appropriate and permitted by
law.
EPA's 2016 "Technical Guidance for Assessing Environmental Justice in Regulatory
Analysis" provides recommendations on conducting the highest quality analysis feasible,
recognizing that data limitations, time and resource constraints, and analytic challenges will vary
by media and regulatory context.40
When assessing the potential for disproportionately high and adverse health or environmental
impacts of regulatory actions on populations of color, low-income populations, tribes, and/or
indigenous peoples, the EPA strives to answer three broad questions: (1) Is there evidence of
potential EJ concerns in the baseline (the state of the world absent the regulatory action)?
Assessing the baseline will allow the EPA to determine whether pre-existing disparities are
associated with the pollutant(s) under consideration (e.g., if the effects of the pollutant(s) are
more concentrated in some population groups). (2) Is there evidence of potential EJ concerns for
the regulatory option(s) under consideration? Specifically, how are the pollutant(s) and its effects
distributed for the regulatory options under consideration? And, (3) do the regulatory option(s)
under consideration exacerbate or mitigate EJ concerns relative to the baseline? It is not always
possible to quantitatively assess these questions.
EPA's 2016 Technical Guidance does not prescribe or recommend a specific approach or
methodology for conducting an environmental justice analysis, though a key consideration is
consistency with the assumptions underlying other parts of the regulatory analysis when
evaluating the baseline and regulatory options. Where applicable and practicable, the Agency
endeavors to conduct such an analysis. Going forward, EPA is committed to conducting
environmental justice analysis for rulemakings based on a framework similar to what is outlined
in EPA's Technical Guidance, in addition to investigating ways to further weave environmental
justice into the fabric of the rulemaking process.
8.3.1 GHG Impacts
In 2009, under the Endangerment and Cause or Contribute Findings for Greenhouse Gases
Under Section 202(a) of the Clean Air Act ("Endangerment Finding"), the Administrator
considered how climate change threatens the health and welfare of the U.S. population. As part
of that consideration, she also considered risks to minority and low-income individuals and
communities, finding that certain parts of the U.S. population may be especially vulnerable based
on their characteristics or circumstances. These groups include economically and socially
disadvantaged communities; individuals at vulnerable lifestages, such as the elderly, the very
young, and pregnant or nursing women; those already in poor health or with comorbidities; the
disabled; those experiencing homelessness, mental illness, or substance abuse; and/or Indigenous
or minority populations dependent on one or limited resources for subsistence due to factors
including but not limited to geography, access, and mobility.
Scientific assessment reports produced over the past decade by the U.S. Global Change
Research Program (USGCRP),41'42 the Intergovernmental Panel on Climate Change
8-17
-------
(IPCC),43'44'45'46 and the National Academies of Science, Engineering, and Medicine47'48 add
more evidence that the impacts of climate change raise potential environmental justice concerns.
These reports conclude that poorer or predominantly non-White communities can be especially
vulnerable to climate change impacts because they tend to have limited adaptive capacities and
are more dependent on climate-sensitive resources such as local water and food supplies, or have
less access to social and information resources. Some communities of color, specifically
populations defined jointly by ethnic/racial characteristics and geographic location, may be
uniquely vulnerable to climate change health impacts in the United States. In particular, the 2016
scientific assessment on the Impacts of Climate Change on Human Health49 found with high
confidence that vulnerabilities are place- and time-specific, lifestages and ages are linked to
immediate and future health impacts, and social determinants of health are linked to greater
extent and severity of climate change-related health impacts.
8.3.1.1 Effects on Specific Populations of Concern
Individuals living in socially and economically disadvantaged communities, such as those
living at or below the poverty line or who are experiencing homelessness or social isolation, are
at greater risk of health effects from climate change. This is also true with respect to people at
vulnerable lifestages, specifically women who are pre- and perinatal, or are nursing; in utero
fetuses; children at all stages of development; and the elderly. Per the Fourth National Climate
Assessment, "Climate change affects human health by altering exposures to heat waves, floods,
droughts, and other extreme events; vector-, food- and waterborne infectious diseases; changes in
the quality and safety of air, food, and water; and stresses to mental health and well-being."50
Many health conditions such as cardiopulmonary or respiratory illness and other health impacts
are associated with and exacerbated by an increase in greenhouse gases and climate change
outcomes, which is problematic as these diseases occur at higher rates within vulnerable
communities. Importantly, negative public health outcomes include those that are physical in
nature, as well as mental, emotional, social, and economic.
To this end, the scientific assessment literature, including the aforementioned reports,
demonstrates that there are myriad ways in which these populations may be affected at the
individual and community levels. Individuals face differential exposure to criteria pollutants, in
part due to the proximities of highways, trains, factories, and other major sources of pollutant-
emitting sources to less-affluent residential areas. Outdoor workers, such as construction or
utility crews and agricultural laborers, who frequently are comprised of already at-risk groups,
are exposed to poor air quality and extreme temperatures without relief. Furthermore, individuals
within EJ populations of concern face greater housing, clean water, and food insecurity and bear
disproportionate economic impacts and health burdens associated with climate change effects.
They have less or limited access to healthcare and affordable, adequate health or homeowner
insurance. Finally, resiliency and adaptation are more difficult for economically disadvantaged
communities: They have less liquidity, individually and collectively, to move or to make the
types of infrastructure or policy changes to limit or reduce the hazards they face. They frequently
are less able to self-advocate for resources that would otherwise aid in building resilience and
hazard reduction and mitigation.
The assessment literature cited in EPA's 2009 and 2016 Endangerment Findings, as well as
Impacts of Climate Change on Human Health,48 also concluded that certain populations and life
stages, including children, are most vulnerable to climate-related health effects. The assessment
8-18
-------
literature produced from 2016 to the present strengthens these conclusions by providing more
detailed findings regarding related vulnerabilities and the projected impacts youth may
experience. These assessments - including the Fourth National Climate Assessment (2018) and
The Impacts of Climate Change on Human Health in the United States (2016) - describe how
children's unique physiological and developmental factors contribute to making them
particularly vulnerable to climate change. Impacts to children are expected from heat waves, air
pollution, infectious and waterborne illnesses, and mental health effects resulting from extreme
weather events. In addition, children are among those especially susceptible to allergens, as well
as health effects associated with heat waves, storms, and floods. Additional health concerns may
arise in low-income households, especially those with children, if climate change reduces food
availability and increases prices, leading to food insecurity within households.
The Impacts of Climate Change on Human Healthalso found that some communities of
color, low-income groups, people with limited English proficiency, and certain immigrant groups
(especially those who are undocumented) live with many of the factors that contribute to their
vulnerability to the health impacts of climate change. While difficult to isolate from related
socioeconomic factors, race appears to be an important factor in vulnerability to climate-related
stress, with elevated risks for mortality from high temperatures reported for Black or African
American individuals compared to White individuals after controlling for factors such as air
conditioning use. Moreover, people of color are disproportionately exposed to air pollution based
on where they live, and disproportionately vulnerable due to higher baseline prevalence of
underlying diseases such as asthma, so climate exacerbations of air pollution are expected to
have disproportionate effects on these communities.
Native American Tribal communities possess unique vulnerabilities to climate change,
particularly those impacted by degradation of natural and cultural resources within established
reservation boundaries and threats to traditional subsistence lifestyles. Tribal communities whose
health, economic well-being, and cultural traditions depend upon the natural environment will
likely be affected by the degradation of ecosystem goods and services associated with climate
change. The IPCC indicates that losses of customs and historical knowledge may cause
communities to be less resilient or adaptable.51 The Fourth National Climate Assessment (2018)
noted that while Indigenous peoples are diverse and will be impacted by the climate changes
universal to all Americans, there are several ways in which climate change uniquely threatens
Indigenous peoples' livelihoods and economies52. In addition, there can institutional barriers to
their management of water, land, and other natural resources that could impede adaptive
measures.
For example, Indigenous agriculture in the Southwest is already being adversely affected by
changing patterns of flooding, drought, dust storms, and rising temperatures leading to increased
soil erosion, irrigation water demand, and decreased crop quality and herd sizes. The
Confederated Tribes of the Umatilla Indian Reservation in the Northwest have identified climate
risks to salmon, elk, deer, roots, and huckleberry habitat. Housing and sanitary water supply
infrastructure are vulnerable to disruption from extreme precipitation events.
NCA4 noted that Indigenous peoples often have disproportionately higher rates of asthma,
cardiovascular disease, Alzheimer's, diabetes, and obesity, which can all contribute to increased
vulnerability to climate-driven extreme heat and air pollution events. These factors also may be
8-19
-------
exacerbated by stressful situations, such as extreme weather events, wildfires, and other
circumstances.
NCA4 and IPCC AR553 also highlighted several impacts specific to Alaskan Indigenous
Peoples. Coastal erosion and permafrost thaw will lead to more coastal erosion, exacerbated risks
of winter travel, and damage to buildings, roads, and other infrastructure - these impacts on
archaeological sites, structures, and objects that will lead to a loss of cultural heritage for
Alaska's Indigenous people. In terms of food security, the NCA discussed reductions in suitable
ice conditions for hunting, warmer temperatures impairing the use of traditional ice cellars for
food storage, and declining shellfish populations due to warming and acidification. While the
NCA also noted that climate change provided more opportunity to hunt from boats later in the
fall season or earlier in the spring, the assessment found that the net impact was an overall
decrease in food security.
In addition, the U.S. Pacific Islands and the indigenous communities that live there are also
uniquely vulnerable to the effects of climate change due to their remote location and geographic
isolation. They rely on the land, ocean, and natural resources for their livelihoods, but face
challenges in obtaining energy and food supplies that need to be shipped in at high costs. As a
result, they face higher energy costs than the rest of the nation and depend on imported fossil
fuels for electricity generation and diesel. These challenges exacerbate the climate impacts that
the Pacific Islands are experiencing. NCA4 notes that Indigenous peoples of the Pacific are
threatened by rising sea levels, diminishing freshwater availability, and negative effects to
ecosystem services that threaten these individuals' health and well-being.
8.3.2 Non-GHG Impacts
In addition to significant climate change benefits, the standards will also impact non-GHG
emissions. In general, we expect small non-GHG emissions reductions from upstream sources
related to refining petroleum fuels. We also expect small increases in emissions from upstream
electricity generating units (EGUs). An increase in emissions from coal- and NG-fired electricity
generation to meet increased EV electricity demand could result in adverse EJ impacts. For on-
road light-duty vehicles, the standards will reduce total non-GHG tailpipe emissions, though we
expect small increases in some non-GHG emissions in the years immediately following
implementation of this rule, followed by growing decreases in non-GHG emissions in later years.
This is due to our projections about the gasoline-fueled LD vehicle population in the final rule
scenario, including decreased scrappage of older vehicles. See Chapter 5.1.2 for more detail on
the estimated non-GHG emissions impacts of the rule.-" As discussed in Section I.A.I of the
Preamble Executive Summary, future EPA actions that would result in increased ZEVs and
emissions reductions from the power sector would more significantly change the non-GHG
impacts of transportation and electricity generation, and those impacts will be analyzed in more
detail in those future actions.
There is evidence that communities with EJ concerns could be impacted by the non-GHG
emissions from light-duty vehicles.54 Numerous studies have found that environmental hazards
such as air pollution are more prevalent in areas where populations of color and low-income
8-20
-------
populations represent a higher fraction of the population compared with the general
population.55'56'57 Consistent with this evidence, a recent study found that most anthropogenic
sources of PM2.5, including industrial sources, and light- and heavy-duty vehicle sources,
disproportionately affect people of color.58
Analyses of communities in close proximity to upstream sources, such as EGUs, have found
that a higher percentage of communities of color and low-income communities live near these
sources when compared to national averages.59 Vulnerable populations near upstream refineries
may experience potential disparities in pollution-related health risk from that source.60 In this
rule we expect that small increases in non-GHG emissions from EGUs and small reductions in
petroleum-sector emissions would lead to small changes in exposure to these non-GHG
pollutants for people living in the communities near these facilities.
There is also substantial evidence that people who live or attend school near major roadways
are more likely to be of a non-White race, Hispanic ethnicity, and/or low socioeconomic
status.61'62 We would expect that communities near roads will benefit from reductions of non-
GHG pollutants as fuel efficiency improves and the use of ZEVs (such as full battery electric
vehicles) increases, though projections about the gasoline-fueled LD vehicle population,
including decreased scrappage, may offset some of these emission reductions, especially in the
years immediately after finalization of the standards.
Although proximity to an emissions source is a useful indicator of potential exposure, it is
important to note that the impacts of emissions from both upstream and tailpipe sources are not
limited to communities in close proximity to these sources. The effects of potential increases and
decreases in emissions from the sources affected by this rule might also be felt many miles away,
including in communities with EJ concerns. The spatial extent of these impacts from upstream
and tailpipe sources depend on a range of interacting and complex factors including the amount
of pollutant emitted, atmospheric chemistry and meteorology.
In summary, we expect this rule will, over time, result in reductions of non-GHG tailpipe
emissions and emissions from upstream refinery sources. We also project that the rule will result
in small increases of non-GHG emissions from upstream EGU sources. Overall, there are
substantial PIvfc.s-related health benefits associated with the non-GHG emissions reductions that
this rule will achieve. The benefits from these emissions reductions, as well as the adverse
impacts associated with the emissions increases, could potentially impact communities with EJ
concerns, though not necessarily immediately and not equally in all locations. For this
rulemaking, the air quality information needed to perform a quantified analysis of the
distribution of such impacts was not available. We therefore recommend caution when
interpreting these broad, qualitative observations. EPA intends to develop a future rule to control
emissions of GHGs, criteria pollutants, and air toxic pollutants from light-duty vehicles for
model years beyond 2026. We are considering how to project air quality impacts from the
changes in non-GHG emissions for that future rulemaking (see Section V.C of the Preamble) and
how to consider potential EJ concerns that may stem from them.
8.4 Affordability and Equity Impacts
Executive Order 13985 defines equity as "the consistent and systematic fair, just, and
impartial treatment of all individuals, including individuals who belong to underserved
communities that have been denied such treatment, such as Black, Latino, and Indigenous and
8-21
-------
Native American persons, Asian Americans and Pacific Islanders and other persons of color;
members of religious minorities; lesbian, gay, bisexual, transgender, and queer (LGBTQ+)
persons; persons with disabilities; persons who live in rural areas; and persons otherwise
adversely affected by persistent poverty or inequality." In the context of transportation, Guo et al.
(2020) consider both horizontal and vertical equity: horizontal equity involves even distribution
of resources in the population, while vertical or social equity "aims to provide services to those
who need them most."63 These definitions suggest that equity involves reducing disparities in
both resources and access across income and demographic groups.
While a full assessment of the effects of these standards on equity is not available, those
impacts depend in part on their effects on the affordability of vehicles and impacts on lower-
income households.
Access to transportation improves the ability of people, including those with low income, to
pursue jobs, education, health care, and necessities of daily life such as food and housing. These
standards might affect affordability of vehicles and their impacts on low-income households in
particular. We acknowledge that vehicles, especially household vehicle ownership, are only a
portion of the larger issue of access to transportation and mobility services, which also takes into
consideration public transportation and urban design. In addition, online working and shopping
may provide alternative ways to accomplish some goals, for those with stable access to internet
services. Though these issues are inextricably linked, the following discussion focuses on effects
related to private vehicle ownership and use. We also acknowledge that the emissions of
vehicles, both local pollutants and greenhouse gases, can have disproportionate impacts on
lower-income and minority communities; see Chapter 8.3 and Preamble Section VII.L. for
further discussion of these topics.
The SAFE rule discussed affordability primarily as relating to the up-front purchase price of a
new vehicle: if the up-front price increased, due to addition of fuel-saving technologies, then
vehicles became less affordable. E.g., "technologies added to comply with fuel economy
standards have an impact on vehicle prices, and, by extension, on the affordability of newer,
safer vehicles, and therefore on the rates at which newer vehicles are acquired and older, less
safe vehicles are retired from use" (85 FR 24742). While this is one use of the concept of
affordability, it is not the only one.
Hutchens et al. (2021)64 and the TSD for the MTE Proposed Determination,65 Chapter 4.3.1,
discuss the lack of specificity in the concept of affordability in academic literature. For instance,
Bradley (2008)66 identified affordability as "a vague concept... When pundits use the word
'afford,' there is no clear definition of affordability; it is at best a subjective notion." Bartl
(2010)67 declared that "affordability is a new 'alien' concept penetrating the field of contract and
consumer law." Researchers have nevertheless grappled with attempting to define the term for
goods including energy, food, telephone service, health insurance, and housing. Some themes
that appear in the different definitions of affordability include:
•	Instead of focusing on the traditional economic concept of willingness to pay, any
consideration of affordability must also consider the ability to pay for a socially
defined minimum level of a good, especially of a necessity.
•	Although the ability to pay is often based on the proportion of income devoted to
expenditures on a particular good, this ratio approach is widely criticized for not
8-22
-------
considering expenditures on other possibly necessary goods, for not considering
quality differences in the good, and for not considering heterogeneity of consumer
preferences for the good.
• Assessing affordability should take into account both the short-term costs and long-
term costs associated with consumption of a particular good.
These themes were all developed in the context of goods typically deemed essential, such as
food and housing. There is very little literature applying the concept of affordability to
transportation, much less to vehicle ownership. Thakuriah and Liao (2006)68 attempted to define
ability to pay for transportation expenditures, but do not offer a definition of affordable
transportation. A report by the Manhattan Strategy Group for the Department of Transportation
and the Department of Housing and Urban Development (HUD) (Schanzenbach and
McGranahan, 2012)69 attempted to create metrics of various types of vehicle costs to be included
in HUD's Location Affordability Index, which considers housing and transportation costs based
on location. However, this report also did not attempt to define vehicle affordability.
It is not clear how to identify the socially acceptable minimum level of transportation service.
It seems reasonable to assume that such a socially acceptable minimum level should allow access
to employment, education, and basic services like buying food, but it is not clear where
consumption of transportation moves from necessity to optional. Normatively defining the
minimum adequate level of transportation consumption is difficult given the heterogeneity of
consumer preferences and living situations. As a result, it is challenging to define how much
residual income should remain with each household after transportation expenditures. It is
therefore not surprising that academic and policy literature have largely avoided attempting to
define transportation affordability.
We do not here offer a quantitative measure of the affordability of new vehicles. Instead, as in
the NPRM, we follow the approach in the Proposed Determination for the Midterm Evaluation70
of considering four questions that relate to the effects of the LDV GHG standards on new vehicle
affordability and equity: how the standards affect low-income households; how the standards
affect the used vehicle market; how the standards affect access to credit; and how the standards
affect the low-priced vehicle segment. These questions are intended to examine some ways in
which the standards might influence the distribution of access to transportation across the public,
especially those who might disproportionately suffer from low access.
8.4.1 Effects on Lower-Income Households
As noted in Chapter 8.4, there is no commonly accepted definition of affordable access to
transportation. Access to transportation involves access to any form of transportation, not only
vehicle ownership and access, but also public transportation, ride-sharing, and ride-hailing
services. Within vehicle ownership, access does not involve only the up-front costs of purchasing
a vehicle, but also the operating and maintenance costs of a vehicle. Trying to define a socially
acceptable minimum level for access to transportation services is even more difficult, because
such requirements will vary with geography and personal needs. People in rural areas are
unlikely to be able to rely on public transit, for instance. Though nutritious food is a generally
acknowledged necessity, people who live in urban food deserts may suffer in health and quality
of life due to the transportation time and cost of accessing adequate and nutritious food. On the
8-23
-------
other hand, those who live in areas with good, inexpensive public transportation and easy access
to stores and other desired destinations may be able to rely on public transportation, bicycling, or
walking to meet their needs, and not need a personally owned vehicle. How the standards might
affect affordable access to transportation is thus a complex question.
A first point to note is that the standards on average are projected to have fuel savings over
the lifetime of the vehicles that exceed the up-front costs (see Chapter 6 and Preamble Section
VII. J). Thus, on average, the standards are expected to reduce the total cost of ownership of new
vehicles subject to the standards. This metric on its own implies that vehicle affordability is
enhanced by the standards. This metric is nevertheless likely to be overly simplistic for the
purposes of understanding the distributional effects of the standards on equitable access to
transportation, and specifically the effects on lower-income households. It does not measure, for
instance, who is likely to get the benefits of the fuel savings, and who bears the increased up-
front costs of the vehicle. If those groups are different, then it is not initially obvious who earns
the net benefits.
It should also be noted that low-income households, defined as households having annual
after-tax income below the current-year's median after-tax income level, are much more likely to
have used vehicles than new ones. For instance, 70 percent of new vehicle buyers have income
above $75,000;71 median household income in 2019 was about $68,700.72 Thus, lower-income
households will eventually feel the effects of reduced fuel consumption in new vehicles over
time, when those vehicles are resold on the used market. Lower-income households are also
more likely to experience the effects of price changes in the used vehicle market as explained in
Chapter 8.4.2, below. As discussed in Preamble Section VII. J., purchasers of used vehicles
subject to these standards are likely to experience greater net gains than the purchasers of new
vehicles, because the up-front cost of a vehicle depreciates much more rapidly than do the fuel
savings.
A few recent papers have asked whether fuel economy standards are progressive or regressive
— that is, having greater beneficial effects or more adverse effect on lower-income households
than on higher-income households. Jacobsen (2013)73 finds, for the flat (not footprint-based)
standard used in the CAFE program before MY 2011, the standards were regressive by implicitly
discouraging more desired larger vehicles. The subsequent use of the footprint-based standard is
intended to reduce the disincentives for larger vehicles. Levinson (2019)74 as well as Davis and
Knittel (2019),75 on the other hand, criticize the use of footprint-based standards for not
providing incentives for people to buy smaller vehicles. These papers argue that the standards are
more harmful to lower-income households than a gasoline tax would be, in part because a
gasoline tax is more economically efficient, in part because higher-income households can better
afford the up-front cost increase, and in part because the revenues from a gasoline tax can be
redistributed in ways to reduce (or eliminate) the regressivity of the tax. Neither of these papers
addresses the reduction in fuel costs associated with the standards, though, and thus they omit a
significant effect of the standards.
The focus on vehicle ownership does not account for how the standards' effects on reduction
in per-mile costs might affect access to transportation for lower-income households. If these
reductions in operating costs are passed along to users, ride-hailing and ride-sharing services
might become less expensive and thus more accessible than before. Vehicles used in these
services are likely to have higher mileage more quickly than personally owned vehicles; as a
8-24
-------
result, up-front costs are likely to be recovered more quickly than the costs for personally owned
vehicles.
Greene and Welch (2018)76 include both fuel consumption and up-front costs in the
calculation of distributional effects. They find that higher-income households experience
decreases in fuel consumption due to the standards before lower-income households, because the
latter are more likely to own used vehicles and thus get the fuel savings with a lag. However,
they estimate that the ratio of fuel savings to costs, as well as the ratio of net savings to income,
is higher for lower-income households than higher-income households.
Vaidyanathan et al. (2021) observe that gasoline burden — the share of gasoline in income —
is more than three times higher for lower-income households than higher-income households.77
Thus, reducing per-mile costs may disproportionately benefit lower-income households, both
through more efficient vehicles gradually entering the used vehicle fleet, as well as through
reduced operating costs for other providers of transportation services. In addition, lower-income
households spent more per year on used vehicles than new ones. As discussed in Preamble
Section VII. J., because used vehicles have lower up-front costs (due to vehicle depreciation) but
still achieve the same reduced fuel consumption of the vehicle when it was new, buyers of used
vehicles will recover up-front costs much more quickly than new vehicle buyers. Expenditures
on fuel also fluctuate more than expenditures on vehicles, suggesting more uncertainty for fuel
costs.
Battery electric vehicles at this time have even higher new-vehicle costs and even lower
operating and maintenance costs than ICEVs. The advent of increased market penetration of
BEVs on lower-income households depends heavily both on how the new vehicle market
responds to those two factors, and on the availability of charging infrastructure for those
households. If EVs prove popular with new vehicle buyers, then the used vehicle market for EVs
will have increased availability; if EVs are slow to enter the new vehicle market, then the used
vehicle market will remain primarily ICEVs. In addition, the cost per mile of ride-sharing and
ride-hailing services are likely to depend on the penetration of EVs into those fleets. With their
higher use than personally owned vehicles, fleet vehicles may get up-front costs paid back more
quickly via reduced operating costs and may be expected to pass some of the reduced operating
costs to customers. Depending on the availability and cost of these services, lower-income
households without vehicles may have increased access to transportation. Challenges arise with
the availability of charging infrastructure for lower-income households: home charging, for
instance, may not be feasible if multi-unit dwellings do not offer charging or do not offer
sufficient charging access, or if people rely on-street parking. The availability of local, publicly
available charging infrastructure may thus influence the decision on whether to purchase an EV.
As a result, the penetration of EVs into lower-income neighborhoods may depend on public and
private decisions over where to place charging infrastructure.
In sum, the effects of the standards on low-income households are likely to be felt primarily
through effects on operating costs, and the effects of the standards on the used vehicle market.
While the standards are projected to reduce per-mile operating costs, and thus potentially
increase access to mobility, increases in new vehicle costs are likely to affect the used vehicle
market as well. This is discussed in Chapter 8.4.2.
8-25
-------
8.4.2 Effects on the Used Vehicle Market
Most people in the market for a vehicle purchase used vehicles: according to the U.S. Bureau
of Transportation Statistics, roughly two-thirds of vehicles sold in recent years have been used
rather than new.78 Further, according to Consumer Federation of America, in 2015 about 92
percent of vehicle purchases by low-income households were used vehicles.79 Thus, the effects
of the standards on many households depend on its impacts, not only in the new vehicle market,
but also in the used vehicle market.
Vehicles are long-lived, durable goods. The vehicles purchased this year are likely to last for
several decades, and their characteristics, including their fuel consumption, will mostly persist
for that time. As discussed in Preamble Section VII. J., recovery of the increase in up-front costs
of a used vehicle by reduced fuel expenditures happens much more rapidly than for a new
vehicle, because the up-front costs are depreciated.
The effect of the standards on the used vehicle market will be related to the effects of the
standards on new vehicle prices, the fuel efficiency of new vehicle models, the fuel efficiency of
used vehicles, and the total sales of new vehicles. On one hand, if the consumer value of fuel
savings resulting from improved fuel efficiency outweighs the average increase in new models'
prices to potential buyers of new vehicles, sales of new vehicles could rise, and the used vehicle
market may increase in volume as new vehicle buyers sell their older vehicles. If this is the case,
lower-income households are likely to benefit from the increased availability of used vehicles.
On the other hand, if potential buyers value future fuel savings resulting from the increased fuel
efficiency of new models at less than the increase in their average selling price, sales of new
vehicles could decline, and the used vehicle market may decrease in volume as people hold onto
their vehicles longer. In this case, lower-income households could face increased costs due to
reduced availability of used vehicles.
As discussed in Chapter 8.1.2, EPA contracted with RTI International to understand better the
connections between the new and the used vehicle market.80 Changes in the new vehicle market
are expected not only to have immediate effects on the prices of used vehicles, but also to affect
the market over time, as the supply of used vehicles in the future depends on how many new
vehicles are sold. As described in that subchapter, considering the interaction with the used
vehicle market leads to a less elastic demand elasticity for new vehicles. This result arises
because, as the price of used vehicles increases, new vehicle buyers are likely to notice that they
may recoup some of the new-vehicle price increase when they sell the vehicle as used. Thus,
buyers are less discouraged from buying new vehicles than if they did not recognize effects in
the used vehicle market. Using a range of parameters from published literature for sensitivity
analyses, this report estimates in Tables 8-2 and 8-3 that a 1 percent increase in new vehicle net
price would increase the price of a 5-year-old vehicle by between 1 and 1.4 percent, and would
reduce overall vehicle stock by 0.04 to 0.11 percent. That is, changes in new vehicle prices are
not expected to have a very large effect on the total stock of vehicles.
If the only effect of the standards on total cost of ownership were the up-front costs, then the
standards might also encourage people to hold onto their used vehicles longer. This effect,
sometimes termed the "Gruenspecht effect" after Gruenspecht (1982),81 would lead to both
slower adoption of vehicles subject to the new standards, and more use of older vehicles not
subject to the new standards, with associated higher emissions. Two older studies examine the
effects of new vehicle prices on scrappage: Miaou (1995)82 estimates an elasticity of -0.2 (that is,
8-26
-------
a 1 percent increase in new vehicle price leads to a 0.2 percent decrease in scrappage), while
Greenspan and Cohen (1999)83 estimate an elasticity of -0.8. Two newer studies estimate the
effects of changes in used vehicle prices on scrappage; Jacobsen and van Benthem (2015)84
estimate an elasticity of -0.7, and Bento et al. (2018)85 estimate an elasticity of -0.4. These
estimates suggest that scrappage rates are likely to change a relatively small amount in response
to changes in the new vehicle market.
The NAS 2021 Report86 (p. 11-357) notes the possibility of the Gruenspecht effect on
scrappage. In addition, it notes that, if people find the reduced fuel consumption of new vehicles
attractive, new vehicle sales would increase, and reduced scrappage would not be expected.
8.4.3 Effects on Access to Credit
Another question is whether higher vehicle prices may exclude some prospective consumers
from the new vehicle market through effects on consumers' ability to finance vehicles. It is
possible that lenders focus solely on the amount of the vehicle loan, the person's current debt,
and the person's income when issuing loans, and not the costs associated with fuel consumption.
If lenders restrict consideration to the amount of the loan, the borrower's debt, and the borrower's
income, then increased up-front costs of new vehicles subject to the standards will reduce
buyers' ability to get loans. However, if fuel savings are factored into lenders' decisions, reduced
fuel costs increase a borrower's capacity to repay a loan and therefore increase the likelihood of
getting a loan. Ignoring fuel savings could prevent a buyer's ability to get a loan, even if fuel
savings exceed the increase in loan payments due to higher purchase price. Thus, if lenders do
not take fuel savings into account in providing loans, households that are borrowing near the
limit of their abilities to borrow will either have to buy a different vehicle than intended, or not
buy a vehicle at all.
On the other hand, some lenders give discounts for loans to purchase more fuel-efficient
vehicles.87 The National Automobile Dealers Association in its comments provided results of
two surveys of financial institutions, which were asked whether they would increase credit for a
more expensive vehicle with lower cost of ownership. With about half of those surveyed
responding, over 80 percent of respondents replied that they would not; the remainder said they
would. These findings do not contradict EPA's finding that some lenders are willing to give
discounts on loans to purchase more fuel-efficient vehicles. In addition, subsidies exist from the
federal government, and some state governments, for plug-in electric vehicles.88 When
automakers comply with the standards through production of plug-in vehicles, these subsidies
reduce the costs of these vehicles and facilitate their purchases. Concerns have been raised that
these subsides go primarily to wealthier households, who are more likely to purchase new
vehicles in general and may be an expensive way to promote adoption of these vehicles. Sheldon
and Dua (2019)89 find that subsidies targeted by income and other factors may be more cost-
effective and progressive financially than untargeted policies.
Hutchens et al. (2021) examined the question whether the higher up-front cost might create an
obstacle if borrowers face a ceiling on the debt-to-income ratio (DTI), which affects a borrower's
access to credit. Evidence previously suggested that lenders may not give loans to consumers
who have a DTI above 36 percent; more recent evidence suggests that lenders consider 43
percent the maximum.90 Hutchens et al., using data from the U.S. Bureau of Labor Statistics'
Consumer Expenditure Survey, found that, from 2007 to 2019, 40 percent of lower-income
households and 8 percent of higher-income households both had a DTI of over 36 percent and
8-27
-------
purchased at least one new vehicle. The 36 percent threshold was maintained for continuity with
previous research. In 2019, 59 percent of lower-income households that purchased either a new
or used vehicle with a DTI of over 36 percent financed their car purchases. Thus, a DTI above 36
percent may not always be a disqualifying threshold in financing a new vehicle
It is worth mentioning that in addition to the factors discussed here, there are other factors that
may influence access to credit, such as race, ethnicity, gender, gender identity, residential
location, religion, or other factors. It is unclear whether or to what extent these possible
limitations on access to credit may affect access to auto loans.
Although access to credit is a potential barrier to purchase of vehicles whose up-front costs
increase, it may be a less impenetrable barrier when those up-front costs come with reduced fuel
consumption.
8.4.4 Effects on Low-Priced Cars
Average transaction price for a new vehicle in February 2021 was $41,000, an increase of 6.5
percent from February 2020. That increase, though, masks great diversity in vehicle prices; for
instance, the average transaction price for subcompacts at the same time was $18,300.91 For that
reason, low-priced vehicles may be considered an entry point for people into buying new
vehicles instead of used ones; automakers may seek to entice people to buy new vehicles through
a low price point. It is possible that higher costs associated with the standards could affect the
ability of automakers to maintain vehicles in this segment.
In the past, when CAFE standards did not vary by footprint, not only was the low-priced
vehicle segment a way to encourage first-time new vehicle purchasers, but it also tended to
include more fuel-efficient vehicles that assisted automakers in achieving CAFE standards.92 The
footprint-based standards, by encouraging improvements in GHG emissions and fuel economy
across the vehicle fleet, reduce the need for smaller, and by extension, low-priced vehicles to be
a primary means of compliance with the standards. This change in incentives for the marketing
of this segment may contribute to the increases in the prices of vehicles previously in this
category. Hutchens et al. (2021) found that, from 2005-2019, the number of vehicles priced
below $20,000 (2019$) has varied from 48 to 66, with 49 such vehicles available in 2019. Both
MotorTrend93 and Car and Driver94 provide a list of the ten least expensive new vehicles for MY
2021. The lowest priced, the Chevrolet Spark, is listed at under $15,000. Car and Driver's list has
prices all below $20,000; MotorTrend includes two of the ten with prices between $20,000 and
$21,000. In addition, these vehicles appear to be gaining more content, such as improved
entertainment systems and electric windows; they may be developing an identity as a desirable
market segment without regard to their historic role in enabling the sales of less efficient vehicles
and compliance with CAFE standards.95 Both MotorTrend and Car and Driver note that these
vehicles come with the latest safety, comfort, and entertainment features. It may be that the
small, fuel-efficient vehicles previously sold with low prices are evolving to fit consumer
demand that prefers content to low prices.
In sum, the low-priced vehicle segment still exists. Whether it continues to exist, and in what
form, may depend on the marketing plans of manufacturers: whether benefits are greater from
offering basic new vehicles to first-time new-vehicle buyers, or from making small vehicles
more attractive by adding more desirable features to them.
8-28
-------
8.4.5 Effects of Electric Vehicles on Affordability
Electric vehicles create some novel questions for affordability. Their up-front costs tend to be
higher than those of comparable gasoline vehicles, and their operating costs tend to be much
lower. Qualitatively, these characteristics are similar to gasoline vehicles subject to the
standards, as discussed above. In addition, though, electric vehicles require access to charging.
Home charging can be very convenient but requires the ability to park where charging is
available; a number of people, such as those who rely on on-street parking, may not have such
access. On the other hand, a recent report from the National Renewable Energy Laboratory
estimated recently that the number of public and workplace charging stations is keeping up with
projected needs, based on Level 2 and fast charging stations per plug-in EV.96 As discussed in
Chapter 4.1.3 under this rule the penetration of plug-in electric vehicles is projected to increase,
from 7 percent in MY 2023 to 17 percent in MY 2026. EPA plans to continue to study and
monitor these concerns as the prevalence of electric vehicles increases.
8.4.6 Summary of Affordability and Equity Effects
As with the effects of the standards on vehicle sales discussed in Chapter 8.1, the effects of
the standards on affordability depend on two countervailing effects: the increase in the up-front
costs of the vehicles, and the decrease in operating costs. The increase in up-front costs has the
potential to increase the prices of used vehicles, to make credit more difficult to obtain, and to
make the least expensive new vehicles less desirable compared to used vehicles. The reduction in
operating costs has the potential to mitigate or reverse all these effects. Lower operating costs on
their own increase mobility (see Chapter 3 for a discussion of rebound driving).
The effects of the standards on lower-income households are of great importance for social
equity and reflect these contrasting forces. The overall effects of the standards on vehicle
ownership, including for lower-income communities, depend heavily, as discussed in Chapter
8.1, on the role of fuel consumption in vehicle sales decisions. At the same time, lower-income
households own fewer vehicles per household and are more likely to buy used vehicles than new
compared to higher-income households, and they spend a higher proportion of their income on
fuel than do higher-income households. As a result, lower-income households may benefit more
from the reduction in operating costs than the increase in up-front costs of either new or used
vehicles. Finally, we note that effects on social equity involve impacts beyond those on lower-
income households. EPA will continue to examine these impacts.
8-29
-------
References for Chapter 8
1	U.S. Environmental Protection Agency (2021). 2020 EPA Automotive Trends Report: Greenhouse Gas Emissions,
Fuel Economy, and Technology since 1975, Chapter 4. EPA-420-R-21-003, https://www.epa.gov/automotive-
trends/download-automotive-trends-report#Full%20Report, accessed 4/15/2021.
2	Jaffe, A.B., and Stavins, R.N. (1994). "The Energy Paradox and the Diffusion of Conservation Technology."
Resource and Energy Economics 16(2): 91-122.
375 FR 25510-25513; 77 FR 62913-62917; U.S. Environmental Protection Agency (2016), Proposed Determination
on the Appropriateness of the Model Year 2022-2025 Light-Duty Vehicle Greenhouse Gas Emissions Standards
under the Midterm Evaluation, EPA-420-R-16-020, Appendix B.1.2; 85 FR 24603-24613.
4	https://www.epa.gov/regulations-emissions-vehicles-and-engines/midterm-evaluation-light-duty-vehicle-
greenhouse-gas
5	National Research Council (2015). Cost, Effectiveness, and Deployment of Fuel Economy Technologies for Light-
Duty Vehicles. Washington, DC: The National Academies Press, https://doi.org/10.17226/21744, p. 1-7.
6	National Academies of Sciences, Engineering, and Medicine (2021). Assessment of Technologies for Improving
Light-Duty Vehicle Fuel Economy 2025-2035. Washington, DC: The National Academies Press.
https://doi.org/10.17226/26092.
7	Helfand, G., et al. (2016). "Searching for Hidden Costs: A Technology-Based Approach to the Energy Efficiency
Gap in Light-Duty Vehicles." Energy Policy 98: 590-606; Huang, H., et al. (2018). "Re-Searching for Hidden Costs:
Evidence from the Adoption of Fuel-Saving Technologies in Light-Duty Vehicles." Transportation Research Part D
65: 194-212.
8	Huang, H., G. Helfand, and K. Bolon (2018a). "Consumer Satisfaction with New Vehicles Subject to Greenhouse
Gas and Fuel Economy Standards." Presentation at the Society for Benefit-Cost Analysis annual conference, March.
https://benefitcostanalysis.Org/docs/G.4_Huang_Slides.pdf, accessed 4/7/2021.
9	Knittel, C. R. (2011). "Automobiles on Steroids: Product Attribute Trade-Offs and Technological Progress in the
Automobile Sector." American Economic Review 101(7): pp. 3368-3399; Klier, T. and Linn, J. (2016). "The Effect
of Vehicle Fuel Economy Standards on Technology Adoption." Journal of Public Economics 133: 41-63;
McKenzie, D. andHeywood, J. B. (2015). "Quantifying efficiency technology improvements in U.S. cars from
1975-2009." Applied Energy 157: 918-928.
10	Moskalik, A., K. Bolon, K. Newman, and J. Cherry (2018). "Representing GHG Reduction Technologies in the
Future Fleet with Full Vehicle Simulation." SAE Technical Paper 2018-01-1273. doi: 10.4271/2018-01-1273.
11	Watten, A., S. Anderson, and G. Helfand (2021). "Attribute Production and Technical Change: Rethinking the
Performance and Fuel Economy Trade-off for Light-duty Vehicles." Working paper.
12	Helfand, G., and R. Dorsey-Palmateer (2015). "The Energy Efficiency Gap in EPA's Benefit-Cost Analysis of
Vehicle Greenhouse Gas Regulations: A Case Study," Journal of Benefit-Cost Analysis 6(2): 432-454.
13	Whitefoot, K., M. Fowlie, and S. Skerlos (2017). "Compliance by Design: Influence of Acceleration Trade-offs
on CO2 Emissions and Costs of Fuel Economy and Greenhouse Gas Regulations." Environmental Science and
Technology 51: 10307-10315.
14	See Chapter 8 Endnote 3.
15	U.S. Environmental Protection Agency (2010). "How Consumers Value Fuel Economy: A Literature Review."
EPA-420-R-10-008, https://clpub.epa.gov/si/si_public_file_download.cfm?p_download_id=499454&Lab=OTAQ
(accessed 4/15/2021); U.S. Environmental Protection Agency (2018). "Consumer Willingness to Pay for Vehicle
Attributes: What is the Current State of Knowledge?" EPA-420-R-18-016,
https ://cfpub .epa.gov/si/si_public_file_download. cfm?p_download_id=536423&Lab=OTAQ (accessed 4/15/2021);
Greene, D., A. Hossain, J. Hofmann, G. Helfand, and R. Beach (2018). "Consumer Willingness to Pay for Vehicle
Attributes: What Do We Know?" Transportation Research Part A 118: 258-279.
16	E.g., Allcott, H. (2013). "The Welfare Effects of Misperceived Product Costs: Data and Calibrations from the
Automobile Market." American Economic Journal: Economic Policy 5: 30-66; Busse, M., C. Knittel, and F.
8-30
-------
Zettlemeyer (2013). "Are Consumers Myopic? Evidence from New and Used Car Purchases." American Economic
Review 103: 220-256; Sallee, J., S. West, and W. Fan (2016). "Do consumers recognize the value of fuel economy?
Evidence from used car prices and gasoline price fluctuations." Journal of Public Economics 135: 61-73.
17	E.g., Gillingham, K., S. House, and A. van Benthem (2019). "Consumer Myopia in Vehicle Purchases: Evidence
from a Natural Experiment." NBER Working Paper 25845, http://www.nber.org/papers/w25845 (accessed
7/28/2021).
18	Sallee, J. (2014). "Rational Inattention and Energy Efficiency." Journal of Law and Economics 57: 781-820.
19	Fischer, C. (2005). "On the Importance of the Supply Side in Demand-Side Management." Energy Economics 27:
165-180; Blumstein, C., and M. Taylor (2013). "Rethinking the Energy-Efficiency Gap: Producers, Intermediaries,
and Innovation." Energy Institute at Haas Working Paper WP 243; Houde, S., and C. Spurlock (2015). "Minimum
Energy Efficiency Standards for Appliances: Old and New Economic Rationales." Economics of Energy &
Environmental Policy 5: 65-83.
20	Fischer, Carolyn (2005). "On the Importance of the Supply Side in Demand-Side Management." Energy
Economics 27:165-180.
21	Abernathy, W. J., and Utterback, J. M. (1978). Patterns of industrial innovation. Technology review, 80, pp. 254-
228; Henderson, R. M., and Clark, K. B. (1990). "Architectural innovation: The reconfiguration of existing product
technologies and the failure of established firms." Administrative science quarterly, 9-30.
99
Blumstein, Carl and Margaret Taylor (2013). "Rethinking the Energy-Efficiency Gap: Producers, Intermediaries,
and Innovation," Energy Institute at Haas Working Paper 243, University of California at Berkeley; Tirole, Jean
(1998). The Theory of Industrial Organization. Cambridge, MA: MIT Press, pp. 400, 402.
90
Popp, D., Newell, R.G., and Jaffe, A.B. (2010). "Energy, the environment and technological change." In
Handbook of the Economics of Innovation 2nd ed. B.H. Hall, and N. Rosenberg, Elsevier; Vollebergh, Herman R.J.,
and Edwin van der Werf (2014). "The Role of Standards in Eco-Innovation: Lessons for Policymakers." Review of
Environmental Economics and Policy 8(2): 230-248.
24	Powell, Walter W., and Eric Giannella (2010). "Collective Invention and Inventor Networks," Chapter 13 in
Handbook of the Economics of Innovation, Volume 1, ed. B. Hall and N. Rosenberg, Elsevier.
25	See Chapter 8 Endnote 6, p. 9-10.
26	Howard, P., and M. Sarinsky (2020). "Turbocharged: How One Revision in the SAFE Rule Economic Analysis
Obscures Billions of Dollars in Social Harms." New York University Institute for Policy Integrity.
https://policyintegrity.org/files/publications/Turbocharged_How_One_Revision_in_the_SAFE_Rule_Economic_An
alysis_Obscures.pdf (accessed 10/07/2021).
27	U.S. Environmental Protection Agency (2021). "The Effects of New-Vehicle Price Changes on New- and Used-
Vehicle Markets and Scrappage." EPA-420-R-21-019,
https://cfpub.epa.gov/si/si_public_record_Report.cfm?dirEntryId=352754&Lab=OTAQ (accessed 10/06/2021).
90
Howard, P., and M. Sarinsky (2020). "Turbocharged: How One Revision in the SAFE Rule Economic Analysis
Obscures Billions of Dollars in Social Harms." New York University Institute for Policy Integrity.
https://policyintegrity.org/files/publications/Turbocharged_How_One_Revision_in_the_SAFE_Rule_Economic_An
alysis_Obscures.pdf (accessed 10/07/2021).
29
Comments of The Center for Biological Diversity, Chesapeake Bay Foundation, Conservation Law Foundation,
Earthjustice, Environmental Law & Policy Center, Natural Resources Defense Council, Public Citizen, Inc., Sierra
Club, and Union of Concerned Scientists Regarding Proposed Rule Regarding Revised 2023 and Later Model Year
Light-Duty Vehicle Greenhouse Gas Emission Standards. EPA-HQ-OAR-2021-0208-0651, Attachment 1, pp. 51-
54.
Morgenstern, R.D.; Pizer, W.A.; and Shih, J.-S. "Jobs Versus the Environment: An Industry-Level Perspective."
Journal of Environmental Economics and Management 43: 412-436. 2002.
o 1
Berman, E. and Bui, L. T. M. (2001). "Environmental Regulation and Labor Demand: Evidence from the South
Coast Air Basin." Journal of Public Economics 79(2): 265-295.
8-31
-------
32	Deschenes, O. (2018). "Balancing the Benefits of Environmental Regulations for Everyone and the Costs to
Workers and Firms." IZA World of Labor 22v2. https://wol.iza.org/uploads/articles/458/pdfs/environmental-
regulations-and-labor-markets.pdf, accessed 4/19/2021.
33	Ehrenberg, R. G., and Smith, R.S. (2000). Modern Labor Economics: Theory and Public Policy, Seventh Edition.
Reading, MA: Addison Wesley Longman, Inc., p. 108.
34	Greenstone, M. (2002). "The Impacts of Environmental Regulations on Industrial Activity: Evidence from the
1970 and 1977 Clean Air Act Amendments and the Census of Manufactures." Journal of Political Economy 110(6):
1175-1219; Ferris, A.; Shadbegian, R.J.; and Wolverton, A. (2014). "The Effect of Environmental Regulation on
Power Sector Employment: Phase I of the Title IV S02 Trading Program." Journal of the Association of
Environmental and Resource Economists 1(4): 521-553; Walker, R.W. (2013). "The Transitional Costs of Sectoral
Reallocation: Evidence From the Clean Air Act and the Workforce." The Quarterly Journal of Economics. 1787-
1835; Curtis, M.E. (2018). "Who Loses Under Cap-and-Trade Programs? The Labor Market Effects of the NOx
Budget Trading Program." The Review of Economics and Statistics 100(1): 151-166; Curtis, M.E. (2020).
"Reevaluating the ozone nonattainment standards: Evidence from the 2004 expansion." Journal of Environmental
Economics and Management 99: 102261
35	Hafstead ,M.A.C. and Williams III, R.C. (2018). "Unemployment and Environmental Regulation in General
Equilibrium." Journal of Public Economics 160: 50-65.
36	Arrow, K., et al. (1996). "Benefit-Cost Analysis in Environmental, Health, and Safety Regulation: A Statement of
Principles." American Enterprise Institute, The Annapolis Center, and Resources for the Future, p. 6.
37	Schmalensee, Richard, and Robert N. Stavins. "A Guide to Economic and Policy Analysis of EPA's Transport
Rule." White paper commissioned by Excelon Corporation, March 2011.
Krisher, T., and Seewer, J. (2021). "Autoworkers face uncertain future in an era of electric cars."
https://abcnews.go.com/US/wireStory/autoworkers-face-dimmer-future-era-electric-cars-75828610, accessed
10/20/2021.
39	Kupper, D., K. Kuhlmann, K. Tominaga, A. Arora, and J. Schlageter (2020). "Shifting Gears in Auto
Manufacturing." https://www.bcg.com/publications/2020/transformative-impact-of-electric-vehicles-on-auto-
manufacturing , accessed 10/20/2021.
40	"Technical Guidance for Assessing Environmental Justice in Regulatory Analysis." Epa.gov, Environmental
Protection Agency, https://www.epa.gOv/sites/production/files/2016-06/documents/ejtg_5_6_16_v5.l.pdf.
41	USGCRP, 2018: Impacts, Risks, and Adaptation in the United States: Fourth National Climate Assessment,
Volume II [Reidmiller, D.R., C.W. Avery, D.R. Easterling, K.E. Kunkel, K.L.M. Lewis, T.K. Maycock, and B.C.
Stewart (eds.)]. U.S. Global Change Research Program, Washington, DC, USA, 1515 pp. doi: 10.7930/NCA4.2018.
42
USGCRP, 2016: The Impacts of Climate Change on Human Health in the United States: A Scientific Assessment.
Crimmins, A., J. Balbus, J.L. Gamble, C.B. Beard, J.E. Bell, D. Dodgen, R.J. Eisen, N. Fann, M.D. Hawkins, S.C.
Herring, L. Jantarasami, D.M. Mills, S. Saha, M.C. Sarofim, J. Trtanj, and L. Ziska, Eds. U.S. Global Change
Research Program, Washington, DC, 312 pp. http://dx.doi.org/10.7930/J0R49NQX
43	Oppenheimer, M., M. Campos, R.Warren, J. Birkmann, G. Luber, B. O'Neill, andK. Takahashi, 2014: Emergent
risks and key vulnerabilities. In: Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and
Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel
on Climate Change [Field, C.B., V.R. Barros, D.J. Dokken, K.J. Mach, M.D. Mastrandrea, T.E. Bilir, M. Chatteijee,
K.L. Ebi, Y.O. Estrada, R.C. Genova, B. Girma, E.S. Kissel, A.N. Levy, S. MacCracken, P.R. Mastrandrea, and
L.L.White (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, pp. 1039-
1099.
44	Porter, J.R., L. Xie, A.J. Challinor, K. Cochrane, S.M. Howden, M.M. Iqbal, D.B. Lobell, and M.I. Travasso,
2014: Food security and food production systems. In: Climate Change 2014: Impacts, Adaptation, and Vulnerability.
Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change [Field, C.B., V.R. Barros, D.J. Dokken, K.J. Mach, M.D. Mastrandrea,
T.E. Bilir, M. Chatteijee, K.L. Ebi, Y.O. Estrada, R.C. Genova, B. Girma, E.S. Kissel, A.N. Levy, S. MacCracken,
8-32
-------
P.R. Mastrandrea, and L.L.White (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York,
NY, USA, pp. 485-533.
45	Smith, K.R., A.Woodward, D. Campbell-Lendrum, D.D. Chadee, Y. Honda, Q. Liu, J.M. Olwoch, B. Revich, and
R. Sauerborn, 2014: Human health: impacts, adaptation, and co-benefits. In: Climate Change 2014: Impacts,
Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth
Assessment Report of the Intergovernmental Panel on Climate Change [Field, C.B., V.R. Barros, D.J. Dokken, K.J.
Mach, M.D. Mastrandrea, T.E. Bilir, M. Chatteijee, K.L. Ebi, Y.O. Estrada, R.C. Genova, B. Girma, E.S.
Kissel,A.N. Levy, S. MacCracken, P.R. Mastrandrea, and L.L.White (eds.)]. Cambridge University Press,
Cambridge, United Kingdom and New York, NY, USA, pp. 709-754.
46	IPCC, 2018: Global Warming of 1.5°C.An IPCC Special Report on the impacts of global warming of 1.5°C
above pre-industrial levels and related global greenhouse gas emission pathways, in the context of strengthening the
global response to the threat of climate change, sustainable development, and efforts to eradicate poverty [Masson-
Delmotte, V., P. Zhai, H.-O. Portner, D. Roberts, J. Skea, P.R. Shukla, A. Pirani, W. Moufouma-Okia, C. Pean, R.
Pidcock, S. Connors, J.B.R. Matthews, Y. Chen, X. Zhou, M.I. Gomis, E. Lonnoy, T. Maycock, M. Tignor, and T.
Waterfield (eds.)]. In Press.
47	National Research Council. 2011. America's Climate Choices. Washington, DC: The National Academies Press.
https://doi.org/10.17226/12781.
48	National Academies of Sciences, Engineering, and Medicine. 2017. Communities in Action: Pathways to Health
Equity. Washington, DC: The National Academies Press, https://doi.org/10.17226/24624.
49	USGCRP, 2016: The Impacts of Climate Change on Human Health in the United States: A Scientific Assessment
50	Ebi, K.L., J.M. Balbus, G. Luber, A. Bole, A. Crimmins, G. Glass, S. Saha, M.M. Shimamoto, J. Trtanj, and J.L.
White-Newsome, 2018: Human Health. In Impacts, Risks, and Adaptation in the United States: Fourth National
Climate Assessment, Volume II [Reidmiller, D.R., C.W. Avery, D.R. Easterling, K.E. Kunkel, K.L.M. Lewis, T.K.
Maycock, and B.C. Stewart (eds.)]. U.S. Global Change Research Program, Washington, DC, USA, pp. 539-571.
doi: 10.7930/NCA4.2018.CH14
51	Porter et al., 2014: Food security and food production systems.
52	Jantarasami, L.C., R. Novak, R. Delgado, E. Marino, S. McNeeley, C. Narducci, J. Raymond-Yakoubian, L.
Singletary, and K. Powys Whyte, 2018: Tribes and Indigenous Peoples. In Impacts, Risks, and Adaptation in the
United States: Fourth National Climate Assessment, Volume II [Reidmiller, D.R., C.W. Avery, D.R. Easterling,
K.E. Kunkel, K.L.M. Lewis, T.K. Maycock, and B.C. Stewart (eds.)]. U.S. Global Change Research Program,
Washington, DC, USA, pp. 572-603. doi: 10.7930/NCA4.2018.CH15
53	Porter et al., 2014: Food security and food production systems.
54	Mohai, P.; Pellow, D.; Roberts Timmons, J. (2009) Environmental justice. Annual Reviews 34: 405-430.
https://doi.org/10.1146/annurev-environ-082508-094348
55	Rowangould, G.M. (2013) A census of the near-roadway population: public health and environmental justice
considerations. Trans Res D 25: 59-67. http://dx.doi.Org/10.1016/j.trd.2013.08.003
56	Marshall, J.D., Swor, K.R.; Nguyen, N.P (2014) Prioritizing environmental justice and equality: diesel emissions
in Southern California. Environ Sci Technol 48: 4063-4068. https://doi.org/10.1021/es405167f
57	Marshall, J.D. (2000) Environmental inequality: air pollution exposures in California's South Coast Air Basin.
Atmos Environ21: 5499-5503. https://doi.Org/10.1016/j.atmosenv.2008.02.005
CO
Tian, N.; Xue, J.; Barzyk. T.M. (2013) Evaluating socioeconomic and racial differences in traffic-related metrics
in the United States using a GIS approach. J Exposure Sci Environ Epidemiol 23: 215-222.
59	Tian, N.; Xue, J.; Barzyk. T.M. (2013) Evaluating socioeconomic and racial differences in traffic-related metrics
in the United States using a GIS approach. J Exposure Sci Environ Epidemiol 23: 215-222.
60	Boehmer, T.K.; Foster, S.L.; Henry, J.R.; Woghiren-Akinnifesi, E.L.; Yip, F.Y. (2013) Residential proximity to
major highways - United States, 2010. Morbidity and Mortality Weekly Report 62(3): 46-50.
8-33
-------
61 Tian, N.; Xue, J.; Barzyk. T.M. (2013) Evaluating socioeconomic and racial differences in traffic-related metrics
in the United States using a GIS approach. JExposure Sci Environ Epidemiol 23: 215-222.
62Boehmer, T.K.; Foster, S.L.; Henry, J.R.; Woghiren-Akinnifesi, E.L.; Yip, F.Y. (2013) Residential proximity to
major highways - United States, 2010. Morbidity and Mortality Weekly Report 62(3): 46-50.
63	Gwo, Y., Z. Chen, A. Stuart, X. Li, and Y. Zhang (2020). "A systematic overview of transportation equity in
terms of accessibility, traffic emissions, and safety outcomes: From conventional to emerging technologies."
Transportation Research Interdisciplinary Perspectives 4: 100091.
64	Hutchens, A., A. Cassidy, G. Burmeister, and G. Helfand. "Impacts of Light-Duty Vehicle Greenhouse Gas
Emission Standards on Vehicle Affordability." Working paper.
65	U.S. Environmental Protection Agency (2016). Proposed Determination on the Appropriateness of the Model
Year 2022-2025 Light-Duty Vehicle Greenhouse Gas Emissions Standards under the Midterm Evaluation:
Technical Support Document. EPA-420-R-16-021. https://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=P100Q3L4.pdf,
accessed 4/26/2021.
66	Bradley, R. (2008). "Comment - Defining health insurance affordability: Unobserved heterogeneity matters."
Journal of Health Economics 27: 1129—1140.
67	Bartl, M. (2010). "The Affordability of Energy: How Much Protection for the Vulnerable Consumers?" Journal of
Consumer Policy 33: 225—245.
68	Thakuriah, P., and Liao, Y. (2006). "Transportation expenditures and ability to pay." Transportation Research
Record: Journal of the Transportation Research Board 1985: 257—265.
69	Schanzenbach, D.W., and McGranahan, L. (2012). "The impact of transportation on affordability: An analysis of
auto cost white paper." Manhattan Strategy Group, accessed 06/09/2014 at:
http://www.locationaffordability.info/downloads/AutoCostResearch.pdf.
70	U.S. Environmental Protection Agency (2016). Proposed Determination on the Appropriateness of the Model
Year 2022-2025 Light-Duty Vehicle Greenhouse Gas Emissions Standards under the Midterm Evaluation. EPA-
420-R-16-020. https://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=P100Q3DO.pdf, accessed 4/26/2021.
71	Cox Automotive (2021). "Cox Automotive Car Buyer Journey Study: Pandemic Edition Media Summary."
https://www.coxautoinc.com/wp-content/uploads/2021/02/Cox-Automotive-Car-Buyer-Journey-Study-Pandemic-
Edition-Summary.pdf, accessed 5/19/2021.
72	Semega, J., et al. (2020). "Income and Poverty in the United States: 2019." U.S. Census Bureau Report P60-207.
https://www.census.gov/library/publications/2020/demo/p60-270.html, accessed 5/19/2021.
73	Jacobsen, M. (2013). "Evaluating U.S. fuel economy standards in a model with producer and household
heterogeneity." American Economic Journal: Economic Policy 5: 148-87.
74	Levinson, A. (2019). "Energy Efficiency Standards Are More Regressive Than Energy Taxes: Theory and
Evidence." Journal of the Association of Environmental and Resource Economists 6: S7- S36.
75	Davis, L., and C. Knittel (2019). "Are Fuel Economy Standards Regressive?" Journal of the Association of
Environmental and Resource Economists 6: S37-S63.
76	Greene, D., and J. Welch (2018). "Impacts of fuel economy improvements on the distribution of income in the
U.S." Energy Policy 122: 528-541.
77	Vaidyanathan, S., P. Huether, and B. Jennings (2021). "Understanding Transportation Energy Burdens."
Washington, DC: American Council for an Energy-Efficient Economy White Paper, https://www.aceee.org/white-
paper/2021/05/understanding-transportation-energy-burdens , accessed 5/24/2021.
78	U.S. Department of Transportation, Bureau of Transportation Statistics. "New and Used Passenger Car and Light
Truck Sales and Leases." National Transportation Statistics Table 1-17. https://www.bts.gov/content/new-and-used-
passenger-car-sales-and-leases-thousands-vehicles, accessed 11/3/2021.
79	See Chapter 8 Endnote 71.
80	See Chapter 8 Endnote 26.
8-34
-------
O 1
Gruenspecht, H. (1982). "Differentiated Regulation: The Case of Auto Emissions Standards." American
Economic Review 72: 328-331.
oo
Miaou, S.-P. (1995). Factors Associated with Aggregate Car Scrappage Rate in the United States: 1966-1992.
Transportation Research Record 1475.
83 Greenspan, A., and D. Cohen (1999). "Motor Vehicle Stocks, Scrappage, and Sales." Review of Economics and
Statistics 81(3): 369-383.
OA
Jacobsen, M., and A. vanBenthem (2015). "Vehicle Scrappage and Gasoline Policy." American Economic
Review 105: 1312-1338.
85	Bento, A., K. Roth, and Y. Zuo (2018). "Vehicle Lifetime Trends and Scrappage Behavior in the U.S. Used Car
Market." The Energy Journal 39: 159-183.
86	See Chapter 8 Endnote 6.
on
Helfand, Gloria (2021). "Memorandum: Lending Institutions that Provide Discounts for more Fuel Efficient
Vehicles." U.S. EPA Office of Transportation and Air Quality, Memorandum to the Docket.
88	U.S. Department of Energy and U.S. Environmental Protection Agency. "Federal Tax Credits for New All-
Electric and Plug-in Hybrid Vehicles." https://www.fueleconomy.gov/feg/taxevb.shtml, accessed 4/28/2021.
89	Sheldon, T., and R. Dua (2019). "Measuring the cost-effectiveness of electric vehicle subsidies." Energy
Economics 84: 1-11.
90	Consumer Financial Protection Bureau (2019). "What is a debt-to-income ratio? Why is the 43% debt-to-income
ratio important?" https://www.consumerfinance.gov/ask-cfpb/what-is-a-debt-to-income-ratio-why-is-the-43-debt-to-
income-ratio-important-en-1791/, accessed 5/18/2021; Murphy, C. (2021). "Debt-to-income (DTI) Ratio."
Investopedia https://www.investopedia.eom/terms/d/dti.asp, accessed 5/18/2021.
91	Kelley Blue Book (2021). "Press Release: Average New-Vehicle Prices Jump More Than 6% Year-Over-Year;
Most Luxury Segments Decline, According to Kelley Blue Book." https://mediaroom.kbb.com/2021-03-l 1-
Average-New-Vehicle-Prices-Jump-More-Than-6-Year-Over-Year-Most-Luxury-Segments-Decline-According-to-
Kelley-Blue-Book, accessed 5/19/2021.
92	Austin, D., and T. Dinan (2005). "Clearing the Air: The Costs and Consequences of Higher CAFE Standards and
Increased Gasoline." Journal of Environmental Economics and Management 50(3): 562—82; Kleit, A. (2004).
"Impacts of Long-Range Increases in the Fuel Economy (CAFE) Standard." Economic Inquiry 42(2): 279-294.
93	Motortrend (2021). "These Are the 10 Cheapest Cars You Can Buy in 2021."
https://www.motortrend.com/features-collections/top-10-cheapest-new-cars/, accessed 4/28/2021.
94	Irwin, A (2020). "10 Cheapest New Cars for 2021." https://www.caranddriver.eom/features/g34908888/10-
cheapest-new-cars-for-2021/, accessed 4/28/2021.
95	See Chapter 8 Endnote 64.
96	Brown, A., A. Schayowitz, and E. Klotz (2021). "Electric Vehicle Infrastructure Trends from the Alternative
Fueling Station Locator: First Quarter 2021." National Renewable Energy Laboratory Technical Report NREL/TP-
5400-80684,
https://afdc.energy.gOv/files/u/publication/electric_vehicle_charging_infrastructure_trends_first_quarter_2021.pdf,
accessed 11/3/2021.
8-35
-------
Chapter 9: Small Business Flexibilities
The Regulatory Flexibility Act, as amended by the Small Business Regulatory Enforcement
Fairness Act of 1996 (SBREFA), generally requires an agency to prepare a regulatory flexibility
analysis of any rule subject to notice-and-comment rulemaking requirements under the
Administrative Procedure Act or any other statute. As a part of this analysis, an agency is
directed to convene a Small Business Advocacy Review Panel (SBAR Panel or 'the Panel'),
unless the agency certifies that the rule will not have a significant economic impact on a
substantial number of small entities. During such a Panel process, the agency would gather
information and recommendations from Small Entity Representatives (SERs) on how to reduce
the impact of the rule on small entities. As discussed below, EPA is certifying that this rule will
not have a significant economic impact on a substantial number of small entities, and thus we
have not conducted an SBAR Panel for this final rulemaking.
The following discussion provides an overview of small entities in the vehicle market. Small
entities include small businesses, small organizations, and small governmental jurisdictions. For
the purposes of assessing the impacts of the rule on small entities, a small entity is defined as: (1)
a small business that meets the definition for business based on the Small Business
Administration's (SBA) size standards (see Table 9-1); (2) a small governmental jurisdiction that
is a government of a city, county, town, school district or special district with a population of less
than 50,000; and (3) a small organization that is any not-for-profit enterprise which is
independently owned and operated and is not dominant in its field. This chapter provides an
overview of the primary SBA small business categories potentially affected by this regulation.
Table 9-1: Primary Vehicle SBA Small Business Categories
Industry a
Defined as Small Entity by
SBA if Less Than or Equal to:
NAICS Codes b
Vehicle manufacturers (including small
volume manufacturers)
1,500 employees
336111,336112
Independent commercial importers
$8 million annual sales
$27 million annual sales
250 employees
811111,811112,811198
441120
423110
Alternative Fuel Vehicle Converters
1,000 employees
1,250 employees
$8 million annual sales
336312,336322,336399
335312
811198
Notes:
a.	Light-duty vehicle entities that qualify as small businesses are not be subject to this rule. We are exempting
small business entities from the GHG standards.
b.	North American Industrial Classification System
We compiled a list of vehicle manufacturers, independent commercial importers (ICIs), and
alternative fuel converters that would be potentially affected by the rule from our 2019 and 2021
model year certification databases. These companies are already certifying their vehicles for
compliance with applicable EPA emissions standards (e.g., Tier 3). We then identified
companies that appear to meet the definition of small business provided in the table above. We
were able to identify companies based on certification information and previous rulemakings
where we conducted Regulatory Flexibility Analyses.
Based on this assessment, EPA identified a total of about 19 entities that appear to fit the
Small Business Administration (SBA) criterion of a small business. EPA estimates there are
9-1
-------
about 7 small vehicle manufacturers, 4 independent commercial importers (ICIs), and 8
alternative fuel vehicle converters in the light-duty vehicle market which may qualify as small
businesses (see Table 9-2 for a list of current entities). Independent commercial importers (ICIs)
are companies that hold a Certificate (or Certificates) of Conformity permitting them to import
nonconforming vehicles and to modify these vehicles to meet U.S. emission standards. ICIs are
not required to meet the emission standards in effect when the vehicle is modified, but instead
they must meet the emission standards in effect when the vehicle was originally produced (with
an annual production cap of a total of 50 light-duty vehicles and trucks). Alternative fuel vehicle
converters are businesses that convert gasoline or diesel vehicles to operate on alternative fuel
(e.g., compressed natural gas), and converters must seek a certificate for all of their vehicle
models. Model year 1993 and newer vehicles that are converted are required to meet the
standards applicable at the time the vehicle was originally certified. Converters serve a niche
market, and these businesses primarily convert vehicles to operate on compressed natural gas
(CNG) and liquefied petroleum gas (LPG), on a dedicated or dual fuel basis.
Table 9-2: Small Business Entities
Small Vehicle
Manufacturers
Alternative Fuel Converters
Independent Commercial Importers
Ineos Automotive
Karma Automotive
Koenigsegg
Pagani
RUF
Workhorse Group
Rimac
AGA Systems, LLC
Agility Powertrain Systems, LLC
Altech-Eco Corporation
Blossman Services, Inc.
Eco Vehicle Systems, LLC
Encore TEC LLC
Landi Renzo USA Corporation
Westport Dallas, Inc
DRPC, LLC
G&K Automotive Conversions, Inc
Wallace Environmental Testing Laboratories, Inc
JK Technologies, LLC
EPA is exempting from the GHG standards any manufacturer, domestic or foreign, meeting
SBA's size definitions of small business as described in 13 CFR 121.201. EPA adopted the same
type of exemption for small businesses in the MY 2012-2016 rulemaking.1 Together, we
estimate that small entities comprise less than 0.1 percent of total annual vehicle sales and
exempting them will have a negligible impact on the GHG emissions reductions from the
standards. In light of our exempting small businesses from the GHG standards, we are certifying
in the preamble to the final rule that the rule will not have a significant economic impact on a
substantial number of small entities. Therefore, EPA has not conducted a Regulatory Flexibility
Analysis or a SBREFA SBAR Panel for the rule.
EPA allows small businesses to voluntarily waive their small business exemption and
optionally certify to the GHG standards. This will allow small entity manufacturers to earn CO2
credits under the GHG program, if their actual fleetwide CO2 performance is better than their
fleetwide CO2 target standard. Manufacturers waiving their small business exemption are
required to meet all aspects of the GHG standards and program requirements across their entire
product line. However, the exemption waiver is optional for small entities and thus we believe
that manufacturers would only opt into the GHG program if it is economically advantageous for
them to do so, for example to generate and sell CO2 credits. Therefore, EPA believes having this
voluntary option does not affect EPA's determination that the standards will impose no
significant adverse impact on a substantial number of small entities.
9-2
-------
References for Chapter 9
1 75 FR 25424, May 7, 2010.
9-3
-------
Chapter 10: Summary of Costs and Benefits
This chapter presents a summary of costs, benefits, and net benefits of the final program and
the alternatives. This rule is not expected to have measurable inflationary or recessionary effects.
10.1 Final Rule
Table 10-1 shows the estimated annual monetized costs of the final program for the indicated
calendar years. The table also shows the present values (PV) of those costs for the calendar years
2021-2050 using both 3 percent and 7 percent discount rates.56 The table includes an estimate of
foregone consumer sales surplus, which measures the loss in benefits attributed to consumers
who would have purchased a new vehicle in the absence of the proposed standards.
Table 10-1: Costs Associated with the Final Program ($Billions of 2018 dollars)
Calendar
Year
Foregone Consumer
Sales Surplus Til
Technology
Costs
Congestion
Noise
Fatality
Costs
Non-fatal
Crash Costs
Total
Costs
2023
$0,029
$5.6
$0.03
$0,000
45
$0.13
$0.23
$6.1
2026
$0.11
$16
$0.12
$0,002
$0.42
$0.7
$17
2030
$0,093
$17
$0.4
$0,006
7
$0.44
$0.73
$19
2035
$0,078
$17
$0.68
$0,011
$0.27
$0.44
$19
2040
$0,063
$16
$0.84
$0,014
$0.15
$0.25
$17
2050
$0,052
$15
$0.9
$0,015
$0.16
$0.25
$16
PV, 3%
$1.3
$280
$9.6
$0.16
$4.9
$8.1
$300
PV, 7%
$0.84
$160
$4.8
$0.08
$3.2
$5.3
$180
Annualiz
ed, 3%
$0,069
$14
$0.49
$0,008
2
$0.25
$0.42
$15
Annualiz
ed, 7%
$0,068
$13
$0.39
$0,006
5
$0.26
$0.43
$14
[1] "Foregone Consumer Sales Surplus" refers to the difference between a vehicle's price and the buyer's
willingness to pay for the new vehicle; the impact reflects the reduction in new vehicle sales described in
Chapter8.1. See Section 8 of CAFE Model Documentation FR 2020.pdf in the docket for more information.
Table 10-2 shows the undiscounted annual monetized fuel savings of the program. The table
also shows the present value of those fuel savings for the same calendar years using both 3
percent and 7 percent discount rates. The aggregate value of fuel savings is calculated using pre-
tax fuel prices since savings in fuel taxes do not represent a reduction in the value of economic
resources utilized in producing and consuming fuel. Note that the fuel savings shown in Table
10-2 result from reductions in fleet-wide fuel use and include rebound effects, credit usage and
advanced technology multiplier use. Thus, fuel savings grow over time as an increasing fraction
of the fleet is projected to meet the final standards.
56 For the estimation of the stream of costs and benefits, we assume that after implementation of the proposed MY
2023-2026 standards, the 2026 standards apply to each year thereafter.
10-1
-------
Table 10-2: Fuel Savings Associated with the Final Program ($Billions of 2018 dollars)
Calendar Year
Retail Fuel Savings
Fuel Tax Savings
Total Fuel Savings
2023
$0.94
$0.31
$0.62
2026
$5.1
$1.7
$3.3
2030
$16
$4.5
$12
2035
$28
$7.1
$21
2040
$37
$8.5
$29
2050
$42
$8.6
$33
PV, 3%
$420
$100
$320
PV, 7%
$210
$51
$150
Annualized, 3%
$21
$5.1
$16
Annualized, 7%
$17
$4.1
$12
Table Note:
Electricity expenditure increases are included.
Table 10-3 presents estimated annual monetized benefits from non-emission sources for the
indicated calendar years. The table also shows the present value of those benefits for the calendar
years 2021-2050 using both 3 percent and 7 percent discount rates.
Table 10-3: Benefits from Non-Emission Sources for the Final Rule (SBillions of 2018 dollars)
Calendar Year
Drive
Value
Refueling Time
Savings
Energy Security
Benefits
Total Non-Emission
Benefits
2023
$0,035
-$0.0052
$0,031
$0,061
2026
$0.14
-$0.12
$0.18
$0.2
2030
$0.55
-$0.27
$0.51
$0.79
2035
$1
-$0.47
$0.92
$1.5
2040
$1.3
-$0.67
$1.3
$1.9
2050
$1.5
-$0.83
$1.6
$2.3
PV, 3%
$15
$-7.4
$14
$21
PV, 7%
$7.2
$-3.6
$7
$11
Annualized,
3%
$0.75
$-0.38
$0.73
$1.1
Annualized,
7%
$0.58
$-0.29
$0.56
$0.85
Table 10-4 presents estimated annual monetized benefits from emission sources for the
indicated calendar years. The table also shows the present value of those benefits for the calendar
years 2021-2050 using both 3 percent and 7 percent discount rates.
Table 10-5 shows the benefits of reduced GHG emissions, and consequently the annual
quantified benefits (i.e., total GHG benefits), for each of the four interim social cost of GHG
(SC-GHG) values estimated by the interagency working group. As discussed in RIA Chapter 3.3
there are some limitations to the SC-GHG analysis, including the incomplete way in which the
integrated assessment models capture catastrophic and non-catastrophic impacts, their
incomplete treatment of adaptation and technological change, uncertainty in the extrapolation of
damages to high temperatures, and assumptions regarding risk aversion.
10-2
-------
Table 10-4: PIVh.s-related Emission Reduction Benefits of the Final Rule (SBillions of 2018 dollars)
Calendar Year
Tailpipe
Benefits
Upstream Benefits
Total PM2 5-related Benefits
3% DR
7% DR
3% DR
7% DR
3% DR
7% DR
2023
-$0.0034
-$0.0031
$0.02
$0,018
$0,016
$0,015
2026
$0,018
$0,016
$0,097
$0,088
$0.11
$0.1
2030
$0.15
$0.13
$0.45
$0.41
$0.6
$0.54
2035
$0.44
$0.4
$0.79
$0.72
$1.2
$1.1
2040
$0.68
$0.62
$1
$0.95
$1.7
$1.6
2050
$0.89
$0.8
$1.4
$1.3
$2.3
$2.1
PV
$6.7
$2.8
$12
$5.3
$19
$8.1
Annualized
$0.34
$0.22
$0.61
$0.43
$0.96
$0.65
Notes:






a Note that the non-GHG impacts associated with the standards presented here do not include the full complement
of health and environmental effects that, if quantified and monetized, would increase the total monetized benefits.
Instead, the non-GHG benefits are based on benefit-per-ton values that reflect only human health impacts
associated with reductions in PM2 5 exposure.




b Calendar year non-GHG benefits presented in this table assume either a 3 percent or 7 percent discount rate in
the valuation of PM-related premature mortality to account for a twenty-year segmented cessation lag. Note that
annual benefits estimated using a 3 percent discount rate were used to calculate the present and annualized values
using a 3 percent discount rate and the annual benefits estimated using a 7 percent discount rate were used to
calculate the present and annualized values using a 7 percent discount rate.


Table 10-5: Climate Benefits from Reduction in GHG Emissions (SBillions of 2018 dollars)
Calendar Year
Discount Rate and Statistic
5% Average
3% Average
2.5% Average
3% 95th percentile
2023
$0,081
$0.27
$0.4
$0.8
2026
$0.48
$1.6
$2.3
$4.7
2030
$1.5
$4.6
$6.7
$14
2035
$2.8
$8.4
$12
$25
2040
$3.9
$11
$16
$34
2050
$5.5
$14
$20
$44
PV
$31
$130
$200
$390
Annualized
$2
$6.6
$9.5
$20
Notes:
Climate benefits are based on changes (reductions) in CO2, CH4, and N20 emissions and are calculated using four
different estimates of the social cost of carbon (SC-CO2), the social cost of methane (SC-CH4), and the social cost
of nitrous oxide (SC-N20) (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, SC-CH4, and SC-N20 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.
The same discount rate used to discount the value of damages from future emissions (SC-GHGs at 5, 3, 2.5
percent) is used to calculate the present value of SC-GHGs for internal consistency Annual benefits shown are
undiscounted values.	
Table 10-6 presents estimated annual net benefits for the indicated calendar years. The table
also shows the present value of those net benefits for the calendar years 2021-2050 using both 3
percent and 7 percent discount rates. The table includes the benefits of reduced GHG emissions
(and consequently the annual net benefits) for each of the four SC-GHG values considered by
EPA. We estimate that the proposed program would result in a net present value of benefits that
10-3
-------
ranges between $27-$450 billion; that is, the total benefits would far exceed the costs of the
program.
Table 10-6: Net Benefits (Emission Benefits + Non-Emission Benefits + Fuel Savings - Costs) Associated with
the Final Program ($Billions of 2018 dollars)
Calendar
Year
Net Benefits,
with Climate
Benefits based on
5% discount rate
Net Benefits,
with Climate
Benefits based on
3% discount rate
Net Benefits,
with Climate
Benefits based
on 2.5% discount
rate
Net Benefits,
with Climate Benefits based
on 3% discount rate, 95th
percentile SC-GHG
2023
-$5.3
-$5.1
-$5
-$4.6
2026
-$13
-$12
-$11
-$9.1
2030
-$4.6
-$1.4
$0.63
$7.9
2035
$7.8
$13
$17
$30
2040
$19
$26
$31
$49
2050
$27
$36
$41
$66
PV, 3%
$88
$190
$260
$450
PV, 7%
$27
$120
$190
$390
Annualized,
3%
$4.9
$9.5
$12
$23
Annualized,
7%
$1.7
$6.2
$9.2
$20
Notes:
Climate benefits are based on changes (reductions) in CO2, CH4, and N20 emissions and are calculated using four
different estimates of the social cost of carbon (SC-CO2), the social cost of methane (SC-CH4), and the social cost
of nitrous oxide (SC-N20) (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, SC-CH4, and SC-N20 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. The same discount rate used to discount the value
of damages from future emissions (SC-GHG at 5, 3, 2.5 percent) is used to calculate present value of SC-GHGs
for internal consistency, while all other costs and benefits are discounted at either 3 percent or 7 percent. Annual
costs and benefits shown are undiscounted values. Note that the non-GHG impacts associated with the standards
included here do not include the full complement of health and environmental effects that, if quantified and
monetized, would increase the total monetized benefits. Instead, the non-GHG benefits are based on benefit-per-
ton values that reflect only human health impacts associated with reductions in PM2 5 exposure.
10.2 Comparison to Proposal (Less Stringent Alternative)
The same series of tables as presented in Chapter 10.1 for the final standards are presented
here for the Proposal. Note that Table 10-7 includes an estimate of foregone consumer sales
surplus, which measures the loss in benefits attributed to consumers who would have purchased
a new vehicle in the absence of the proposed standards.
10-4
-------
Table 10-7: Costs Associated with the Proposal (SBillions of 2018 dollars)
Calendar
Year
Foregone
Consumer
Sales
Surplusm
Technology
Costs
Congestion
Noise
Fatality
Costs
Non-fatal
Crash
Costs
Total Costs
2023
$0,025
$4.7
$0,014
$0.00021
$0.11
$0.18
$5
2026
$0,061
$10
$0,086
$0.0014
$0.32
$0.54
$11
2030
$0,046
$11
$0.3
$0.0049
$0.28
$0.46
$12
2035
$0,038
$11
$0.5
$0.0082
$0.16
$0.27
$12
2040
$0,031
$9.7
$0.61
$0.0099
$0,089
$0.14
$11
2050
$0,025
$9.1
$0.65
$0,011
$0,072
$0.12
$9.9
PV, 3%
$0.72
$170
$7
$0.11
$3.2
$5.3
$190
PV, 7%
$0.46
$100
$3.5
$0,057
$2.1
$3.5
$110
Annualized
,3%
$0,037
$8.9
$0.35
$0.0058
$0.16
$0.27
$9.8
Annualized
,7%
$0,037
$8.4
$0.28
$0.0046
$0.17
$0.29
$9.2
Notes:
[1] "Foregone Consumer Sales Surplus" refers to the difference between a vehicle's price and the buyer's
willingness to pay for the new vehicle; the impact reflects the reduction in new vehicle sales described in Chapter
8.1. See Section 8 of CAFE Model Documentation FR 2020.pdf in the docket for more information.
Table 10-8 shows the undiscounted annual monetized fuel savings of the Proposal. The table
also shows the present value of those fuel savings for the same calendar years using both 3
percent and 7 percent discount rates. The aggregate value of fuel savings is calculated using pre-
tax fuel prices since savings in fuel taxes do not represent a reduction in the value of economic
resources utilized in producing and consuming fuel. Note that the fuel savings shown in Table
10-8 result from reductions in fleet-wide fuel use. Thus, fuel savings grow over time as an
increasing fraction of the fleet is projected to meet the standards.
Table 10-8: Fuel Savings Associated with the Proposal (SBillions of 2018 dollars)
Calendar Year
Retail Fuel Savings
Fuel Tax Savings
Total Fuel Savings
2023
$0.62
$0.23
$0.39
2026
$3.5
$1.2
$2.3
2030
$11
$3
$7.7
2035
$18
$4.6
$14
2040
$24
$5.4
$18
2050
$27
$5.5
$22
PV, 3%
$270
$65
$210
PV, 7%
$130
$33
$100
Annualized, 3%
$14
$3.3
$11
Annualized, 7%
$11
$2.7
$8.2
Table 10-9 presents estimated annual monetized benefits from non-emission sources for the
indicated calendar years. The table also shows the present value of those benefits for the calendar
years 2021-2050 using both 3 percent and 7 percent discount rates.
10-5
-------
Table 10-9: Benefits from Non-Emission Sources Associated with the Proposal (SBillions of 2018 dollars)
Calendar Year
Drive
Refueling Time
Energy Security
Total Non-Emission

Value
Savings
Benefits
Benefits
2023
$0,013
-$0,019
$0,023
$0,017
2026
$0,085
-$0.12
$0.13
$0,094
2030
$0.38
-$0.19
$0.34
$0.54
2035
$0.72
-$0.29
$0.6
$1
2040
$0.93
-$0.34
$0.81
$1.4
2050
$1
-$0.49
$1
$1.6
PV, 3%
$10
$-4.4
$9.3
$15
PV, 7%
$5
$-2.3
$4.6
$7.3
Annualized,
$0.52
$-0.22
$0.47
$0.77
3%




Annualized,
$0.4
$-0.18
$0.37
$0.59
7%




Table 10-10 presents estimated annual monetized benefits from emission sources for the
indicated calendar years. The table also shows the present value of those benefits for the calendar
years 2021-2050 using both 3 percent and 7 percent discount rates.
Table 10-10: PlVhs-related Emission Reduction Benefits Associated with the Proposal (SBillions of 2018
dollars)
Calendar Year
Tailpipe
Benefits
Upstream Benefits
Total PlVhs-related Benefits
3% DR
7% DR
3% DR
7% DR
3% DR
7% DR
2023
-$0.0023
-$0.0021
$0.0017
$0.0016
-$0.00061
-$0.00053
2026
$0,012
$0,011
$0,048
$0,045
$0,061
$0,056
2030
$0.1
$0,094
$0.29
$0.27
$0.4
$0.36
2035
$0.29
$0.26
$0.52
$0.47
$0.81
$0.73
2040
$0.43
$0.39
$0.68
$0.62
$1.1
$1
2050
$0.56
$0.51
$0.91
$0.82
$1.5
$1.3
PV
$4.3
$1.8
$7.7
$3.4
$12
$5.2
Annualized
$0.22
$0.14
$0.39
$0.27
$0.61
$0.42
Notes:






a Note that the non-GHG impacts associated with the standards presented here do not include the full complement
of health and environmental effects that, if quantified and monetized, would increase the total monetized benefits.
Instead, the non-GHG benefits are based on benefit-per-ton values that reflect only human health impacts
associated with reductions in PM2 5 exposure.




b Calendar year non-GHG benefits presented in this table assume either a 3 percent or 7 percent discount rate in
the valuation of PM-related premature mortality to account for a twenty-year segmented cessation lag. Note that
annual benefits estimated using a 3 percent discount rate were used to calculate the present and annualized values
using a 3 percent discount rate and the annual benefits estimated using a 7 percent discount rate were used to
calculate the present and annualized values using a 7 percent discount rate.


Table 10-11 shows the benefits of reduced GHG emissions, and consequently the annual
quantified benefits (i.e., total benefits), for each of the four interim social cost of GHG (SC-
GHG) values estimated by the interagency working group. As discussed in RIA Chapter 3.3
there are some limitations to the SC-GHG analysis, including the incomplete way in which the
integrated assessment models capture catastrophic and non-catastrophic impacts, their
incomplete treatment of adaptation and technological change, uncertainty in the extrapolation of
damages to high temperatures, and assumptions regarding risk aversion.
10-6
-------
Table 10-11: Climate Benefits from Reduction in Greenhouse Gas Emissions Associated with the Proposal
($Billions of 2018 dollars)
Calendar Year
Discount Rate and Statistic
5% Average
3% Average
2.5% Average
3% 95th percentile
2023
$0,055
$0.18
$0.27
$0.55
2026
$0.34
$1.1
$1.6
$3.3
2030
$0.99
$3.1
$4.5
$9.3
2035
$1.8
$5.5
$7.8
$17
2040
$2.5
$7.2
$10
$22
2050
$3.5
$9.2
$13
$28
PV
$20
$83
$130
$250
Annualized
$1.3
$4.3
$6.2
$13
Notes:
Climate benefits are based on changes (reductions) in CO2, CH4, and N20 emissions and are calculated using four
different estimates of the social cost of carbon (SC-CO2), the social cost of methane (SC-CH4), and the social cost
of nitrous oxide (SC-N20) (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, SC-CH4, and SC-N20 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.
The same discount rate used to discount the value of damages from future emissions (SC-GHGs at 5, 3, 2.5
percent) is used to calculate the present value of SC-GHGs for internal consistency Annual benefits shown are
undiscounted values.
Table 10-12 presents estimated annual net benefits for the indicated calendar years. The table
also shows the present value of those net benefits for the calendar years 2021-2050 using both 3
percent and 7 percent discount rates. The table includes the benefits of reduced GHG emissions
(and consequently the annual net benefits) for each of the four SC-GHG values considered by
EPA.
10-7
-------
Table 10-12: Net Benefits (Emission Benefits + Non-Emission Benefits + Fuel Savings - Costs) for the
Proposal (SBillions of 2018 dollars)
Calendar
Year
Net Benefits,
with Climate
Benefits based on
5% discount rate
Net Benefits,
with Climate
Benefits based on
3% discount rate
Net Benefits,
with Climate
Benefits based
on 2.5% discount
rate
Net Benefits,
with Climate Benefits based
on 3% discount rate, 95th
percentile SC-GHG
2023
-$4.6
-$4.4
-$4.3
-$4.1
2026
-$8.6
-$7.8
-$7.3
-$5.7
2030
-$2.2
-$0,077
$1.3
$6.2
2035
$5.9
$9.5
$12
$21
2040
$13
$18
$20
$32
2050
$18
$24
$27
$43
PV, 3%
$62
$130
$170
$300
PV, 7%
$19
$82
$130
$250
Annualized,
3%
$3.5
$6.4
$8.3
$15
Annualized,
7%
$1.2
$4.2
$6.1
$13
Notes:
a Climate benefits are based on changes (reductions) in CO2, CH4, and N20 emissions and are calculated using
four different estimates of the social cost of carbon (SC-CO2), the social cost of methane (SC-CH4), and the social
cost of nitrous oxide (SC-N20) (model average at 2.5 percent, 3 percent, and 5 percent discount rates; 95th
percentile at 3 percent discount rate). We emphasize the importance and value of considering the benefits
calculated using all four SC-CO2, SC-CH4, and SC-N20 estimates. As discussed in the Technical Support
Document: Social Cost of Carbon, Methane, and Nitrous Oxide Interim Estimates under Executive Order 13990
(IWG 2021), a consideration of climate benefits calculated using discount rates below 3 percent, including 2
percent and lower, are also warranted when discounting intergenerational impacts. The same discount rate used to
discount the value of damages from future emissions (SC-GHG at 5, 3, 2.5 percent) is used to calculate present
value of SC-GHGs for internal consistency, while all other costs and benefits are discounted at either 3 percent or
7 percent. Annual costs and benefits shown are undiscounted values.
10.3 Comparison to the More Stringent Alternative 2 Minus 10
The same series of tables as presented in Chapter 10.1 for the final standards are presented
here for the more stringent standards, referred to as Alternative 2 minus 10. Note that Table
10-13 includes an estimate of foregone consumer sales surplus, which measures the loss in
benefits attributed to consumers who would have purchased a new vehicle in the absence of the
standards.
10-8
-------
Table 10-13: Costs Associated with the Alternative 2 minus 10 ($Billions of 2018 dollars)
Calendar
Year
Foregone Consumer
Sales Surplus Til
Technology
Costs
Congestion
Noise
Fatality
Costs
Non-fatal
Crash Costs
Total
Costs
2023
$0.06
$10
$0,051
$0,000
8
$0.24
$0.4
$11
2026
$0.12
$17
$0.18
$0,002
9
$0.5
$0.83
$19
2030
$0,096
$18
$0.45
$0,007
6
$0.46
$0.75
$19
2035
$0,081
$17
$0.72
$0,012
$0.25
$0.41
$19
2040
$0,066
$16
$0.86
$0,014
$0.13
$0.22
$17
2050
$0,054
$15
$0.9
$0,015
$0.13
$0.21
$16
PV, 3%
$1.5
$290
$10
$0.17
$5.3
$8.7
$320
PV, 7%
$0.94
$180
$5.2
$0,088
$3.6
$5.9
$190
Annualiz
ed, 3%
$0,075
$15
$0.52
$0,008
7
$0.27
$0.44
$16
Annualiz
ed, 7%
$0,076
$14
$0.42
$0,007
1
$0.29
$0.48
$15
[1] "Foregone Consumer Sales Surplus" refers to the difference between a vehicle's price and the buyer's
willingness to pay for the new vehicle; the impact reflects the reduction in new vehicle sales described in Chapter
8.1. See Section 8 of CAFE Model Documentation FR 2020.pdf in the docket for more information.
Table 10-14 shows the undiscounted annual monetized fuel savings of Alternative 2 minus
10. The table also shows the present value of those fuel savings for the same calendar years using
both 3 percent and 7 percent discount rates. The aggregate value of fuel savings is calculated
using pre-tax fuel prices since savings in fuel taxes do not represent a reduction in the value of
economic resources utilized in producing and consuming fuel. Note that the fuel savings shown
in Table 10-14 result from reductions in fleet-wide fuel use. Thus, fuel savings grow over time as
an increasing fraction of the fleet is projected to meet the standards.
Table 10-14: Fuel Savings Associated with Alternative 2 minus 10 ($Billions of 2018 dollars)
Calendar Year
Retail Fuel Savings
Fuel Tax Savings
Total Fuel Savings
2023
$1.4
$0.47
$0.94
2026
$6
$2
$4
2030
$17
$4.7
$12
2035
$29
$7.3
$21
2040
$37
$8.5
$29
2050
$42
$8.7
$33
PV, 3%
$430
$100
$320
PV, 7%
$210
$53
$160
Annualized, 3%
$22
$5.2
$16
Annualized, 7%
$17
$4.2
$13
Table 10-15 presents estimated annual monetized benefits from non-emission sources for the
indicated calendar years. The table also shows the present value of those benefits for the calendar
years 2021-2050 using both 3 percent and 7 percent discount rates.
10-9
-------
Table 10-15: Benefits from Non-Emission Sources Associated with Alternative 2 minus 10 ($Billions of 2018
dollars)
Calendar Year
Drive
Value
Refueling Time
Savings
Energy Security
Benefits
Total Non-Emission
Benefits*
2023
$0.06
-$0,015
$0,047
$0,092
2026
$0.22
-$0,094
$0.21
$0.33
2030
$0.63
-$0.27
$0.54
$0.9
2035
$1.1
-$0.45
$0.94
$1.6
2040
$1.4
-$0.55
$1.3
$2.1
2050
$1.5
-$0.86
$1.6
$2.3
PV, 3%
$16
$-6.7
$15
$24
PV, 7%
$7.9
$-3.3
$7.2
$12
Annualized,
3%
$0.81
$-0.34
$0.75
$1.2
Annualized,
7%
$0.64
$-0.27
$0.58
$0.95
Table 10-16 presents estimated annual monetized benefits from emission sources for the
indicated calendar years. The table also shows the present value of those benefits for the calendar
years 2021-2050 using both 3 percent and 7 percent discount rates.
Table 10-16: PIVh.s-related Emission Reduction Benefits Associated with Alternative 2 minus 10 ($Billions of
2018 dollars)
Calendar Year
Tailpipe
Benefits
Upstream Benefits
Total PM2 5-related Benefits
3% DR
7% DR
3% DR
7% DR
3% DR
7% DR
2023
-$0.0076
-$0.0068
$0,031
$0,028
$0,023
$0,021
2026
$0,018
$0,016
$0.15
$0.13
$0.17
$0.15
2030
$0.16
$0.14
$0.5
$0.45
$0.66
$0.6
2035
$0.46
$0.42
$0.81
$0.74
$1.3
$1.2
2040
$0.71
$0.64
$1.1
$0.95
$1.8
$1.6
2050
$0.93
$0.84
$1.4
$1.3
$2.3
$2.1
PV
$7
$2.9
$12
$5.5
$19
$8.4
Annualized
$0.36
$0.23
$0.63
$0.45
$0.99
$0.68
Notes:






" Note that the non-GHG impacts associated with the standards presented here do not include the full complement
of health and environmental effects that, if quantified and monetized, would increase the total monetized benefits.
Instead, the non-GHG benefits are based on benefit-per-ton values that reflect only human health impacts
associated with reductions in PM2 5 exposure.




b Calendar year non-GHG benefits presented in this table assume either a 3 percent or 7 percent discount rate in
the valuation of PM-related premature mortality to account for a twenty-year segmented cessation lag. Note that
annual benefits estimated using a 3 percent discount rate were used to calculate the present and annualized values
using a 3 percent discount rate and the annual benefits estimated using a 7 percent discount rate were used to
calculate the present and annualized values using a 7 percent discount rate.


Table 10-17 shows the benefits of reduced GHG emissions, and consequently the annual
quantified benefits (i.e., total benefits), for each of the four interim social cost of GHG (SC-
GHG) values estimated by the interagency working group. As discussed in RIA Chapter 3.3
there are some limitations to the SC-GHG analysis, including the incomplete way in which the
integrated assessment models capture catastrophic and non-catastrophic impacts, their
incomplete treatment of adaptation and technological change, uncertainty in the extrapolation of
damages to high temperatures, and assumptions regarding risk aversion.
10-10
-------
Table 10-17: Climate Benefits from Reduction in Greenhouse Gas Emissions Associated with Alternative 2
minus 10 ($Billions of 2018 dollars)
Calendar Year
Discount Rate and Statistic
5% Average
3% Average
2.5% Average
3% 95th percentile
2023
$0.12
$0.41
$0.6
$1.2
2026
$0.56
$1.8
$2.7
$5.5
2030
$1.6
$4.9
$7.1
$15
2035
$2.9
$8.6
$12
$26
2040
$3.9
$11
$16
$34
2050
$5.5
$14
$20
$44
PV
$32
$130
$200
$400
Annualized
$2.1
$6.7
$9.7
$20
Notes:
Climate benefits are based on changes (reductions) in CO2, CH4, and N20 emissions and are calculated using four
different estimates of the social cost of carbon (SC-CO2), the social cost of methane (SC-CH4), and the social cost
of nitrous oxide (SC-N20) (model average at 2.5 percent, 3 percent, and 5 percent discount rates; 95th percentile
at 3 percent discount rate). We emphasize the importance and value of considering the benefits calculated using
all four SC-CO2, SC-CH4, and SC-N20 estimates. As discussed in the Technical Support Document: Social Cost
of Carbon, Methane, and Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG 2021), a
consideration of climate benefits calculated using discount rates below 3 percent, including 2 percent and lower,
are also warranted when discounting intergenerational impacts.
The same discount rate used to discount the value of damages from future emissions (SC-GHGs at 5, 3, 2.5
percent) is used to calculate the present value of SC-GHGs for internal consistency Annual benefits shown are
undiscounted values.
Table 10-18 presents estimated annual net benefits for the indicated calendar years. The table
also shows the present value of those net benefits for the calendar years 2021-2050 using both 3
percent and 7 percent discount rates. The table includes the benefits of reduced GHG emissions
(and consequently the annual net benefits) for each of the four SC-GHG values considered by
EPA.
10-11
-------
Table 10-18: Net Benefits (Emission Benefits + Non-Emission Benefits + Fuel Savings - Costs) Associated
with Alternative 2 minus 10 ($Billions of 2018 dollars)
Calendar
Year
Net Benefits,
with Climate
Benefits based on
5% discount rate
Net Benefits,
with Climate
Benefits based on
3% discount rate
Net Benefits,
with Climate
Benefits based
on 2.5% discount
rate
Net Benefits,
with Climate Benefits based
on 3% discount rate, 95th
percentile SC-GHG
2023
-$9.6
-$9.3
-$9.1
-$8.5
2026
-$14
-$12
-$12
OO
OO
&
1
2030
-$4.1
-$0.77
$1.4
$9.1
2035
$8.1
$14
$17
$31
2040
$19
$26
$31
$49
2050
$27
$36
$41
$65
PV, 3%
$80
$180
$250
$450
PV, 7%
$19
$120
$190
$390
Annualized,
3%
$4.5
$9.2
$12
$23
Annualized,
7%
$1.1
$5.7
$8.7
$19
Notes:
Climate benefits are based on changes (reductions) in CO2, CH4, and N20 emissions and are calculated using four
different estimates of the social cost of carbon (SC-CO2), the social cost of methane (SC-CH4), and the social cost
of nitrous oxide (SC-N20) (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, SC-CH4, and SC-N20 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. The same discount rate used to discount the value
of damages from future emissions (SC-GHG at 5, 3, 2.5 percent) is used to calculate present value of SC-GHGs
for internal consistency, while all other costs and benefits are discounted at either 3 percent or 7 percent. Annual
costs and benefits shown are undiscounted values.
10.4 Sensitivities
We have conducted the following sensitivities:
•	AEO high oil price (AEO high)
•	AEO low oil price (AEO low)
•	Allow HCR2 in MY 2025 and later (Allow HCR2)
•	Battery costs higher (uses NPRM battery costs)
•	Battery costs lower (battery costs roughly 24 percent lower than the updated FRM
costs)
•	Sales demand elasticity of-0.15
•	Sales demand elasticity of-1.0
•	Mass safety coefficients at the 5th percentile (Mass safety 5th pctile)
•	Mass safety coefficients at the 95th percentile (Mass safety 95th pctile)
•	No further applicaton of mild or string hybrid technology (no hybrids)
•	Perfect trading, which allows perfect trading of CO2 credits between manufacturers57
57 To simulate perfect trading, the entire fleet is attributed to a single manufacturer, dubbed "Industry," in the market
input file.
10-12
-------
•	Rebound rate of -5 percent
•	Rebound rate of -15 percent
Each sensitivity is compared to its own no action scenario. In other words, the no action
standards were used but the no action scenario was run using the same set of sensitivity
parameters as used for the action scenario.
Table 10-19 Monetized Discounted Costs, Benefits, and Net Benefits of the Proposed Program and each
Sensitivity for Calendar Years through 2050 ($Billions of 2018 dollars, 3 percent Discount Rate)a'b'c'd

Final
AEO
High
AEO
Low
Allow
HCR2
Higher
Battery
Costs
Lower
Battery
Costs
Demand
elast. of -
0.15
Demand
elast. of -
1.0
Mass
safety at
5th
pctile
Mass
safety at
95
pctile
No
Hybrids
Perfect
Trading
Rebound
of 15%
Rebound
of 5%
Costs
$300
$260
$330
$300
$360
$240
$300
$300
$280
$320
$310
$330
$310
$290
Fuel
Savings
$320
$510
$170
$320
$330
$310
$320
$310
$320
$320
$310
$360
$320
$320
Benefits
$170
$150
$190
$170
$190
$160
$170
$170
$170
$170
$160
$190
$180
$160
Net
Benefits
$190
$400
$30
$190
$150
$230
$190
$180
$210
$160
$160
$220
$190
$190
Notes:
a Values rounded to two significant figures; totals may not sum due to rounding. Present and annualized values are based on the stream of
annual calendar year costs and benefits included in the analysis (2021 - 2050) and discounted back to year 2021.
b Climate benefits are based on reductions in C02, CH4 and N20 emissions and are calculated using four different estimates of the social cost of
each greenhouse gas (SC-GHG model average at 2.5 percent, 3 percent, and 5 percent discount rates; 95th percentile at 3 percent discount rate),
which each increase over time. For the presentational purposes of this table, we show the benefits associated with the average SC-GHGs at a 3
percent discount rate, but the Agency does not have a single central SC-GHG point estimate. We emphasize the importance and value of
considering the benefits calculated using all four SC-GHG estimates and present them later in this RIA. As discussed in Chapter 3.3 of the RIA,
a consideration of climate benefits calculated using discount rates below 3 percent, including 2 percent and lower, are also warranted when
discounting intergenerational impacts.
c The same discount rate used to discount the value of damages from future GHG emissions (SC-GHGs at 5, 3, and 2.5 percent) is used to
calculate the present and annualized values of climate benefits for internal consistency, while all other costs and benefits are discounted at either
3 percent or 7 percent.
d Net benefits reflect the fuel savings plus benefits minus costs.
e Non-GHG impacts associated with the standards presented here do not include the full complement of health and environmental effects that, if
quantified and monetized, would increase the total monetized benefits. Instead, the non-GHG benefits are based on benefit-per-ton values that
reflect only human health impacts associated with reductions in PM2.5 exposure.	
10-13
-------
Table 10-20 Monetized Discounted Costs, Benefits, and Net Benefits of the Proposed Program and each
Sensitivity for Calendar Years through 2050 ($Billions of 2018 dollars, 7 percent Discount Rate)a'b'c'd

Final
AEO
High
AEO
Low
Allow
HCR2
Higher
Battery
Costs
Lower
Battery
Costs
Demand
elast. of -
0.15
Demand
elast. of -
1.0
Mass
safety at
5th
pctile
Mass
safety at
95
pctile
No
Hybrids
Perfect
Trading
Rebound
of 15%
Rebound
of 5%
Costs
$180
$160
$190
$180
$220
$150
$180
$180
$170
$190
$180
$190
$180
$170
Fuel
Savings
$150
$250
$83
$160
$160
$150
$160
$150
$150
$150
$150
$170
$150
$160
Benefits
$150
$130
$170
$150
$160
$140
$150
$150
$150
$150
$140
$170
$150
$140
Net
Benefits
$130
$230
$55
$130
$95
$150
$130
$120
$140
$110
$110
$150
$120
$120
Notes:
a Values rounded to two significant figures; totals may not sum due to rounding. Present and annualized values are based on the stream of
annual calendar year costs and benefits included in the analysis (2021 - 2050) and discounted back to year 2021.
b Climate benefits are based on reductions in CO2, CH4 and N20 emissions and are calculated using four different estimates of the social cost of
each greenhouse gas (SC-GHG model average at 2.5 percent, 3 percent, and 5 percent discount rates; 95th percentile at 3 percent discount rate),
which each increase over time. For the presentational purposes of this table, we show the benefits associated with the average SC-GHGs at a 3
percent discount rate, but the Agency does not have a single central SC-GHG point estimate. We emphasize the importance and value of
considering the benefits calculated using all four SC-GHG estimates and present them later in this RIA. As discussed in Chapter 3.3 of the RIA,
a consideration of climate benefits calculated using discount rates below 3 percent, including 2 percent and lower, are also warranted when
discounting intergenerational impacts.
c The same discount rate used to discount the value of damages from future GHG emissions (SC-GHGs at 5, 3, and 2.5 percent) is used to
calculate the present and annualized values of climate benefits for internal consistency, while all other costs and benefits are discounted at either
3 percent or 7 percent.
d Net benefits reflect the fuel savings plus benefits minus costs.
e Non-GHG impacts associated with the standards presented here do not include the full complement of health and environmental effects that, if
quantified and monetized, would increase the total monetized benefits. Instead, the non-GHG benefits are based on benefit-per-ton values that
reflect only human health impacts associated with reductions in PM2.5 exposure.	
10-14
-------