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
SEPA
United States
Environmental Protection
Agency
-------
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
U.S. Environmental Protection Agency
Assessment and Standards Division
United States
Environmental Protection
^1 Agency
EPA-420-R-17-002
January 2017
-------
Table of Contents
Introduction 1
Chapter 1: General and Process 4
1.1 General Comments 4
1.2 Legal Process and Timing 5
Chapter 2: Technology Assessment 12
2.1 Effectiveness Assessment: General Comments, Technology Packages, Penetrations, 12
and Sufficiency of Non-Electrified Technologies 12
2.2 Effectiveness Modeling and Quality Assurance 25
2.2.1 ALPHA Model 25
2.2.2 Lumped Parameter Model (LPM) 26
2.2.3 Quality Assurance / Plausibility Checks, ALPHA-LPM Calibration 29
2.2.4 Vehicle Classifications 40
2.2.5 Performance Neutrality 41
2.3 Estimated Costs (Technology Costs, Total Costs, Learning) 43
2.4 Lead Time 46
2.5 Individual Technologies 49
2.5.1 Atkinson Cycle Engine 49
2.5.2 Turbo Downsizing 55
2.5.3 Other Engine Technologies (Cylinder Deactivation, Cam Phasing, Variable Valve 64
Lift) 64
2.5.4 Transmissions 66
2.5.5 Battery Technology / Cost 71
2.5.6 Non-battery Technology / Cost 73
2.5.7 Stop-Start 74
2.5.8 Mild Hybrid (48V) 75
2.5.9 Strong Hybrid 76
2.5.10 Plug-in Vehicles 77
2.5.10.1 Plug-in Hybrid Electric Vehicles (PHEVs) 77
2.5.10.2 Battery Electric Vehicles (BEVs) 78
2.5.11 Fuels/Octane 80
2.5.12 Mass Reduction 82
2.5.13 Aerodynamics 87
2.5.14 Tire Rolling Resistance 89
2.5.15 Low Drag Brakes 90
2.5.16 Air Conditioning 91
2.5.16.1 A/C Efficiency Credits 91
2.5.16.2 A/C Refrigerant Credits 93
2.6 Criteria Emissions / Tier 3 94
2.7 Baseline Fleet 95
2.7.1 Technologies in Baseline 97
2.7.2 The ZEV Program in the OMEGA Analysis Fleet 99
2.8 OMEGA 102
Chapter 3: Economic / Consumer / Other Factors 120
3.1 Consumer Response 120
3.2 Consumer Impacts of New Technologies 128
-------
3.3 Affordability 132
3.4 Employment 138
3.5 Economic and Other Key Inputs 143
3.6 Safety 146
3.7 Alternative Fuel (PEV) Infrastructure 148
3.8 Standards Design 150
3.9 Credits, Incentives, and Flexibilities 152
3.10 Harmonization 157
Chapter 4: Climate Science and the Further GHG Reductions Beyond 2025 160
4.1 Climate Science 160
4.2 Post-2025 Standards 162
-------
List of Acronyms
2MHEV
2-Mode Hybrid
ABS
Anti-lock Braking System
ABT
Averaging, Banking, and Trading
AC
Alternating Current
A/C
Air Conditioning
ACEEE
American Council for an Energy-Efficient Economy
AEO
Annual Energy Outlook
AER
All-Electric Range
AFDC
Alternative Fuels Data Center
AGM
Absorbent Glass Mat
AHSS
Advanced High Strength Steel
ALPHA
Advanced Light-Duty Powertrain and Hybrid Analysis Tool
AMT
Automated Manual Transmission
ANL
Argonne National Laboratory
ARB
California Air Resources Board
ASI
Area Specific Impedance
ASL
Aggressive Shift Logic
ASM
Annual Survey of Manufacturers
AT
Automatic Transmissions
Avg
Average
AWD
All Wheel Drive
BenMAP
Benefits Mapping and Analysis Program
BEV
Battery Electric Vehicle
BISG
Belt Integrated Starter Generator
BIW
Body-In-White
BLS
Bureau of Labor Statistics
BMEP
Brake Mean Effective Pressure
BOM
Bill of Materials
BSFC
Brake Specific Fuel Consumption
BTE
Brake-Thermal Efficiency
BTU
British Thermal Unit
CAA
Clean Air Act
CAD
Computer Aided Designs
CAD/CAE
Computer Aided Design and Engineering
CAE
Computer Aided Engineering
CAFE
Corporate Average Fuel Economy
CARB
California Air Resources Board
CAVs
Connected and Automated (or autonomous) Vehicles
CBD
Center for Biological Diversity
CBI
Confidential Business Information
CCP
Coupled Cam Phasing
1
-------
CDPF Catalyzed Diesel Particulate Filter
CEC California Energy Commission
cEGR Cooled Exhaust Gas Recirculation
CES Consumer Expenditure Survey
CFD Computational Fluid Dynamics
CFR Code of Federal Regulations
CH4 Methane
CISG Crank Integrated Starter Generator
CNG Compressed Natural Gas
CO Carbon Monoxide
COI Conflict Of Interest
CO2 Carbon Dioxide
C02eq CO2 Equivalent
COP Coefficient of Performance
CSM Conceptual Site Model
CSV Comma-separated Values
CUV Crossover Utility Vehicles
CVT Continuously Variable Transmission
CY Calendar Year
DC Direct Current
DCFC Direct Carbon Fuel Cell
DCP Dual Cam Phasing
DCT Dual Clutch Transmission
DEAC Cylinder Deactivation
DFMA Design for Manufacturing and Assembly
DGS California Department of General Services
DICE Dynamic Integrated Climate and Economy
DMC Direct Manufacturing Costs
DoE Department of Energy
DOE Design of Experiments
DOHC Dual Overhead Camshaft Engines
DOT Department of Transportation
DRI Dynamic Research, Inc.
DRLs Daytime Running Lamps
DVVL Discrete Variable Valve Lift
EGR Exhaust Gas Recirculation
EHPS Electrohydraulic Power Steering
Energy Information Administration (part of the U.S. Department of
iii/\ t-> \
Energy)
EISA Energy Independence and Security Act
EIVC Early Intake Valve Closing
EPA Environmental Protection Agency
EPCA Energy Policy and Conservation Act
11
-------
EPRI Electric Power Research Institute
EPS Electric Power Steering
EPS Energy Power Systems
EREV Extended Range El ectri c Vehi cl e
ERM Employment Requirements Matrix
ESC Electronic Stability Control
EV Electric Vehicle
EVSE Electric Vehicle Supply Equipment
FARS Fatality Analysis Reporting System
FCEV Fuel Cell Electric Vehicle
FCPM Fuel Cost Per Mile
FCEV Fuel Cell Electric Vehicle
FE Finite Element
FEV1 Functional Expiratory Volume
FHWA Federal Highway Administration
FMEP Friction Mean Effective Pressure
FMVSS Federal Motor Vehicle Safety Standards
FR Federal Register
FRIA Final Regulatory Impact Analysis
FRM Final Rulemaking
FTP Federal Test Procedure
gal/mi Gallon/Mile
GCWR Gross Combined Weight Rating
GDI Gasoline Direct Injection
GDP Gross Domestic Product
GEM Greenhouse gas Emissions Model
GHG Greenhouse Gases
Greenhouse Gases, Regulated Emissions, and Energy Use in
C t R F, F,1 rp , . •
Transportation
GVW Gross Vehicle Weight
GWP Global Warming Potential
GWU George Washington University
HD Heavy-Duty
HEV Hybrid Electric Vehicle
HFC Hydrofluorocarbon
HFET Highway Fuel Economy Dynamometer Procedure
HIL Hardware-In-Loop
hp Horsepower
hrs Hours
HP/WT Horsepower Divided by Weight
HVAC Heating, Ventilating, And Air Conditioning
hz Hertz
IACC Improved Accessories
111
-------
IAM Integrated Assessment Models
IATC Improved Automatic Transmission Control
IC Indirect Cost
ICCT International Council on Clean Transportation
ICF ICF International
ICM Indirect Cost Multiplier
IHX Internal Heat Exchanger
IMA Improved Mobile Assist
IMAC Improved Mobile Air Conditioning
INL Idaho National Laboratory
IOU Investor Owned Utilities
IPCC Intergovernmental Panel on Climate Change
IPM Integrated Planning Model
ITC Institute of Transportation Studies
IWG Interagency Working Group
k Thousand
kg Kilogram
kW Kilowatt
kWh kilowatt-hour
L Liter
lb Pound
LBNL Lawrence Berkeley National Laboratory
LD Light-Duty
LEV Low-Emission Vehicle
LHD Light Heavy-Duty
LD V Li ght Duty Vehi cl e
LNT Lean NOx Trap
LPM Lumped Parameter Model
LRR Lower Rolling Resistance
LT Light Trucks
LWT Lightweighted Pickup Truck
MAD Minimum Absolute Deviation
MBPD Million Barrels Per Day
MD Medium-Duty
MDPV Medium-Duty Passenger Vehicles
MEMA Motor Equipment Manufacturers Association
Mg Megagram
mg Milligram
MHEV Mild Hybrid Electric Vehicle
mi mile
min minimum
min Minute
MM Million
IV
-------
MMLV
Multi-Material Lightweight Vehicle
MMT
Million Metric Tons
MOVES
Motor Vehicle Emissions Simulator
mpg
Miles per Gallon
mph
Miles per Hour
MPV
Multi-Purpose Vehicle
MSRP
Manufacturer's Suggested Retail Price
MTE
Mid Term Evaluation
MuD
Multi-Unit Development
MY
Model Year
N20
Nitrous Oxide
NA
Not Applicable
NAAQS
National Ambient Air Quality Standards
NADA
National Automobile Dealers Association
NAS
National Academy of Sciences
NCA
National Climate Assessment
NCAP
New Car Assessment Program
NEMS
National Energy Modeling System
NESHAP
National Emissions Standards for Hazardous Air Pollutants
NF3
Nitrogen Trifluoride
NGO
Non-Governmental Organization
NHTSA
National Highway Traffic Safety Administration
NiMH
Nickel Metal-Hydride
NF3
Nitrogen Trifluoride
NOx
Nitrogen Oxides
N02
Nitrogen Dioxide
NOx
Oxides of Nitrogen
NPRM
Notice of Proposed Rulemaking
NRC
National Research Council
NRC-CAN
National Research Council of Canada
NREL
National Renewable Energy Laboratory
NVH
Noise Vibration and Harshness
NVPP
National Vehicle Population Profiles
OAR
EPA's Office of Air and Radiation
OEM
Original Equipment Manufacturer
OECD
Organization for Economic Cooperation and Development
OHV
Overhead Valve
OLS
Ordinary Least Squares
OMB
EPA's Office of Management and Budget
OPEC
Organization of Petroleum Exporting Countries
ORNL
Oak Ridge National Laboratory
OTAQ
EPA's Office of Transportation and Air Quality
PAGE
Policy Analysis of the Greenhouse Effect
V
-------
PC
Passenger Car
P/E
Power-to-Energy
PEF
Peak Expiratory Flow
PEV
Plug-in Electric Vehicle
PFCs
Perfluorocarbons
PFI
Port-fuel-inj ection
PGM
Platinum Group Metal
PHEV
Plug-in Hybrid Electric Vehicle
PLM
Planar Layered Matrix
PM
Particulate Matter
PM2.5
Fine Particulate Matter (diameter of 2.5 |im or less)
PMSMs
Permanent-Magnet Synchronous Motors
PSHEV
Power-split Hybrid
PSI
Pounds per Square Inch
PWM
Pulse-width Modulated
R&D
Research and Development
RFS2
Renewable Fuel Standard 2
RIA
Regulatory Impact Analysis
RPE
Retail Price Equivalent
RPM
Revolutions per Minute
RSM
Response Surface Models
RTI
RTI International (formerly Research Triangle Institute)
SA
Strategic Analysis, Inc.
SAB
Science Advisory Board
SAB-
Science Advisory Board Environmental Economics Advisory
EEAC
Committee
SAE
Society of Automotive Engineers
SCO3
Soak Control third iteration
see
Social Cost of Carbon
SCR
Selective Catalyst Reduction
sf6
Sulfur Hexafluoride
SGDI
Stoichiometric Gasoline Direct Injection
SHEV
Strong Hybrid Electric Vehicles
SI
Spark-Ignition
SIDI
Spark Ignition Direct Injection
SIL
Software-In-Loop
SMDI
Steel Market Development Institutes
SNAP
Significant New Alternatives Policy
SNPRM
Supplemental Notice of Proposed Rulemaking
S02
Sulfur Dioxide
SOx
Sulfur Oxides
SOC
State of Charge
VI
-------
SOHC
Single Overhead Cam
SOL
Small Overlap
SPR
Strategic Petroleum Reserve
Std
Standard
SUV
Sport Utility Vehicle
TAR
Technical Assessment Report
TC
Total Costs
TCIP
Tire Consumer Information Program
TDC
Top Dead Center
Tds
Direct Solar Transmittance
TFECIP
Tire Fuel Efficiency Consumer Information Program
TPE
Total Primary Energy
TRBDS
Turbocharging and Downsizing
TSD
Technical Support Document
UMTRI
University of Michigan Transportation Research Institute
UTQGS
Uniform Tire Quality Grading Standards
V2V
Vehicle-To-Vehicle
VGI
Vehicle Grid Integration
VIF
Variance Inflation Factor
VMT
Vehicle Miles Traveled
VOC
Volatile Organic Compound
VSL
Vehicle Speed Limiter
VVL
Variable Valve Lift
VVT
Variable Valve Timing
WT/FP
Weight Divided by Footprint
ZEV
Zero Emission Vehicle
Vll
-------
Introduction
The 2012 rulemaking establishing the National Program for federal greenhouse gas (GHG)
emissions and corporate average fuel economy (CAFE) standards for model years (MY) 2017-
2025 light-duty vehicles included a regulatory requirement for the Environmental Protection
Agency (EPA) to conduct a Midterm Evaluation (MTE) of the GHG standards established for
MY2022-2025.1 In the Final Determination that this document accompanies, the Administrator
is making a final adjudicatory determination (hereafter "Final Determination") that, based on her
evaluation of extensive technical information available to her and significant input from the
industry and other stakeholders, and in light of the factors listed in the 2012 final rule
establishing the MY2017-2025 standards, the MY2022-2025 standards remain appropriate under
section 202(a)(1) of the Clean Air Act. The Final Determination leaves those standards entirely
as they now exist, unaltered.
The Final Determination follows the November 2016 Proposed Determination issued by the
EPA Administrator and the July 2016 release of a Draft Technical Assessment Report (TAR),
issued jointly by the EPA, the National Highway Traffic Safety Administration (NHTSA), and
the California Air Resources Board (CARB). Opportunities for public comment were provided
for both the Draft TAR and the Proposed Determination. In the Draft TAR, the agencies
examined a wide range of issues relevant to GHG emissions standards for MY2022-2025, and
shared with the public their initial technical analyses of those issues. The Draft TAR was
required by EPA's regulations as the first step in the Midterm Evaluation process. In developing
the Proposed Determination, the Administrator considered public comments on the Draft TAR
and EPA updated its analyses where appropriate in response to comments and to reflect the latest
available data. The Administrator has likewise considered public input on the Proposed
Determination in developing the Final Determination.
EPA received more than 100,000 public comments on the Proposed Determination, with
comments from about 60 organizations and the rest from individuals, the vast majority of which
are from mass comment campaigns. These public comments have informed the Administrator's
Final Determination, and EPA has responded to the comments in this Response to Comments
(RTC) document. Many of the comments received included the same or similar information as
that we received on the Draft TAR, to which we previously responded in the Proposed
Determination document and its accompanying Technical Support Document (TSD). This RTC
document, together with the Final Determination, Proposed Determination, the Appendices to the
Proposed Determination, and the TSD to the Proposed Determination should be considered
collectively as EPA's response to all of the significant comments received on EPA's Midterm
Evaluation.2
1 40 CFR 86.1818-12(h).
2 The Final Determination, this RTC document, the Proposed Determination, the Appendices to the Proposed
Determination, and the TSD to the Proposed Determination are contained in EPA Docket ID No. EPA-HQ-OAR-
2015-0827 and can be found at https://www.epa.gov/regulations-emissions-vehicles-and-engines/midterm-
evaluation-light-duty-vehicle-greenhouse-gas-ghg.
1
-------
List of Commenters for the Proposed Determination
The following represents the list of commenter who submitted comments on the Proposed
Determination:
Acadia Center et al.
Achates Power, Inc.
Adsorbed Natural Gas Products, Inc. (ANGP)
Advanced Biofuels USA
Alliance of Automobile Manufacturers
American Chemistry Council (ACC)
American Coalition for Ethanol (ACE)
American Council for an Energy-Efficient Economy (ACEEE)
American Iron and Steel Institute (AISI)
American Lung Association et al.
BlueGreen Alliance
BMW Group
Boyden Grey & Associates PLLC on behalf of Energy Future Coalition (EFC) and Urban
Air Initiative (UAI)
Business for Innovative Climate and Energy Policy (BICEP), a project of Ceres
California Air Resources Board (CARB)
California State Legislature et al.
California State Teachers' Retirement System et al.
Carnegie Mellon University et al.
Center for Automotive Research (CAR)
Center for Biological Diversity (CBD)
Consumer Federation of America (CFA)
Consumers Union (CU)
Denso International America, Inc.
Edison Electric Institute (EEI)
Enhanced Protective Glass Automotive Association (EPGAA)
Environmental Defense Fund (EDF) et al.
Faraday Future
Fiat Chrysler Automobiles U.S. LLC (FCA)
Ford Motor Company
Fuel Freedom Foundation
General Motors
Growth Energy
Heritage Foundation
High Octane, Low Carbon (HOLC)
Honda Motor Co., Inc.
Honeywell
Institute for Energy Research (IER)
International Council on Clean Transportation (ICCT)
Manufacturers of Emission Controls Association (MECA)
Mercedes-Benz USA, LLC and Daimler AG
Michigan League of Conservation Voters (LCV)
2
-------
Mike DeWine, Ohio Attorney General and Bill Schuette, Michigan Attorney General
Minnesota Corn Growers Association (MCGA) and the
Illinois Corn Growers Association (ICGA)
Motor & Equipment Manufacturers Association (MEMA)
National Association of Clean Air Agencies (NACAA)
National Association of Manufacturers (NAM)
National Automobile Dealers Association (NADA)
National Corn Growers Association (NCGA)
Natural Resources Defense Council (NRDC)
New York State Department of Environmental Conservation (NYSDEC)
Nissan North America, Inc.
Northeast States for Coordinated Air Use Management (NESCAUM)
Novation Analytics
Pittsburgh Glass Works, LLC
Renewable Fuels Association (RFA)
Suburu
The Association of Global Automakers, Inc. (Global Automakers)
Toyota Motor North America, Inc.
U.S Chamber of Commerce
Union of Concerned Scientists
United Automobile, Aerospace and Agricultural Implement Workers of America (UAW)
Utah Physicians for a Healthy Environment (UPHE)
Mass Comment Campaign sponsored by Environment America
Mass Comment Campaign sponsored by Environmental Defense Fund (EDF)
Mass Comment Campaign sponsored by Fuel Freedom Foundation
Mass Comment Campaign sponsored by Moms Clean Air Force
Mass Comment Campaign sponsored by Natural Resources Defense Council (NRDC)
Mass Comment Campaign sponsored by Union of Concerned Scientists (UCS)
Mass Comment Campaign sponsored by Washington Environmental Council (WEC)
4 Mass Comment Campaigns sponsored by unknown organizations
149 unique private citizen comments
65 anonymous public comments
3
-------
Chapter 1: General and Process
1.1 General Comments
We received a large number of broad comments on the Proposed Determination that are not
on any specific aspect of the Proposed Determination, but rather are directed generally at the
Proposed Determination finding that the MY2022-2025 standards remain appropriate under
section 202(a)(1) of the Clean Air Act. These general comments include those from both
organizations and private citizens.
Many comments generally supported the Proposed Determination finding that the MY2022-
2025 standards remain appropriate. Examples of these types of comments include those from
Achates Power, American Council for an Energy-Efficient Economy (ACEEE), Consumers
Union, Edison Electric Institute, Honeywell, Manufacturers of Emission Controls Association,
the California Legislature, the California Air Resources Board (CARB), the BlueGreen Alliance,
and others. Some commenters further agreed that the record supports strengthening the
standards, including the International Council on Clean Transportation (ICCT), the Natural
Resources Defense Council (NRDC), the National Association of Clean Air Agencies
(NACAA), Northeast States for Coordinated Air Use Management (NESCAUM), Environmental
Defense Fund, the Union of Concerned Scientists (UCS), and others.
Other comments generally disagreed with the Proposed Determination; examples of these
types of comments include those from the Alliance of Automobile Manufacturers (Alliance),
Global Automakers, Ford, General Motors (GM), FCA, Toyota, Nissan, Subaru, the American
Iron and Steel Institute (AESI), Motor Equipment Manufacturers Association (MEMA), National
Automobile Dealers Association (NADA), U.S. Chamber of Commerce, National Association of
Manufacturers (NAM), Adsorbed Natural Gas Produces (ANGP), High Octane Low Carbon,
National Corn Growers, Minnesota and Illinois Corn Growers Associations, and the Energy
Future Coalition.
We appreciate the time and effort taken by the commenters in developing their comments,
both on the Proposed Determination specifically, and during the many opportunities for public
input throughout the Midterm Evaluation process. We have carefully considered public input on
the Proposed Determination, and these public comments have informed the Administrator's Final
Determination.
Based on her evaluation of extensive technical information available to her and significant
input from stakeholders, and in light of the factors listed in the 2012 final rule establishing the
MY2017-2025 standards, the Administrator is making a final adjudicatory determination that the
MY2022-2025 standards remain appropriate under section 202(a)(1) of the Clean Air Act. We
continue to believe that making the Final Determination now recognizes that long-term
regulatory certainty and stability are important for the automotive industry and contributes to the
continued success of the program, which in turn will reduce emissions, improve fuel economy,
deliver significant fuel savings to consumers, and benefit public health and welfare.
We appreciate the support for the Proposed Determination expressed by many of the
commenters. In our consideration of comments that expressed general opposition to the Proposed
4
-------
Determination, we find no information presented in the public comments on the Proposed
Determination that leads us to change the Agency's analysis in support of the Proposed
Determination. In fact, in many cases, we received similar or identical comments for the Draft
TAR and we responded to them in the Proposed Determination. We respond to comments on the
Midterm Evaluation process, including the Proposed Determination process and timing, in
Section 1.2 below, and to comments on specific aspects of the Proposed Determination
throughout other sections of this Response to Comments (RTC) document.
1.2 Legal Process and Timing
Summary of Comments on the Draft TAR addressed in the Proposed Determination
The Executive Summary and Section I.A of the Proposed Determination provided an
overview of the Midterm Evaluation (MTE) process established by Environmental Protection
Agency (EPA) regulations. EPA did not receive significant comments on the Draft TAR's
description of the MTE process.
Summary of Comments on the Proposed Determination
EPA received many comments relating to the process that the Administrator used in issuing
the Proposed Determination. Several NGO and state government commenters stated that they
believe the process that the Administrator used for this action was appropriate and/or supported
the Administrator moving forward with a Final Determination. These included the
Environmental Defense Fund, which described its legal conclusions that the Administrator's
action is not a rulemaking and that "given the strong and consistent factual record on which the
finding is based, it was likewise appropriate for the Administrator to move forward with her
proposed adjudicatory determination." Other commenters that agreed that the process was
appropriate include the following: Achates Power, Edison Electric Institute, National Association
of Clean Air Agencies, Washington Environmental Council, American Council for an Energy-
Efficient Economy, International Council on Clean Transportation, Natural Resources Defense
Council, Union of Concerned Scientists, and Consumer Federation of America. The Michigan
League of Conservation Voters generally supported the Administrator's action to maintain the
existing GHG standards, although they would have preferred a 30-day extension of the comment
period. The UAW also said it would have preferred "a lengthier midterm review," but also said
that the shortened process "should reduce the likelihood of competing standards and will provide
certainty to an industry that needs ample lead time to plan for production."
Auto manufacturers and their trade groups strongly expressed concerns about several aspects
of the process the Administrator used in issuing the Proposed Determination. Other
organizations, such as some automotive suppliers, echoed some of these process concerns raised
by the auto manufacturing industry. We describe these comments in more detail and respond to
them below. In general, many commenters assert that they were not afforded the required or
expected procedures under the Midterm Evaluation. Several commenters also describe why they
believe the Final Determination is a rule, and thus they believe it is subject to more procedural
requirements than a non-rulemaking action. Commenting on one or more of these issues were
the following automobile manufacturers and their trade organizations: The Alliance of
Automobile Manufacturers (Alliance), Global Automakers, Ford, GM, FCA, Toyota, Nissan,
Subaru, BMW, and Mercedes-Benz. Other commenters expressing similar views on these issues
included DENSO, American Iron and Steel Institute, Motor Equipment Manufacturers
5
-------
Association, National Automobile Dealers Association, U.S. Chamber of Commerce, National
Association of Manufacturers, Adsorbed Natural Gas Products, High Octane Low Carbon,
National Corn Growers Association, Minnesota and Illinois Corn Growers Associations, and the
Energy Future Coalition.
Some commenters, including the Alliance and Global Automakers, commented that by taking
action separate from and prior to MTE-related regulatory action by NHTSA and CARB, EPA's
current action creates a conflict with the agencies' basic principle of a One National Program
(ONP), that allows manufacturers to build a single national light-duty vehicle fleet that complies
with EPA, NHTSA, and CARB standards. FCA specifically stated their concern that EPA's
treatment of occupant safety comments was related to a perceived lack of coordination with
NHTSA.
Response to Comments on the Proposed Determination
Many of the commenters expressed concerns that they had not been afforded the procedures
required under the rules establishing the MTE, and that the procedures in any case were not in
accord with stakeholders' legitimate expectations. The main contention was that the process was
precipitate and afforded inadequate opportunity to properly evaluate and discuss the many
technical issues arising under the MTE. Commenters pointed to various preamble statements
from the 2012 final rulemaking (FRM) regarding the need for an iterative, data-driven process,
other preamble statements indicating that EPA and NHTSA intended to act on concurrent time
frames, EPA statements, including web postings, suggesting plans for a lengthier process, all of
which commenters view as inconsistent with EPA's process. The Alliance stated that the
process has "precluded consideration by EPA of pending studies and more current information."
EPA has followed and complied with all of the procedural steps set out in the rules. In the
Notice of Intent describing the second phase of the National Program, which describes the
midterm evaluation process, EPA indicated (see 76 FR 48672-673, Aug. 9, 2011) that:
• EPA would conduct a mid-term evaluation of the MY2022-2025 standards to
determine whether those standards are appropriate under section 202(a) of the act, and
must make a final determination no later than April 1, 2018;
• EPA, NHTSA, and CARB would jointly prepare a Draft Technical Assessment Report
(TAR) to inform EPA's determination, and there would be an opportunity for public
comment on the Draft TAR, and appropriate peer review of its underlying analyses;
and that all assumptions and modeling underlying the Draft TAR would be available
for public comment;
• EPA would also seek public comment on whether the standards are appropriate, and
would carefully consider and respond to those comments in taking any final action;
• EPA and NHTSA would consult and coordinate in developing EPA's determination of
whether the MY2022-2025 standards are appropriate;
• EPA's determination is to be based on a comprehensive, integrated assessment of all
the results of its review, as well as any public comments received during the
evaluation, taken as a whole; the Administrator is to consider a record at least as
robust as that in the rulemaking establishing the standards;
6
-------
• An EPA decision that the MY2022-2025 standards are appropriate would be final
agency action subject to judicial review, and EPA would announce that final decision
and the basis for EPA's decision; however, if EPA determines that the standards are
not appropriate, EPA must initiate rulemaking to amend the standards.
More specifically, the codified rules on the MTE require that EPA complete the following
tasks prior to its final determination: prepare the Draft TAR; seek public comment on the Draft
TAR; and seek public comment on whether the MY2022-2025 standards are appropriate under
section 202(a). See 40 CFR section 86.1818-12(h). The time frame set forth in that rule specified
that the Draft TAR be completed no later than November 15, 2017, and that the Final
Determination be completed no later than April 1, 2018, or a period of only four and one-half
months after the Draft TAR.
EPA has adhered to all of these requirements. The agencies (EPA, NHTSA, and CARB)
prepared the Draft TAR, and sought and received substantial public comment thereon. EPA also
considered all late comments on the Draft TAR. EPA carefully considered and responded in
detail to all of the significant public comments as part of the record for the Proposed
Determination. Part of the response was to make a number of changes urged by commenters.
These included updating the baseline fleet to a MY2015 basis, better accounting for certain
technologies in that baseline fleet, improving the vehicle classification structure to improve the
resolution of cost-effectiveness estimates applied in the OMEGA model, updating effectiveness
estimates for certain advanced transmission technologies, conducting additional sensitivity
analyses (including those where certain advanced technologies are artificially constrained), and
adding quality assurance checks of technology effectiveness into the ALPHA and Lumped
Parameter Model. See Proposed Determination Appendix A at A-l and A-2. EPA consulted
with NHTSA and CARB as part of the process of developing the Proposed Determination. The
Final Determination is based on an administrative record at the very least as robust as that for the
2012 FRM, including extensive state-of-the-art research projects conducted by EPA and
consultants to both agencies, data and input from stakeholders, multiple rounds of public
comment, information from technical conferences, published literature, and studies published by
various organizations. EPA put primary emphasis on the many peer-reviewed studies, as well as
on the National Academy of Sciences 2015 report on fuel economy technologies.
EPA has considered those comments that contend that the process the Administrator has
followed with the Proposed Determination, especially regarding opportunities for stakeholders to
provide meaningful public comment, is not in accord with the stakeholders' legitimate
expectations. EPA believes that the comment period for the Proposed Determination is sufficient
in light of the limited new data and information presented in that document as well as in the
comments we received on the Draft TAR (which formed the technical underpinnings of the
Proposed Determination). The Administrator has moved forward with the Proposed
Determination based on an extensive technical record developed over several years of research,
analysis, and public input, with the recognition that lead time and regulatory certainty are critical
to the auto industry. Regarding pending industry studies that the Alliance believes could have
improved EPA's analysis, having considered extensive input from industry and other sources and
improved the quality of our technical understanding over several years, the sum of all of this
information has reinforced our fundamental conclusions from the 2012 final rule that the
standards are feasible and appropriate, and has even provided evidence to support the potential
strengthening of the standards. The Administrator believes that the likelihood that new,
7
-------
unforeseen data or information of sufficient consequence to alter this determination might come
to light in the near future is very small, and has concluded that the existing record fully supports
a decision to move forward with the Final Determination.
Several commenters maintained that the Final Determination is a rulemaking, and therefore
EPA must follow the rulemaking procedures in the Administrative Procedure Act ("APA"), or in
section 307(d) of the Act, or both. These comments are mistaken. As noted in the Proposed and
Final Determinations, this action is not a rulemaking. Rulemaking procedures are not legally
required, and EPA has properly exercised its discretion to proceed by adjudication.
None of the EPA's own rules, nor the APA or the Clean Air Act (CAA), legally require the
determination be made by rulemaking. First, EPA's own rules do not require rulemaking, unless
EPA acts "to revise the standards" upon finding the existing standards "not appropriate." See 40
C.F.R. 86.1818-12(h). But here, EPA is finding the existing standards appropriate. Had EPA
instead found the existing standards not appropriate, it would have initiated a rulemaking to
revise them. See 77 FR 62784 (Oct. 15, 2012) (stating that if EPA concludes the standards are
appropriate it will "announce that final decision and the basis for EPA's decision" and if EPA
decides the standards are not appropriate, it will "initiate a rulemaking to adopt standards that are
appropriate under section 202(a)").
Second, the APA does not require rulemaking. An APA rulemaking is defined as
"formulating, amending, or repealing a rule." Kennecott Utah Copper Corp. v. United States
DOI, 88 F.3d 1191, 1208 (1996) (citing 5 U.S.C. § 551(5)). By contrast, an agency's decision not
to revise an existing rule after consideration of new information is not a rulemaking. See
National Mining Ass'n v. MSHA ("NMA"), 599 F.3d 662, 670-71 (D.C. Cir. 2010); ICORE v.
FCC, 985 F.2d 1075, 1082 (D.C. Cir. 1993). Here, as mNMA and ICORE, EPA considered new
information and chose not to revise the existing standards. Thus, EPA was not required to, and
did not, engage in rulemaking pursuant to the APA.
Third, the CAA does not require rulemaking. CAA § 307(d)(l)(K) imposes certain
rulemaking procedures on the "promulgation or revision of regulations under section [202] and
test procedures for new motor vehicles or engines under section [206], and the revision of a
standard under section [202(a)(3)]." CAA § 202(a) also directs EPA to prescribe "by regulation"
motor vehicle emission standards. But these directives are inapposite, because EPA is not
promulgating a new emission standard (or test procedure) or revising an existing standard.
Instead, EPA has decided not to revise an existing standard.
In the absence of any statutory or regulatory requirement to conduct rulemaking, "the choice
between rulemaking and adjudication lies in the first instance within the agency's discretion."
POM Wonderful, LLC v. FTC, 777 F.3d 478, 497 (2015) (citing NLRB v. Bell Aerospace Co.,
416 U.S. 267, 294 (1974)); see also Ark. Power & Light Co. v. Interstate Commerce Com., 725
F.2d 716, 723 (1984) ("court will compel an agency to institute rulemaking proceedings only in
extremely rare instances"); Richard J. Pierce, Administrative Law Treatise 503 (5th ed. 2010)
("On the federal level, and outside the unusual context of statutorily mandated exclusive reliance
on rulemaking, all [judicial challenges to compel rulemaking over adjudication] have failed.").
EPA has exercised its discretion to proceed by adjudication. The agency believes that doing so
here is especially suitable for several independent reasons.
8
-------
Here, EPA is not promulgating any new policies or standards. Rather, EPA has chosen to not
revise its existing standards, after undertaking the process set forth in an existing rule. See 40
C.F.R. § 86.1818-12(h). Applying that rule's processes, EPA evaluated the factual record
concerning existing standards, considering technical factors such as practicability, feasibility,
technology effectiveness, impacts on the automobile industry and consumers, and safety. See id.
§ 86.1818-12(h)(l)(i)-(viii) (listing the factors). In order to do so, EPA compiled a thorough
record, see id. § 86.1818-12(h)(2), and issued a Draft Technical Assessment Report, see id. §
86.1818-12(h)(3). And in this Final Determination, EPA has "set forth in detail the basis for the
determination," id. § 86.1818-12(h)(4), and deemed the existing standards "appropriate," id. §
86.1818-12(h). Agencies regularly evaluate factual records through adjudication. See POM
Wonderful, 777 F.3d at 497-98; Safari Club Int'l v. Jewell, 2016 U.S. Dist. LEXIS 136235, at
*33-34 (D.D.C. Sept. 30, 2016); cf. also Shays v. FEC, 528 F.3d 914, 930 (D.C. Cir. 2008)
(ruling that an agency "has authority to flesh out its rules through adjudications and advisory
opinions" (citing Shalala v. Guernsey Mem'IHosp., 514 U.S. 87, 96 (1995))). Moreover, this
action has no new future effects and disturbs no reliance interests. In some cases, rulemaking
may be suitable where adjudication would unduly disturb reasonable reliance interests. See Bell
Aerospace, 416 U.S. at 295. But this action does not change the existing standards; it creates no
new rights, liabilities or rules of conduct different from those already established by the 2012
rule. Finally, EPA has historically not regarded analogous mid-course evaluations as legally
requiring rulemaking. The agency routinely conducts mid-course evaluations of its standards,
particularly those that have long lead times. In these prior cases, the agency did not find
rulemaking legally required, even though as here, the actions closely reexamined facts relating to
existing rules.3 EPA continues this practice here, further justifying its exercise of discretion. See
Drake v. FAA, 291 F.3d 59, 69 (2002) ("Where the agency's litigation position is consistent with
its past statements and actions, there is good reason for the court to defer, for then the position
seems simply to articulate an explanation of longstanding agency practice").
EPA is not persuaded by the commenters' arguments to the contrary. Global Automakers and
the Alliance commented that EPA must proceed by rulemaking because this action would have
future effect on the industry and necessarily involve policy considerations. Not so. As the D.C.
Circuit has repeatedly held, "the fact that an order rendered in an adjudication may affect agency
policy and have general prospective application does not make it rulemaking subject to APA
section 553 notice and comment." POM Wonderful, 777 F.3d at 497 (citing Conference Grp.,
LLC v. FCC, 720 F.3d 957, 966 (D.C. Cir. 2013)). Indeed, "adjudicated cases may and do serve
as vehicles for the formulation of agency policies, which are applied and announced therein, and
. . . such cases generally provide a guide to action that the agency may be expected to take in
future cases." Bell Aerospace, 416 U.S. at 294 (citing NLRB v. Wyman-Gordon Co., 394 U.S.
3 For example, in the final rule for heavy-duty engine standards (66 FR 5063, January 18, 2001), EPA announced
regular biennial reviews of the status of the key emission control technology. EPA subsequently issued those
reviews in 2002 and 2004, without going through rulemaking. See EPA Report 420-R-02-016; EPA Report 420-
R-04-004. Or for instance, in the final rule for the Nonroad Tier 3 standards (63 FR 56983, Oct 23, 1998), EPA
committed to reviewing the feasibility of the standards by 2001 and to adjust them by rulemaking if necessary. In
2001, without engaging in rulemaking, EPA published a report (see EPA Report 420-R-01-052) accepted
comments, and concluded in a memorandum placed in the docket that the standards remained technologically
feasible (Memorandum: "Comments On Nonroad Diesel Emissions Standards: Staff Technical Paper," from Chet
France to Margo Oge, June 4, 2002).
9
-------
759, 765-66 (1969)). And an agency may in its "very broad discretion" use adjudication to
formulate orders broadly applicable to an entire industry. Qwest Servs. Corp. v. FCC, 509 F.3d
531, 536 (D C. Cir. 2007).
EPA agrees that this action, like virtually any administrative action, may implicate some
policy considerations relevant to the industry. The determination that the existing standards are
appropriate, however, does not alter agency policy; the policies of the 2012 rule remain in place.
And as already noted, this action does not change the existing legal rights and obligations of
regulated parties.4
The Alliance additionally commented that EPA's decision to provide public notice and
comment necessarily transforms this action into a rulemaking. The Alliance cites no legal
authority for this claim, and EPA is unaware of any. To the contrary, "[ajgencies are free to grant
additional procedural rights in the exercise of their discretion." Vt. Yankee Nuclear Power Corp.
v. Nat. Res. Def. Council, Inc., 435 U.S. 519, 524 (1978). EPA is thus free to proceed by
adjudication with enhanced procedures, such as notice and opportunity to comment.
The Alliance further commented that EPA must proceed by rulemaking because EPA has
stated that the authority for the MTE is found in CAA 202(a), and because EPA has "reopened"
the prior rule and its record, citing cases like General Motors Corp. v. EPA, 363 F.3d 442, 449-
50 (D.C. Cir. 2004) and National Mining Association v. Department of Interior, 70 F.3d 1345,
1352 (D.C. Cir. 1995). In the alternative, the Alliance argues further that the Proposed (and now
Final) Determination constitutes a reconsideration of the MY2022-2025 standards under section
307(d)(7)(B) of the Act.
These arguments fundamentally mistake the nature of the MTE. EPA established the MTE
process by regulation. See 40 CFR 86.1818-12(h). EPA continues to believe the authority for
that regulation derives from the authority to establish appropriate standards pursuant to CAA
section 202. See 77 FR at 62786. That regulation requires the Administrator to make a
determination whether the MY2022-2025 standards are appropriate, after an opportunity for
public comment. See 40 CFR 86.1818-12(h). The regulation further requires that //the
Administrator determines the standards are not appropriate, then the Administrator will initiate a
rulemaking to revise the standards. Id.
Thus, the Final Determination is not "reopening" or "reconsidering" the 2012 rule. Rather,
the Administrator is undertaking an examination of the factual record currently before her
pursuant to the 2012 rule, and that rule does not require any further rulemaking when she
determines that the standards are appropriate. In fact, the commenter acknowledges that in
promulgating the 2012 rule EPA rejected the argument that the MTE would constitute a
"reconsideration" of the rule under CAA 307(d). See 77 FR at 62786. Section 307(d)(7)(B)
applies to situations where EPA is required to reconsider a rule on the basis of new information
raised by a petitioner to the agency which information could not have been available during the
rulemaking. Here EPA is carrying out the provisions of the rule, by assessing the
appropriateness of the standards. Section 307(d)(7)(B) is entirely inapplicable in these
circumstances.
4 As noted above, although not relevant here, agencies are generally permitted to change policy, with prospective
effect on regulated entities, through adjudication.
10
-------
Moreover, the fact that the agency reviewed new factual information as part of the
determination does not itself trigger any requirement to undertake rulemaking. To the contrary,
where the agency decides not to revise its existing standards, it need not proceed by rulemaking,
even if it considers new information. See 7VM4, 599 F.3d at 670-71; ICORE, 985 F.2d at 1082. In
both NMA and ICORE, as here, the agency expressly considered substantial new information in
issuing the challenged determinations. Nonetheless, in both cases, the court upheld the agency's
choice to not proceed by rulemaking. SeeNMA, 599 F.3d at 670-71; ICORE, 985 F.2d at 1082.
It is also worth noting that the case law on "reopening" rules is further inapplicable because it
simply addresses when a court may consider a challenge to a long-standing rule that would
otherwise be time-barred. Relatedly, Global Automakers commented that EPA must institute
rulemaking because of the "substance of what the [agency] has purported to do and has done,"
citing Center for Auto Safety v. National Highway Traffic Safety Administration, 710 F. 2d 842,
846 (1983).
These authorities are inapposite. They address whether a court has jurisdiction to review
agency action, not whether an agency must proceed by rulemaking or adjudication.5 Here, EPA
has stated that the Final Determination is a reviewable, final agency action. See 77 FR at 62784.
The question here is not whether a court would have power to hear a petition for review, but
whether EPA is required to follow rulemaking procedures. The authorities cited by commenters
do not address this issue.
Commenters also stated that the Proposed and Final Determination should have been
submitted to the Office of Management and Budget pursuant to Executive Order (EO) 12866,
arguing that section 6 of that Executive Order requires review of "significant regulatory
action[s]," meaning actions "likely to result in a rule that may have an annual effect on the
economy of $100 million or more." The Final Determination is not an action subject to EO
12866 review and has no economic effect. It determines that standards previously promulgated
in a final rule (which was subject to review per the Executive Order) remain appropriate, and
leaves the current regulatory status-quo unaltered.
5 Moreover, as noted above, EPA has not reopened the 2012 rule—it has fulfilled its obligations under the 2012 rule
and concluded that the standards are appropriate.
11
-------
Chapter 2: Technology Assessment
2.1 Effectiveness Assessment: General Comments, Technology Packages, Penetrations,
and Sufficiency of Non-Electrified Technologies
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Some comments received on the Draft TAR were critical of EPA's assessment of technology
effectiveness and its compliance projections, while others were supportive. Upon examination of
specific comments to this effect, both broadly and with respect to specific technologies examined
throughout the Proposed Determination and the Technical Support Document (TSD), EPA
concluded in its assessment that the effectiveness values developed for the Draft TAR were
largely accurate representations of benefits achievable by manufacturers within the MY2022-
2025 time frame. EPA also noted that this was not to state or imply that every manufacturer that
had added a technology had already achieved the effectiveness estimated in the Draft TAR.
Some technologies that are currently in their first or second design iteration may improve in
effectiveness in successive iterations. One example provided was the emerging use of integrated
and cooled exhaust manifolds and the resulting improved effectiveness from turbo-charged
downsized engines. Additionally, we noted that some manufacturers that have adopted
technology may have used some of the benefit to improve other vehicle attributes, rather than
solely to improve fuel economy;6 but when these technologies are combined with the sole intent
of improving vehicle efficiency, our analyses continue to show that significant improvements
from the baseline fleets are broadly achievable using conventional powertrains (see Section 2.1
of the TSD at p. 2-1 to 2-2).
Some auto industry commenters stated generally that the EPA models and/or effectiveness
assumptions are overly optimistic, while other commenters recommended higher technology
effectiveness values than we estimated in the Draft TAR. In some cases, the commenters either
did not provide any supporting evidence, or provided evidence that was incomplete, not
applicable, or not relevant to an assessment of the cost, effectiveness, and implementation
feasibility in MYs 2022-2025. In particular, the conclusion drawn by the Alliance of Automobile
Manufacturers that "MY2021 and MY2025 targets cannot be met with the suite of technologies
at the deployment rates projected by the Agencies in the 2012 FRM" is based on the premise that
the only possible technology available in MY2025 will be represented by technology already
contained in the Draft TAR's MY2014 baseline fleet, and that technology will not improve in
efficiency. See TSD App. A. In response, EPA disagreed with this assertion, noting that it is not
plausible that the best gasoline powertrain efficiencies of today represent the limit of achievable
efficiencies in the future. Id. at A.l. Even setting aside the assumption that the best available
technologies today will undergo no improvement in future years (a premise the auto industry has
disproved time and again), the methodology used in the Alliance-contracted study (which was
not peer reviewed) does not even allow for the recombination of existing technologies, and thus
severely and unduly limits potential effectiveness increases obtainable by MY2025. Id. at A.2.
Further, EPA disagreed with this assumption that the only technology combinations available in
MY2025 are those that are present in the MY2014 fleet. EPA noted that events had already
6 For example, the DeFour Group analysis cited by the Alliance in its comments on the Proposed Determination
alluded to manufacturers of strong hybrids allocating fuel efficiency gains to improved performance or towing
capacity rather than fuel economy (DeFour Group attachment to the Alliance comments, p. 14).
12
-------
disproven this assumption and provided, as one specific example, a Ford-introduced 10-speed
automatic transmission on the MY2017 F150 paired with a turbocharged downsized engine
which represents a technology combination that was not previously available and was therefore
not considered (and would be deemed impossible) by the Alliance-contracted study. Id. In
contrast, EPA's Proposed Determination projections of effectiveness through MY2025 included
technology packages that are achievable and cost-effective, but did not exist in the MY2014
fleet. For example, a 24 bar turbocharged downsized engine with cooled EGR, or a high
compression ratio Atkinson cycle engine with cylinder deactivation and cooled EGR paired with
an efficient high speed, high efficiency, high ratio spread transmission. EPA's approach for
evaluating technology effectiveness was and still is based on detailed data for individual
technologies and physics-based vehicle modeling of combinations of technologies. In the
Proposed Determination, EPA stated its assessment that these particular comments by the
Alliance with respect to future technology effectiveness were drawn from an approach that was
overly simplistic, lacked rigor, and therefore did not call into question EPA's determination that
the technology assessment supported the Proposed Determination that the MY2022-2025
standards remain appropriate. EPA's detailed response to the Alliance-contracted study is found
in the TSD (Chapter 2.3.3 and Appendix A).
In comments on the Draft TAR, several commenters, including many NGOs, state and local
government organizations, and consumer groups, supported EPA's assessment in the Draft TAR
as a robust assessment of technology availability showing multiple cost-effective paths
(compliance paths more cost-effective than those considered by the agencies in the 2012 FRM)
to comply with the 2025 standards. Some groups believed our assessment to be overly
conservative; for example, the International Council on Clean Transportation (ICCT) expressed
the view that there are some key areas where the Draft TAR analysis "is still somewhat behind
what is already happening in the market," and the American Council for an Energy-Efficient
Economy (ACEEE) stated that additional technology options are "developing rapidly and are
likely to result in multiple options at least as cost effective as those represented in the agencies'
analysis." Both ICCT and ACEEE cited examples of technologies that EPA did not model, like
e-boost, variable compression ratio, and dynamic cylinder deactivation, which they stated are
currently undergoing active development and are likely to contribute to cost-effective paths for
compliance in the MY2022-2025 time frame.
Regarding the Draft TAR's estimated penetration rates of electrified vehicle technologies, the
Alliance, Global Automakers, and several individual automakers commented that more strong
hybrids and electric vehicles would be needed to achieve the standards (MY2025 in particular)
than projected by the agencies. This is the corollary to the comment summarized above that EPA
was overly optimistic in assessing efficiencies and availability of advanced gasoline engine and
other technologies. As described above, EPA responded that the premise underlying this
comment was unfounded, undocumented, and already inconsistent with market developments.
Thus, EPA's initial response on the issue of amount of electrification needed to comply with the
MY2022-2025 standards continued to be that the standards are achievable using minimal
amounts of strong hybrid and all-electric vehicles.
Building on their premise that more electrification would be needed (which EPA did not
accept), the various auto industry commenters went on to state that sales of hybrid (HEV) and
plug-in electric vehicles (PEV) have fallen due to current low gasoline prices. With gasoline
prices not expected to rise rapidly in the time frame of the Midterm Evaluation, they were
13
-------
concerned that they will not be able to sell the vehicles they assert to be needed to meet the
standards. In contrast, comments by Tesla Motors, the International Council on Clean
Transportation (ICCT), Nextgen Climate America, Consumer Federation of America (CFA), and
Faraday Future suggested that consumer acceptance of electrified vehicles is rising rapidly,
especially with longer-range PEVs becoming less expensive. Tesla suggested that EPA should
increase the stringency of the standards to encourage both advanced gasoline technologies and
PEVs. Faraday Future and Consumer Federation of America cited survey evidence that interest is
growing in PEVs, especially among young people. ICCT pointed out that the prospects for PEVs
have improved in recent years, and that many companies are deploying this technology. Nextgen
Climate America said that PEVs can offer greater benefits than assumed in the Draft TAR. The
National Association of Clean Air Agencies also commented, pointing to rapid growth in sales of
hybrid and electric vehicles in the states that have adopted California's Zero Emission Vehicle
program, as well as other states. Given that EPA identified multiple compliance pathways, all
only minimally dependent on use of PEVs, EPA did not consider this debate as weighing
significantly on the subject of the Proposed Determination, viz. whether the standards remain
appropriate (See Section B. 1.5.2 of the Proposed Determination Appendix and Sections B.1.5
and C. 1.2 of the Proposed Determination Appendix).
OEM commenters also aimed criticism at differences in projected penetrations of individual
technologies between the Draft TAR and the 2012 FRM, characterizing these differences as
evidence that the agencies' analysis approach was unsound. In response, EPA pointed out that
these differences are not evidence of a flawed analysis but are a natural result of the Draft TAR
having recognized and included innovations and improved efficiencies that occurred since the
2012 FRM, the very sorts of improvements that the Alliance contractor report assumed would
not occur between now and 2025. See Proposed Determination at p. 24. In addition, the
technologies reflected in the Draft TAR and Proposed Determination analyses reflect many of
the technology changes that have been introduced in the fleet since the 2012 FRM. Thus, EPA
would be remiss to not consider these technologies within the context of the MTE. The 2015
NAS report also recognized these important emerging and changing technologies, such as
Atkinson cycle engines and CVTs, and recommended that the agencies consider these
technologies in their future analyses. EPA thus disagreed that such differences in projected
technology penetrations indicate in any way that the analysis and analytic approach were
unsound. On the contrary, the incorporation of new technologies and unforeseen applications
since the 2012 FRM would necessarily influence the cost-effective pathway modeled by EPA.
Id. at p. 25. For example, the application of direct injection Atkinson cycle engines in non-
hybrids, greater penetration of continuously variable transmissions (CVTs), and 48-volt mild
hybridization have all influenced projected technology penetrations, as have developments in
downsized turbo-charged engines, cylinder deactivation, and electrification. EPA also noted the
consistently low level of strong electrification projected in the 2010 TAR, 2011 NPRM, 2012
FRM, and 2016 Draft TAR, as further corroborated by the 2015 National Academy of Sciences
(NAS) study. This consistency has persisted even as EPA's technology assessment and
compliance analysis has undergone many updates and improved in its precision over the past six
years, further supporting EPA's determination that the MY2022-2025 standards remain
appropriate. Id.
Commenters on the Draft TAR also asserted that differences between the 2012 FRM and
Draft TAR with respect to the technologies considered and their projected penetrations suggest
14
-------
that the analyses were flawed. For example, the Global Automakers and its members commented
that "the agencies should investigate and document why their previous predictions (from the
FRM) were inaccurate." EPA responded that, in fact, if the differences were inaccuracies, they
only represented the failure to anticipate the success with which the industry has innovated to
increase efficiencies in the intervening years between the FRM and Draft TAR. EPA did not
agree that variations in modeled technology penetrations from the FRM to the Draft TAR were
an indication that the analysis and analytic approach were unsound. EPA further pointed out that
incorporating new technologies and unforeseen applications that had emerged since the 2012
FRM would be expected to have an impact on the penetrations of technologies in the cost-
effective pathway modeled by OMEGA. EPA cited examples such as the application of direct
injection Atkinson cycle engines in non-hybrids, greater penetration of continuously variable
transmissions (CVT), and 48-volt mild hybridization which would all tend to influence projected
technology penetrations. EPA also noted the consistency with which only low levels of strong
electrification were projected in the 2010 TAR, the 2011 NPRM, the 2012 FRM, the 2016 TAR
and the Proposed Determination as evidence that the analyses were robust, and further cited the
2015 National Academy of Sciences (NAS) study which also found that the 2025 standards
would be achieved largely through improvements to gasoline technologies without extensive
electrification.
Regarding the projected penetration of higher compression ratio, naturally aspirated gasoline
engines (Atkinson 2), the Alliance stated in their comments on the Draft TAR that they did not
believe that the projected market penetration of Atkinson 2 technology (at over 40 percent) was
likely or feasible. EPA noted (in part) that the Proposed Determination analysis projected a
reduced penetration of Atkinson 2 (at 27 percent), and that this reduction was the result of
refinements in EPA's effectiveness modeling that better reflect the relative improvements
allocated to advanced engines and transmissions in powertrain packages. See for example
Section IV.A.3 of the Proposed Determination at p. 39, and Section C. 1.1.3.2 of the Proposed
Determination Appendix at p. A-132. EPA also presented sensitivity analyses, one of which
artificially constrained Atkinson 2 technology to 10 percent penetration. This sensitivity
demonstrated that cost-effective compliance paths using primarily other advanced gasoline
engine technologies continue to exist even under this scenario, at only modestly increased costs
(see Section C. 1.2.1.4 of the Proposed Determination Appendix at p. A-144 and p. A-147).
Significantly, even those increased cost estimates remain lower than the agencies projected in the
2012 FRM, which the agencies have already evaluated as being reasonable. EPA provided
rationale for the feasibility of Atkinson 2 including responses to lead time arguments in Section
A.2.3.1 of the Proposed Determination Appendix at p. A-7 and in Chapter 2.3.4.1.8.3 of the TSD
at p. 2-308 to 2-311. Comments relating to lead time for deployment of the Atkinson 2
technology are also discussed in Chapters 2.5.1 and 4.3 of this RTC document.
Summary and Response to Comments on the Proposed Determination
In comments on the Proposed Determination, many NGOs repeated their disagreement with
the prevailing stance of many of the auto industry commenters that the standards are not
achievable with advanced gasoline technologies and would require much higher levels of
electrification than EPA projects. For example, ICCT supported EPA's Proposed Determination
but continues to believe that EPA's analysis utilized conservative assumptions for the cost and
effectiveness of many technologies. In addition, the Environmental Defense Fund (EDF)
commented that more stringent standards for MY2022-2025 are feasible, and shared an analysis
15
-------
to support this view. In order to assess the cost and technology penetration implications of
setting more stringent standards, EDF's analysis, conducted using EPA's OMEGA model,
includes four scenarios that are 10, 20, 30, and 40 g/mi more stringent than the current MY2025
target of 173 g/mi. EDF noted that the 20 and 30 g/mi more-stringent scenarios demonstrated
that the standards could be met cost-effectively with the same advanced technology pathways
projected to be utilized in EPA's analysis of the existing MY2022-2025 standards, and with very
low levels of strong hybrids and electric vehicles. This analysis also indicated that the lifetime
fuel savings benefits to consumers (assuming AEO 2016 reference case fuel prices, as in EPA's
analysis) would more than outweigh the projected increase in vehicle cost. EPA appreciates this
informative analysis.
In contrast, some industry commenters repeated the suggestion that variations in projected
technology penetrations were evidence that the analyses were unsound.
Global Automakers, The Alliance, and Toyota stated that it was unclear how EPA arrived at
significant changes in technology penetrations between the two analyses, with specific reference
to the reduction in projected Atkinson 2 engine penetration (from 44 percent to 27 percent).
In response, as mentioned above, EPA noted in the Proposed Determination that this
reduction was in part the result of refinements in EPA's effectiveness modeling that better reflect
the relative improvements allocated to advanced engines and transmissions in powertrain
packages. Another factor was the adoption of a modeled increase in engine displacement of 5
percent to ensure that acceleration performance is not degraded due to knock protection
measures when using regular grade gasoline. This change in EPA's assessment for Atkinson
technology was in direct response to comments received on the Draft TAR. (See for example
Section IV. A.3 of the Proposed Determination at p. 39, Section C. 1.1.3.2 of the Proposed
Determination Appendix at p. A-132, and Chapter 2.3.4.1.8.1 of the Technical Support
Document at p. 2-298). The change is also discussed in Chapter 2.8 in the discussion of
OMEGA outputs where more context is provided.
BMW stated that EPA underestimated the current penetration of advanced powertrain and
lightweighting in its fleet, saying that BMW has already included these technologies in current
vehicles, and that EPA overestimates the potential for further improvements, leading to the need
for higher levels of electrification, especially in light of lower fuel prices than anticipated in the
2012 final rule. BMW's general comment is substantively the same as the comments on the need
for greater electrification that the Alliance and other manufacturers made on the Draft TAR, and
our response to those comments in the Proposed Determination applies to these comments as
well. Regarding the specific example BMW gives of their current technology offerings, we
believe we have accurately incorporated BMW's situation into our fleetwide modeling.
Global Automakers referred to the changes in Atkinson 2 penetration and several individual
manufacturer's technology costs from the Draft TAR to the Proposed Determination as evidence
of general volatility in EPA's model, positing that EPA had made "significant revisions in the
course of a few months" and "[tjhese radical changes from one analysis to the other belie the
claim ... that there was a 'Robust Technical Analysis'..." EPA disagrees, noting that contrary to
the assertion of modeling volatility, EPA's assumptions of technology package cost-
effectiveness considered in the OMEGA model have remained highly stable between the Draft
TAR and Proposed Determination assessments.
16
-------
While the technology types considered and their projected penetrations have indeed changed
over the time span between the 2012 FRM, Draft TAR, and Proposed Determination, these
changes are largely due to the innovation of the automotive industry being reflected in our
updated analyses. Furthermore, even as these technologies and penetrations have changed, the
stability of EPA's estimated costs for complying with the standards support the conclusion that
there are several viable cost-effective alternative pathways to meeting the MY2025 standards and
that substantial levels of electrification will not be required. For example, transmission
technology is one example of how competing technologies may evolve due to innovation to
produce parallel options with little difference in cost or effectiveness. While the compliance
costs will tend to be stable due to competition between multiple similarly cost-effective
technologies, when a minor change in cost-effectiveness for one technology does occur (e.g., due
to innovation), the projected penetration of the various competing technologies can change, in
some case significantly. In the 2012 FRM, based in part on input from the auto industry and
other stakeholders, EPA's assessment was that dual clutch transmissions (DCTs) would provide
a better opportunity for significant improvements in vehicle efficiency than continuously
variable transmissions (CVTs), due to indications that CVTs demonstrated characteristics that
were unacceptable to U.S. consumers. However, since the 2012 FRM, it became clear that early
implementations of DCTs were experiencing some consumer resistance, while CVTs were
becoming well accepted in the market due to ongoing improvements (for example,
improvements to control strategies, such as the implementation of indexed shifting to simulate
the feel of a conventional automatic transmission). As a result, penetration of CVTs had become
much greater than originally expected. In addition, improvements have been made in each type
of automatic transmission such that the relative difference in efficiency between transmission
architectures is rapidly diminishing. EPA believes that the changes reflected in each of EPA's
analyses are the natural result of our representing in these analyses the continuing innovation in
the light-duty market, and are not indicative of volatility, instability, or unsoundness as some
commenters suggest.
Further, an examination of the cost-minimizing technology pathways also supports the
stability of the assessments. Figure 2-1 through Figure 2-5 below show the curves that define the
cost-minimizing technology package at each level of effectiveness ('frontier curves') for small
car, standard car, cross-over utility, sport utility, and pickup truck vehicle classes. In these
Figures, the technology cost-effectiveness estimated by EPA is shown to have generally
improved in the Draft TAR and Proposed Determination (lowered frontier curves) relative to the
FRM assessment, consistent with lower cost and/or higher effectiveness values that were
identified in some cases by EPA when considering additional technologies and updated
information in these most recent assessments. We note in these figures that conventional, non-
electrified gasoline technology packages reside primarily in the region below 45 percent
effectiveness, which is also the range of effectiveness values that will generally enable
manufacturers to achieve the 2022-2025 standards. Within this critical range of effectiveness
values, it can be seen that technology cost-effectiveness has remained within a narrow band
between the Draft TAR and Proposed Determination - a finding that directly contradicts the
commenter's assertion of "radical changes" that would call into question EPA's conclusions
regarding the sufficiency of conventional non-electrified technologies. On the contrary, the
consistency between the Draft TAR and Proposed Determination frontier curves shown in Figure
2-1 through Figure 2-5 is a direct indication of the general stability in EPA's modeling.
17
-------
FRM
Draft TAR
PD
Includes
| Strong
Electrification
Does riot
Include Strong
Electrification
25% 30% 35% 40% 45% 50% 55%
Effectiveness
Figure 2-1 Most Cost-Effective Technology Packages Considered by EPA: Vehicle Type 1, Low Power-to-
Weight, Low Road Load Vehicles (Small Car in Draft TAR) w/14 DOHC
$5000
$4000
$3000
4-1
M
O •
U
$2000
$1000
I"RM
Draft TAR
PD
Includes
j Strong
1 Electrification
Does not
I Include Strong
Electrification
$5000
$4000
$3000
¦—>
in
O -
U
$2000
$1000
25% 30% 35% W% 45% 50% 55%
Effectiveness
Figure 2-2 Most Cost-Effective Technology Packages Considered by EPA: Vehicle Type 13, Mid Power-to-
Weight, Low Road Load Vehicles (Standard Car, Vehicle Type 3 in Draft TAR) w/ V6 DOHC
18
-------
FRM
Draft TAR
PD
Includes
| Strong
Electrification
Does riot
Include Strong
Electrification
$5000
$4000
$3000
4-1
M
O ¦
U
$2000
$1000
25% 30% 35% 40% 45% 50% 55%
Effectiveness
Figure 2-3 Most Cost-Effective Technology Packages Considered by EPA: Vehicle Type 4, Low Power-to-
Weight, High Road Load Vehicles (Small MPV, Vehicle Type 7 in Draft TAR) w/14 DOHC
I"RM
Draft TAR
PD
Includes
j Strong
1 Electrification
Does not
I Include Strong
Electrification
$5000
$4000
$3000
¦—>
in
O *
U
$2000
$1000
Figure 2-4 Most Cost-Effective Technology Packages Considered by EPA: Vehicle Type 11, Mid Power-to-
Weight, High Road Load Vehicles (Large MPV, Vehicle Type 9 in Draft TAR) w/ V6 SOHC
25% 30% 35% W% 45% 50% 55%
Effectiveness
19
-------
~
~
Does not
Include Strong
Electrification
Includes
Strong
Electrification
TRM
Draft TAR
25% 30% 35% 40% 45% 50% 55%
Effectiveness
Figure 2-5 Most Cost-Effective Technology Packages Considered by EPA: Vehicle Type 29, Pickup Trucks
(Vehicle Type 16 in Draft TAR) w/ V8 DOHC
One of the main arguments in the comments received on the Draft TAR for why the standards
would have an adverse impact on the industry was reiterated by several manufacturers in
comments on the Proposed Determination; that the standards, although achievable, would require
manufacturers to adopt extensive electrification, resulting in more expensive vehicles - and
emerging technologies - which commenters assert consumers will be reluctant to purchase. As in
comments on the Draft TAR, the conclusion regarding the extent of electrification required
followed logically, in the view of the commenters, from their comm ents rei terating that EPA was
overly optimistic in assessing efficiencies and availability of advanced gasoline engine and other
conventional technologies. A few manufacturers shared confidential business information
il lustrating technology walks, which show the cumulative effects of the appli cati on of various
technologies applied to a given vehicle model. However, while the technology walks provided
include some of the same advanced technologies considered by EPA, none of them included a
fuller range of conventional technologies in the combinations described in the Proposed (and
Final) Determination. Some are missing very reasonable vehicle technologies, some are missing
very reasonable engine technologies, and some are missing very reasonable transmission
technologies. Because the example technology walks supplied by the manufacturers don't
include all technologies in the appropriate combinations, and in some cases don't include the
appropriate credit values, the examples show a shortfall in achieving the MY2025 CO2 targets
(as would be expected) of about 20-40 g/mi depending on the vehicle. This resulting gap
between the EPA and manufacturer-supplied projections would be eliminated if a broader set of
the available technologies described in the Proposed and Final Determination were included in
their analysis and appropriate credit values were used.
20
-------
In response, EPA's conclusion that the standards can be achieved using relatively small
penetration rates of strong hybrid and all-electric vehicles has been reinforced in the Draft TAR
and Proposed Determination assessments with the incorporation of information from the most
recent market implementations of technologies, additional benchmarking data of recent
production vehicles, extensive reviews of the literature, and refined modeling approaches. This
conclusion is also supported by the 2015 NAS study and a number of sensitivity analyses
conducted by EPA that assumed, among other things, significantly less use of the Atkinson
engine technology. See Table ES-1 and the Proposed Determination Section IV.A.3 and
Appendix C.l. Thus, EPA's response on the issue of the sufficiency of conventional gasoline
technologies and the amount of electrification needed to comply with the MY2022-2025
standards remains that the standards are achievable using very low amounts of strong hybrid and
all-electric vehicles.
While EPA's assessment in the Proposed Determination of non-electrified technologies
reflect a number of updates since the 2012 FRM and the Draft TAR, EPA incorporated the
details regarding new technology in two different ways. Some technology was updated and fully
modeled and simulated. Other technology changes were identified as supporting of our
conclusions but not fully simulated. For example, while EPA cited information that was
published by manufacturers for several new highly efficient engines that had recently entered
production or were production ready, these data (which included engine maps) were not included
directly in EPA's effectiveness modeling. In response to stakeholder comments on the Proposed
Determination regarding EPA's effectiveness estimates for advanced gasoline engine
technologies, EPA has utilized this publicly available information to further corroborate our
assessment regarding the sufficiency of conventional gasoline technologies by showing that
vehicles equipped with these existing gasoline engine technologies, along with improved
transmissions and road load reduction technologies, can support compliance with the MY2025
targets. The process used by EPA was very similar to that described in a paper published in
2016.7
For these technology walks, EPA first selected five production vehicles each from two full-
line manufacturers, which are representative of important vehicle classes: small car, midsize car,
cross-over utility vehicle, sport-utility vehicle, and pickup truck. The vehicle characteristics were
drawn from the MY2015 EPA Test Car List, to ensure that emissions values, test weights, and
road load coefficients were representative of actual tested vehicles without the application of any
adjustment or averaging as may be the case for certification values. The characteristics of these
tested vehicles are described in Table 2-1.
7 Kargul, J., Moskalik, A., Barba, D., Newman, K. et al., "Estimating GHG Reduction from Combinations of
Current Best-Available and Future Powertrain and Vehicle Technologies for a Midsized Car Using EPA's
ALPHA Model," SAE Technical Paper 2016-01-0910, 2016, doi:10.4271/2016-01-0910.
21
-------
Table 2-1 Technology Walks with Existing Engines: Baseline Vehicle Specifications
MY 2015 Actual Vehicles
Footprint
(sq. ft.)
ETW
(lbs.)
A Coeff.
(Ibf)
B Coeff.
(Ibf/mph)
C Coeff.
(Ibf/mph2)
Rated Horse
Power
Corolla
44.1
3125
29.834
-0.08450
0.021121
132
Camry
47.2
3500
27.232
0.04319
0.019374
178
RAV4 AWD
44.9
3875
33.417
0.07314
0.026719
176
Highlander AWD
49.0
4750
39.939
0.04131
0.030299
270
Tundra AWD
68.7
5500
37.347
0.63046
0.039122
381
Fiesta
40.8
2875
22.880
0.25500
0.019160
120
Fusion
49.0
3750
18.880
0.30750
0.015990
169
Escape AWD
45.6
4000
25.100
0.42490
0.023360
173
Explorer AWD
52.5
5000
36.190
0.84250
0.022530
290
F150 AWD
68.1
5250
31.040
0.35380
0.036860
365
Next, using the ALPHA model, EPA constructed two independent technology walks for each
of the vehicles with characteristics described in Table 2-1, using data recently published by
manufacturers for two highly efficient gasoline engines. The first technology walk series is based
on Toyota's published efficiency map for a 2.5L Atkinson cycle engine with cooled EGR.8 The
second technology walk series is based on the published map of Honda's 1.5L four-cylinder
turbo engine.9
Each technology walk begins with the MY2015 vehicle and sweeps through a series of five
different technology packages:
• The initial technology package in each technology walk includes the efficient gasoline
engine (either the Toyota or the Honda), an existing high ratio spread transmission
(the EPA-benchmarked MY2014 8-speed, or TRX21), and stop-start technology if not
already present on the baseline vehicle.
• In the next technology package, the automatic transmission was improved to reflect
future efficiency improvements (TRX22) and improved accessory loads, reduced from
390 to 290 watts consistent with the approach used in the Proposed Determination
assessment.
8 Toyota: Eiji Murase and Rio Shimizu, "Innovative Gasoline Combustion Concepts for Toyota New Global
Architecture," 25th Aachen Colloquium Automobile and Engine Technology 2016.
9 Honda: Wada, Y., Nakano, K., Mochizuki, K., and Hata, R., "Development of a New 1.5L 14 Turbocharged
Gasoline Direct Injection Engine," SAE Technical Paper 2016-01-1020, 2016, doi:10.4271/2016-01-1020.
(supplemented with data publicly available during the 2016 SAE World Congress).
22
-------
• In the final three technology packages of the technology walks, increasing levels of
aerodynamic, tire, and mass road load improvements are applied from 10 percent in
the low load reduction case to 20 percent in the high load reduction case.10
The modeled results of the technology walks are presented below. Table 2-2 shows tailpipe
CO2 values of the modeled packages for the 10 modeled vehicles, while Table 2-3 shows the
difference between these tailpipe CO2 values and the footprint target values for the MY2025
standards. Importantly, Table 2-3 shows that 6 of the 10 vehicles are able to meet or exceed their
respective MY2025 target values with only low or moderate levels of load reduction based on
the Toyota engine, while 7 of the 10 vehicles are able to meet or exceed target values with
moderate or high levels of road load reduction based on the Honda engine.
Overall, these findings corroborate EPA's conclusion that the standards are achievable with
primarily non-electrified technologies. First, the fact that roughly half of the vehicles in these
tech walks are able to generate credits in MY2025 using current engine technology and moderate
road load reduction is indicative of a favorable compliance scenario with fleet average standards
such as these, since not every vehicle in a manufacturer's fleet would need to meet its individual
footprint target. Second, due to a number of conservative assumptions made when conducting
this technology walk analysis, the opportunity for conventional technologies to contribute to
achieving the standards will likely be even greater than indicated by these results. The first of
these conservative assumptions is the effective double counting of transmission neutral-drag
losses. Specifically, since EPA had not quantified these losses for each specific vehicle, the road
load coefficients were not adjusted, resulting in an average 3 percent greater CO2 value for the
10 modeled baseline vehicles than the actual tested vehicles; an overestimation that is likely
propagated to some extent through the subsequent technology packages in each techwalk.
Additional conservative assumptions made by EPA in these techwalks include the assumption
that there will be no further improvements in engine technologies beyond these two existing
engines, and the lack of consideration of off-cycle credits beyond the stop-start credit. In reality,
a manufacturer's actual compliance opportunities will include the potential for engine
technology improvements beyond those that exist today, the potential for off-cycle credits
beyond stop-start credits, and the potential for some level of mild hybridization. Further, to the
extent manufacturers do choose to apply strong hybridization or electrification, these
technologies can provide a significant compliance benefit at even at low penetration levels.
10 Note that the moderate load reduction case includes 10 percent mass reduction and 20 percent tire and aero
improvements. In each of the low, moderate, and high load reduction cases, improvements are measured relative
the mass, aero, and tire levels assigned to the corresponding vehicles in EPA's MY2015 baseline fleet. As one
example, since EPA's PD assessment applies 12.5 percent mass reduction to the baseline F150, additional mass
reduction is not applied in this technology walk for the low and moderate load reduction cases.
23
-------
Table 2-2 Technology Walks with Existing Engines: Modeled Tailpipe CO2 Values (g CCh/mi)
Corolla
Camry
RAV4
AWD
Highlander
AWD
Tundra
AWD
Fiesta
Fusion
Escape
AWD
Explorer
AWD
F150
AWD
2025 target (g C02/mi, using 2015 footprint)
141
151
173
188
258
131
157
176
201
256
2025 Stop-start off-cycle credits (g C02/mi)
2.5
2.5
4.4
4.4
4.4
2.5
2.5
4.4
4.4
4.4
2025 AC credits (g C02/mi)
18.8
18.8
24.4
24.4
24.4
18.8
18.8
24.4
24.4
24.4
2015 actual vehicle, pertest car list
205
232
268
335
456
216
228
271
362
362
MY2015 modeled
217
239
278
337
465
228
234
285
367
364
Delta: MY2015 modeled - actual veh
6%
3%
4%
1%
2%
6%
2%
5%
1%
1%
Technology Walk#l: Published Toyota engine. 2.5L Atkinson + CEGR. scaled to vehicle size and performance neutral
+ existing 8-speed and stop-start
187
201
233
288
400
183
194
231
313
331
+ trans and accessory improvements
173
184
217
265
369
170
179
216
291
308
+ low load reduction
164
170
199
246
340
154
169
197
267
300
+ mid load reduction
153
165
190
234
329
148
164
190
258
290
+ high load reduction
146
155
179
213
302
140
154
180
243
277
Technology Walk #2: Published Honda engine, 1.5LTurbo, scaled to vehicle size and performance neutral
+ existing 8-speed and stop-start
191
210
241
295
416
184
202
241
322
346
+ trans and accessory improvements
174
194
227
274
384
172
184
226
300
324
+ low load reduction
166
179
208
249
351
159
172
203
276
316
+ mid load reduction
155
174
199
239
339
152
167
195
266
301
+ high load reduction
147
159
184
226
315
145
157
183
248
282
Table 2-3 Technology Walks with Existing Engines: Gap to MY2025 GHG Target (g C02/mi)
Corolla
Camry
RAV4
AWD
Highlander
AWD
Tundra
AWD
Fiesta
Fusion
Escape
AWD
Explorer
AWD
F150
AWD
2015 actual vehicle, per test car list
-43
-60
-66
-118
-169
-64
-50
-66
-133
-77
Technology Walk #1: Published Toyota engine, 2.5L Atkinson + CEGR, scaled to vehicle size and performance neutral
+ existing 8-speed and stop-start
-25
-29
-31
-71
-112
-31
-16
-26
-83
-46
+ trans and accessory improvements
-11
-12
-15
-48
-82
-18
-2
-12
-62
-23
+ low load reduction
-2
2
3
-29
-53
-2
8
8
-37
-15
+ mid load reduction
9
7
12
-17
-42
4
14
15
-29
-4
+ high load reduction
16
17
23
3
-14
12
24
24
-14
8
Technology Walk#2: Published Honda engine, 1.5LTurbo, scaled to vehicle size and performance neutral
+ existing 8-speed and stop-start | -29
-38
-39
-78
-129
-33
-24
-36
-92
-61
+ trans and accessory improvements j -13
-22
-25
-57
-96
-20
-6
-21
-71
-39
+ low load reduction | -4
-7
-6
-32
-64
-7
5
1
-46
-31
+ mid load reduction 7
-2
3
-22
-51
-1
11
10
-37
-16
+ high load reduction 15
13
18
-9
-27
7
21
22
-19
3
Note: Assumes the application of available AC credits and stop-start off-cycle credits shown in Table 2-2. Values in
green indicate numerical CO2 values lower than (or approaching) the footprint target GHG values.
24
-------
2.2 Effectiveness Modeling and Quality Assurance
2.2.1 ALPHA Model
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Most comments on the Draft TAR that related to the ALPHA model were directed toward
specific ALPHA inputs that EPA used for the analysis, rather than the model itself. Of the
comments relating to the model itself, EPA received positive comments from the Union of
Concerned Scientists, Environmental Defense Fund, NRDC, and others, pointing out the
importance of using a physics-based, full vehicle simulation model such as ALPHA, and
commending EPA's decision to make ALPHA and all of its inputs fully transparent and freely
available to the public. Other comments on the ALPHA model included suggestions that EPA
use the Autonomie model in place of the ALPHA model, on the grounds that industry is more
familiar with Autonomie. EPA responded (TSD p. 2-268) that commercially available tools such
as Autonomie cannot be made fully transparent and therefore are not the most suitable models to
use for regulatory purposes, where transparency and replicability are critical and highly desirable
elements. An additional comment was received regarding quality control and quality assurance
parameters that can be used to verify the validity of model results in all output files. This topic is
addressed in Chapter 5.3.3.2.3 of the Draft TAR and in the public release of the ALPHA
model.11
Comments from vehicle manufacturers regarding the effectiveness values modeled in ALPHA
for various individual technologies were addressed in the respective subsections of TSD Chapter
2.3.4. Additional comments regarding issues with EPA's engine sizing and performance were
addressed in TSD Chapter 2.3.1.2, in which EPA further explains our methodology and how this
relates to OEM product realities. Comments from industry regarding top gear gradeability were
addressed in TSD Chapter 2.3.4.2, with the additional discussion that manufacturers are not
currently maintaining top gear gradeability due to the inherent advantages of advanced
transmissions. Accessory load assumptions were also raised in the comments. TSD Chapter
2.3.3.3.6 provided a further discussion of EPA's use of multiple values of accessory load values
based on the vintage of the vehicle being modeled and how these values were derived from
actual vehicle testing.
Summary and Response to Comments on the Proposed Determination
In comments on the Proposed Determination, NRDC highlighted a key finding of the 2015
NAS Committee, stating that the Committee found that the agencies' original analysis was
"thorough and high caliber on the whole." NRDC also cited the position of a member of that
Committee, that the Draft TAR and Proposed Determination analysis is "extremely thorough and
of high caliber since its methodologies are consistent with the NAS recommendations to increase
the use of these approaches" (referring to EPA's use of full simulation modeling combined with
lumped parameter model, vehicle testing, and tear-down studies).
In reference to the peer review of the ALPHA model, the Alliance noted that the findings of
the peer review were not available until October 2016 (when the peer review was completed and
the report was published), and stated that the comment period on the Draft TAR therefore did not
11 The public release of the ALPHA model is available at: https://www.epa.gov/regulations-emissions-vehicles-and-
engines/advanced-light-duty-powertrain-and-hybrid-analysis-alpha.
25
-------
provide stakeholders an opportunity to participate in the review or examine the model. In
response, as indicated in the Draft TAR, all of the materials provided to the ALPHA model peer
reviewers had been publicly posted on the EPA website, including the fully functioning ALPHA
models (see Draft TAR p. 5-256); therefore, interested stakeholders in fact did have an
opportunity to examine the ALPHA model since May 2016 when this information was posted.
Additional fully functioning ALPHA models were publicly posted at the Draft TAR release,
including the engine maps, transmission maps, and complete vehicle information used in the
Draft TAR analysis. Further, in response to the Alliance's comment that they were not provided
an opportunity to participate in the peer review, in conducting independent peer reviews EPA
follows Science and Technology Policy Council guidelines, which specify that, "The Agency
should not be involved, however, in the selection of individual peer reviewers and should avoid
commenting on the contractor's selection of peer reviewers other than to determine whether the
reviewers, once selected, meet the qualifications established."
In reference to the Alliance suggestion in their comments on the Draft TAR that EPA add a
specific set of proposed quality control checking parameters, the Alliance suggested that EPA's
response in the Proposed Determination was unclear as to whether ALPHA calculates these
parameters, and criticized EPA's suggestion that stakeholders could modify ALPHA to add such
parameters as desired. In response, we note that ALPHA includes extensive energy auditing
measures which serve as a quality control checking mechanism. The energy auditing topic is
addressed in Chapter 5.3.3.2.3 of the Draft TAR, Chapter 2.3.3.3.3 in the TSD, and in the public
release of the ALPHA model.12
2.2.2 Lumped Parameter Model (LPM)
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Comments on the Draft TAR regarding the Lumped Parameter Model (LPM) mostly focused
on the applicability of the model and the accuracy of the efficiency results it projects.
The most detailed comments received were regarding the LPM modeling methods, powertrain
efficiency, and quality control (QC) checking. Some comments from vehicle manufacturers s
asserted that EPA's modeling methods overestimate the effectiveness of technologies at the
vehicle level, and thus underestimate the required penetration rates of advanced technology
required to meet the 2022-2025 standards. Specifically, these comments cite a study by Novation
Analytics, contracted by the Alliance and Global Automakers, and a similar study done at Oak
Ridge National Laboratory. EPA's response, found in Appendix A of the TSD, provided a
detailed analysis of the shortcomings in the Novation analysis (submitted as an Attachment by
The Alliance in their comments on the Draft TAR), and a clarification on the apparent
misinterpretation by commenters of the conclusions of the Oak Ridge study. Among the
deficiencies there noted, the Novation study assumes a priori that the MY2014 powertrain
efficiency will define maximum achievable efficiency. Among other things, this assumption
ignores possibilities of combinations of existing technology packages and subsystems, as well as
likely technological improvements. The analysis, for example, failed to account for such
technological developments that have already occurred, such as 24 bar turbocharged downsized
(TDS) engines, cooled exhaust gas recirculation (cEGR), or the high expansion ratio Atkinson
12 The public release of the ALPHA model is available at https://www.epa.gov/regulations-emissions-vehicles-and-
engines/advanced-light-duty-powertrain-and-hybrid-analysis-alpha.
26
-------
cycle engine with cEGR. These comments further stated that the LPM should incorporate "key
vehicle and powertrain parameters which determine powertrain efficiency," by which the
commenters mean accounting for the engine displacement and power in relation to the energy
expended over the test cycles. These comments were addressed in TSD Chapters 2.3.3.2 and
2.3.3.5.4, which explained how the Proposed Determination included consideration of the
powertrain efficiency metric as a quality control (QC) tool. Further comments about adopting
QC checks to determine the plausibility of results are addressed in TSD Appendix B, which
explained how EPA adopted an additional layer of QC check based on powertrain efficiency, as
suggested in the comments.
EPA also received comments questioning how the LPM accounts for the baseline efficiency
of vehicles. These comments are addressed in TSD Chapter 2.3.3.5.1, which further explains
how the baseline vehicle technologies are fully accounted for in the analysis. It should be noted
that this comment relates to identifying a proper regulatory baseline, rather than to the Lumped
Parameter Model itself. The LPM identifies incremental improvements to that baseline.
Comments received on the Draft TAR and Proposed Determination regarding the baseline fleet
are reviewed in Chapters 2.7, 2.7.1, and 2.7.2 of this RTC document.
Summary and Response to Comments on the Proposed Determination
As previously stated with respect to the ALPHA model, the NRDC comments listed the LPM
as one of the modeling approaches that the NAS Committee recommended for continued use in
combination with full vehicle simulation.
The American Council for an Energy-Efficient Economy (ACEEE) commented, "For the PD,
EPA undertook substantial additional analysis to further investigate those topics. In order to
capture variations in power train technology effectiveness, EPA i) altered its vehicle
classification to reflect variations in power-to-weight ratio and road load power and ii) used a
power-to-weight correction factor within each class to adjust the effectiveness values produced
by the lumped parameter model before those values were input to OMEGA. To address the QC
point, EPA backed out power train efficiency for a representative set of vehicles in the
compliance package and found that the resulting efficiencies were in fact reasonable. The
agency's results support the conclusion that the 2025 compliance scenario presented in the PD is
plausible."
The Alliance asserted that EPA failed to adequately document the steps taken to calibrate the
LPM, did not provide the executable version, and did not provide clear directions for use of the
spreadsheet version of the LPM. FCA reiterated its position that the LPM should be verified in
some way with real-world data, and suggested that the lack of this verification makes it more
critical that EPA should fully document the steps that were taken to calibrate the LPM.
The method for calibrating the LPM is basically unchanged from the 2012 FRM, although the
inputs used to calibrate the LPM have been continuously refined based on the latest available
data. For the 2012 FRM, the LPM was calibrated using data from multiple rounds of full vehicle
simulation from Ricardo, under contract to EPA, along with real world data and other sources
such as the National Academy of Science reports. For the Draft TAR, the ALPHA full vehicle
simulation model was introduced to provide an additional level of detail and transparency to
EPA's analyses. Transparency and underlying technical details were also increased through the
addition of engine, transmission, and vehicle benchmarking still relying on the LPM to
27
-------
differentiate vehicle types. For the Proposed Determination analysis, the ALPHA model
provides the calibration data across all vehicle types, and moves the LPM into a simpler role of
providing effectiveness values between ALPHA and OMEGA. In addition, in response to
comments from the Alliance on the Draft TAR, the Proposed Determination analysis further
differentiates individual vehicles using the particular characteristics of each baseline vehicle to
expand upon the resolution provided by the ALPHA calibration data. The calibration of the
LPM is described in Chapter 2.3.3.5.2 of the TSD. The description of the ALPHA full vehicle
simulation model along with the real-world data used to calibrate ALPHA is described in
Chapters 2.3.3 and 2.3.4 of the TSD. Over 100 ready-to-run ALPHA models used to tune the
LPM are available at the EPA ALPHA website.13 Contrary to the Alliance's comments, EPA
believes that the information provided in the TSD adequately describes the function of the LPM
as well as how the LPM output represents incremental effectiveness.
The executable version of the LPM has never been used in the OMEGA analysis and was
originally provided with the FRM as an aid for stakeholders to build OMEGA packages resulting
in effectiveness values derived from the Ricardo simulations. As the LPM has expanded in
scope since the Draft TAR, this tool would require the user to have specific knowledge to
provide specific inputs that is well beyond the original simple intent of the tool and therefore is
no longer supported. The appropriate reference files to examine OMEGA technology packages
and LPM output are contained in the master set and machine files located in the OMEGA pre-
processors. 14 These files were available at the time of the Draft TAR and Proposed
Determination releases, and contain several hundred thousand technology combinations across
all vehicle types, providing all possible technology packages considered in the OMEGA process
without any user input required.
FCA revisited comments received by EPA on the Draft TAR, originally made by Global,
regarding the ability of the LPM to predict the CO2 emissions of vehicles from the MY2014
fleet. EPA had responded in part that the LPM should not be expected to predict absolute CO2
emissions because it is not designed for that purpose.
As discussed in TSD Chapter 2.3.3.5, the LPM does not predict the absolute CO2 emissions
for specific vehicles in the baseline fleet. During preprocessing for the Proposed Determination
OMEGA analysis, the LPM used results from the ALPHA full vehicle simulation model to
estimate a specific net effectiveness value for each of the specific technology packages that
OMEGA will be using for its analysis. For each baseline vehicle in its analysis, the OMEGA
model starts with the vehicle's actual certified CO2 value and applies the net effectiveness value
for the specific technology package applied to that vehicle to arrive at an estimate of the
improved vehicle's CO2. This process has not changed since the 2012 FRM.
Docket memo EPA-HQ-OAR-2015-0827-5918 describes additional documents that were
publicly available in support of the Proposed Determination detailing inputs used in the ALPHA
full vehicle simulation model. These models and their inputs (engine maps, transmission maps,
road loads, etc.) are completely transparent for examination and further analysis by stakeholders.
13 The public release of the ALPHA model is available at https://www.epa.gov/regulations-emissions-vehicles-and-
engines/advanced-light-duty-powertrain-and-hybrid-analysis-alpha.
14 The cited materials can be found at: https://www.epa.gov/regulations-emissions-vehicles-and-
engines/optimization-model-reducing-emissions-greenhouse-gases.
28
-------
Global Automakers commented on what it describes as an error in the LPM that causes EPA's
projected CO2 savings for vehicles of different sizes to be the same when the same technology
combinations are applied, which Global characterized as evidence of an error, on the grounds
that large vehicles would be expected to show a greater CO2 reduction than smaller vehicles.
Global cited the specific example of the combination of Atkinson 1 with cylinder deactivation,
going on to say that the raw data provided as part of the Proposed Determination suggested that
this 'error' went uncorrected from the Draft TAR.
In response to the comment by Global Automakers, the Atkinson 1 engine technology is
reserved for strong hybrid applications, and cylinder deactivation is not considered as an option.
The Atkinson 2 engine technology used in advanced powertrain technology packages considers
cylinder deactivation. The master set file cited in the comments submitted
(MS_Control_in2025AB_20161118_icm_aeoR) does not contain the technology combination of
Atkinson 1 and cylinder deactivation, and therefore would not be considered in the OMEGA
analysis. EPA agrees that vehicles with a higher power-to-weight ratio will typically result in
more effectiveness for a given package of technologies. Consider the following technology
package in the same file (MS_Control_in2025AB_20161118_icm_aeoR), similar to many
vehicles in the 2015 baseline:
LUB+EFR1+LRRT1+IACC1+EP S+Aero 1 +LDB+DCP+WR5%+TRX 11
When this technology package is applied to a lower power-to-weight ratio vehicle (Type 1),
the effectiveness improvement is 20.7 percent. This same technology package applied to a
higher power-to-weight ratio vehicle (Type 15) has an effectiveness improvement of 23.7
percent. These effectiveness improvements apply to the exemplar vehicles for these vehicle
types and are further adjusted based on the characteristics of the individual baseline vehicles, as
described in TSD Chapter 2.3.3.2.
2.2.3 Quality Assurance / Plausibility Checks, ALPHA-LPM Calibration
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Some comments received on the Draft TAR were critical of the processes used by EPA to
assure the reliability and accuracy of the modeling tools. In a contracted study referenced in The
Alliance comments, Novation Analytics stated that "[N]o procedure or methodology is currently
in place to check the outcomes of the [LPM's] technology effectiveness projection process
against logical efficiency metrics and limits. Without such checks, the outcomes can exceed
plausible limits" (pg. 44, Alliance comments). In the Proposed Determination, EPA responded
that it did not agree that the processes used in the previous FRM and Draft TAR assessments did
not involve plausibility checks. The LPM had been calibrated to, and was bounded by, the
physics-based full vehicle simulation model results. It was not used to predict anything beyond
the bounds of these fundamental inputs. As described in Appendix A of the TSD of the Proposed
Determination, EPA considered each of the three metrics proposed by the Alliance and did not
find any of them to be appropriate for use as plausibility checks of technology effectiveness. At
the same time, we acknowledged that quality assurance processes are important for ensuring the
validity of any modeling, and EPA adopted the use of the powertrain efficiency metric as a
quality assurance tool for the Proposed Determination as described in TSD Chapter 2.3.3.5.4 and
Appendix B of the TSD.
29
-------
Summary and Response to Comments on the Proposed Determination
EPA received multiple comments on the discussion in Appendix A of the TSD which
addressed the technology effectiveness studies undertaken by the Alliance's contractor, Novation
Analytics. The comments submitted by the Union of Concerned Scientists (UCS), the
International Council on Clean Transportation (ICCT), and the American Council for an Energy-
Efficient Economy (ACEEE) generally concur with EPA's analysis that was presented in
Appendix A of the TSD, and with the conclusion that the additional plausibility limits
recommended by Novation were not justifiable. The comments received from the Alliance of
Automobile Manufacturers (Alliance), the Association of Global Automakers (Global), Toyota,
and Novation Analytics were more critical of EPA's consideration of the Novation work in the
Proposed Determination. These commenters expressed the view that EPA did not provide
sufficient explanation in the Proposed Determination for dismissing the plausibility limits
recommended by Novation. Both Global and Novation commented that the methodologies used
in the EPA and Novation work were fundamentally the same, and that EPA had improperly
characterized Novation's methodology. These comments, while extensive, primarily offer
criticisms of EPA's analysis and conclusions in the TSD Appendix A regarding the Novation
studies and do not provide new information. They do not persuade EPA to alter our previous
conclusion to not accept the recommended plausibility limits. We address the particular concerns
raised in their comments below.
Differing Use of Key Concepts and Terminologies: The comments from Novation
Analytics and the methodologies in their earlier studies use certain critical concepts differently
from EPA, as well as in an inconsistent manner, making it difficult at times to assess or respond
to them in detail. Specifically, the comments borrow some of EPA's terminology but appear to
define or refer to certain key concepts differently compared to how EPA defined or referred to
them in the TSD.
Consistent use of conceptual terminology is vitally important in discussing and describing the
modeling analysis and results that are fundamental in the MTE. Novation expresses
disagreement with EPA's assessment in the TSD of the critical flaws in their studies. Novation
comments that their technical approach to characterizing technology package effectiveness is
fundamentally the same as EPA's, and "by criticizing Novation, EPA is calling into question its
own approach."
EPA disagrees that its modeling approach is the same as Novation's, and believes instead that
the differing conclusions of the Novation study are largely premised on a misuse of key concepts
and terms as applied in engineering models of vehicle operation. In particular, the concepts in the
Novation materials represented by the terms "maps" and "full-vehicle simulation" are
significantly different from the same language and terms used in EPA's technical assessments,
leading to divergent results.
Novation uses the term "map" to mean a representation of the efficiency of a powertrain type
(defined as the ratio of vehicle tractive energy to the fuel energy used) over a test cycle as a
function of displacement specific operating load (a measure of powertrain sizing). EPA uses a
variety of maps in its modeling process, but most commonly EPA uses the term "map" to refer to
a representation of the efficiency of an engine (defined as the ratio of engine work out to fuel
energy in) as a function of operating speed and load applied, with subsequent accounting in the
full-vehicle simulation for the interaction with other components (including, critically, the
30
-------
transmission). Novation has explicitly stated in public descriptions of the ENERGY software
used in its contracted work for the Alliance that the model does not include engine fuel maps,
transmission shift strategies, or alternator maps that would be necessary to model technology
combinations that are not currently available for physical testing. As the term is used by
Novation, powertrain efficiency "maps" are unable to model component level improvements and
alternative component combinations beyond those that exist today, unless some adjustments are
made to incorporate input from component maps of the type used in EPA's full vehicle
simulation to account for future engine, transmission, and other component technology. Novation
has provided no indication that the Alliance-contracted work employs these component-level
"maps" as defined and used by EPA in its TSD.
Novation uses the term "full-vehicle simulation" to mean the combination of a time-step road
load simulation with a powertrain efficiency map to estimate the fuel energy consumed over a
test cycle. Novation states that its road load analysis "does not impose arbitrary constraints, and
thus the simulation enables the adoption of future levels of road load improvements that may not
exist in the fleet currently." However, as explained above, the constraints imposed by Novation's
limited use of powertrain efficiency "maps" preclude the consideration of technology
improvements beyond the components and combinations that exist currently, and as a result, the
results of Novation's full-vehicle simulation are predisposed to be artificially over-constrained.
EPA uses the term "full-vehicle simulation" to mean a time-step simulation of engine,
transmission, and accessory component maps, together with component interaction models and a
road load model to estimate the fuel energy consumed over a test cycle. By modeling individual
components and their interactions, EPA has applied the ALPHA full-vehicle simulation to model
both vehicle-level performance of technology packages that exist today, as well as those that are
expected to be available in the future.
In summary, EPA believes it is inappropriate to replace its analysis of the future fleet of
vehicles with an analysis limited by the constrained modeling concepts implemented by
Novation, as described above, because such an approach inherently would produce results that
ignore both appropriate recombinations of current technologies and any future development, and
show powertrain efficiency values which are constrained by the efficiency of current production
powertrains. Even with the subsequent application of road load reductions by Novation
(constrained to the levels projected by EPA in the FRM), the analysis conducted for the Alliance
was predisposed to show a shortfall in the ability of conventional technology to meet the
MY2025 standards. However, the inconsistent meaning and use of terms and concepts in
Novation's comments make it difficult to compare their analysis to the methodology used in
EPA's work.
General Material: In comments on the TSD, Novation Analytics stated that EPA's
discussion in Appendix A of the TSD was "largely based on blogs15'16 rather than fact-checked
and peer-reviewed sources." This comment seems to imply that EPA did not perform its own
technical assessment of the Novation Analytics work. This assertion is incorrect. The discussion
of the Novation Analytics work detailed in Appendix A of the TSD consists of EPA's own
15 David Cooke, "Five Deceptive Tactics Automakers Are Using to Fight Fuel Economy Standards," July 13, 2016,
Union of Concerned Scientists, http://blog.ucsusa.org/davecooke/automakers-fuel-economy-standards.
16 Alam Baum and Dan Luria, "Why We Believe the Auto Alliance Review of Fuel Economy Standards Misses the
Mark," July 6, 2016, Ceres, https://www.ceres.org/press/blogposts/auto-alliance-review-misses-the-mark/.
31
-------
analysis of Novation's reports, which were submitted by the Alliance in support of their
comments on the Draft TAR, and previous briefings by the Alliance and Novation Analytics to
the EPA. EPA expressed the view that its position was further supported by the fact that other
parties such as UCS and Ceres independently came to similar conclusions about the Novation
reports (as expressed via their official blogs). EPA notes that comments submitted on the
Proposed Determination by UCS, ICCT, and ACEEE generally concur with the analysis that was
presented in Appendix A of the TSD.
The Association of Global Automakers commented that "[b]y using the same inputs and basic
methodology as EPA, Novation should have come to the same conclusions concerning the
technological feasibility of the MY2022-2025 standards [as EPA did]." In response, EPA agrees
that if Novation had actually used the same inputs and the same basic methodology, this would
be the expected result. However, Novation did not do so. Instead, they "performed the study
using a 'top-down' analysis, which evaluates scenarios using the overall energy conversion
efficiency of the powertrain system" in contrast to "starting with a baseline performance value
and adding percentage changes expected for a given technology as would be done in a 'bottom-
up' study" as the EPA performed."17 Thus, not only was the modeling methodology used by
Novation different from EPA's, but also the required inputs used by Novation were different. It
is clear that the Novation reports did not use "the same inputs and basic methodology as EPA" as
Global contends.
ICCT also commented on the Novation studies, stating, "[ojverall, it should be noted that a
'top-down' analysis such as that offered by Novation should cast doubt on a detailed, simulation-
based analysis such as that conducted by EPA only to the extent that the top-down analysis
demonstrates that the simulation-based approach violates fundamental principles. The Novation
report does not make any such demonstration, but rather imposes artificial constraints on how far
and how fast technology can advance." EPA agrees with ICCT's assessment.
Both UCS and ACEEE commented on the Alliance's use of Novation's study to support their
contention that "conventional powertrains will likely not displace the need for more
electrification."18 UCS and ACEEE disagreed with the Alliance's conclusion, with UCS
commenting that "[i]n fact, the [Novation] report identifies two different scenarios where
manufacturers would be able to comply with the 2025 regulations using conventionally powered
vehicles," and that "these scenarios are generally consistent with EPA's technology pathways by
deploying 24-bar turbocharged engines, stop-start, and high-ratio transmissions." ACEEE further
comments that "given Novation's failure to properly account for technology advances, [this
result] supports the conclusion that more stringent standards than those in place for MY 2022-
2025 could be achieved."
In an introduction to their comments on the Proposed Determination, Novation advances two
reasons for disagreeing with what it characterizes as EPA's main argument:
"EPA's main argument is that Novation simply assumed MY2014 technology and levels of
powertrain efficiency, making no consideration for powertrain and vehicle load technology
advancements. On the contrary, the Novation studies assumed: (1) The same powertrain
17 Fleet-Level Assessment, p. 7.
18 Alliance of Automobile Manufacturers Comments on Draft Technical Assessment Report, EPA-HQ-OAR-2015-
0827-5711, p. iii.
32
-------
technology pathways published in the FRM, which included aggressive turbocharging with
engine displacement downsizing, high efficiency and high ratio spread transmissions, stop-
start, and multiple levels of electrification. (2) The same vehicle load reductions published in
the FRM, which included aerodynamic drag and tire rolling resistance reductions of up to
20% in addition to mass reductions of up to 10%."
In response, Novation's comment both misstates EPA's responses in the TSD and confounds
different aspects of their own work. On the first point, Novation considered a number of
powertrain "technology bundles," some of which do not exist in the MY2014 fleet (for example,
27 bar turbocharged engines with high ratio spread transmissions and stop-start). However,
Novation confounds their inclusion of these technology bundles with their failure to properly
assess potential technological advancement within each technology bundle. Within each bundle,
Novation simply assumed, without providing substantiation, that the average powertrain
efficiency in the future will be tied to the efficiency distribution within the MY2014 fleet, and
improvements within a technology bundle are due strictly to "learning." In fact, there are
multiple individual sub-technologies that can be applied to a powertrain which do not change its
"bundle" as defined by Novation, but do increase the powertrain efficiency - for example,
Atkinson or Miller cycle engines. The Novation process ascribes powertrain efficiency
improvements due to the incorporation of additional technologies not as quantifiable advances,
but as progression along a statistical "learning" curve.
However, all combinations of sub-technologies do not exist in the MY2014 fleet, and thus
potential powertrain efficiency improvements exceed what is currently in the fleet. The statistical
representation used by Novation for each technology bundle, which is tied to the efficiency of
existing combinations in the MY2014 fleet, thus systematically underestimates potential future
improvements due to new technology or recombinations of technologies already included in the
powertrains within the bundle. This artificial limitation on technology improvement within each
bundle was what was noted by EPA in the TSD as a fundamental inadequacy of the Novation
study, not the existence (or lack thereof) of downsized turbocharged engines in Novation's
studies.
On the second point, Novation again misstates EPA's discussion in the TSD. EPA pointed out
that Novation did not consider changes in the penetration rate of vehicle load reductions
published in the FRM, specifically in the portion of their analysis where it would be appropriate
to do so. Moreover, Novation confounds the separate sections of their own analysis: EPA
acknowledges that when Novation attempted to evaluate EPA's projected powertrain efficiency
numbers, they appropriately maintained EPA's projected vehicle load reductions. However, this
was not the point made by EPA in the TSD. EPA noted in the TSD that Novation did not
consider changing the projected vehicle load reductions (or other non-powertrain aspects) in the
latter part of their analysis where it would be appropriate to do so. Specifically, rather than
consider possibly cost-effective decreases in road loads when evaluating "alternative technology
deployment pathways that could allow the fleet to comply with the agencies' future model year
standards,"19 Novation unnecessarily maintained EPA's projected vehicle load reductions and
considered only more advanced powertrain technology such as costlier HEV or BEV packages.
By not considering additional vehicle load reductions as part of the alternative pathway, the
19 Novation Analytics, Technology Effectiveness - Phase I: Fleet-Level Assessment, version 1.1, prepared for the
Alliance of Automobile Manufacturers & Association of Global Automakers, October 19, 2015, p.64.
33
-------
analysis is predisposed to require more expensive powertrain technology and therefore project
higher costs and higher levels of technology usage.
Both of these comments are considered in further detail below.
Constraints on Technology Combinations: In Appendix A of the TSD, EPA stated that the
methodology in the Novation report essentially assumes "that all possible technology available in
2025 can be represented by technology already contained in the MY2014 baseline fleet." In their
comments, Novation disagreed with EPA's characterization, stating that they "used the current
powertrains as a foundation upon which it added the technologies assumed by the FRM."
Furthermore, they stated, "[t]his is fundamentally the same process that the agencies use:
measure the performance of current production powertrains and powertrain components to
establish a baseline, then add those technologies and technology combinations that do not exist
in the fleet today. The difference is simply system-level analysis versus component-level
analysis."
However, this is an example of where Novation inappropriately uses similar terminology to
refer to different concepts in an attempt to draw a parallel between their process and EPA's. In
fact, the specifics of the Novation process bear little resemblance to processes used by EPA, and
their reference to a "system-level analysis versus component-level analysis" merely attempts to
mask the fact that Novation's process systematically fails to account for the existence (and
effect) of sub-technologies within their technology bundles, and thereby tends to under-predict
potential improvements in technology effectiveness.
When EPA adds technologies to the baseline fleet, EPA uses multiple data sources, as
described in Chapter 2 of the TSD, to determine effectiveness values for specific technologies
alone and in combination, including some combined powertrain packages that do not exist in the
fleet today (although most or all of the individual sub-technologies do exist). These individual
technologies include, for example, variable valve lift and timing, Atkinson cycle engines, engine
friction reduction, early torque converter lockup, gearbox efficiency improvements, and others.
In contrast, when Novation adds technologies to the baseline, they consider only broad
categories of powertrain technology "bundles" (see Novation Comments at pp. 6 and 7), and set
the efficiency range of these bundles such that the 50th percentile of powertrain efficiency
represents current fleet efficiency levels.20 The potential existence of sub-technologies such as
Atkinson cycle engines, engine friction reduction, early torque converter lockup, or gearbox
efficiency improvements is not represented. This failure to consider the individual effects of
known technologies is a critical and inherent shortcoming of the report.
In their comments, Novation points out, as evidence that their report accounted for technology
advancement, that they included in their analysis "powertrain combinations [which] are not in
production," specifically "advanced spark-ignition (SI) based powertrains [i.e., 24 bar
turbocharged / downsized engines with cooled EGR] with high ratio spread transmissions and
stop start." While true, this comment confounds the mere existence of powertrain bundles not in
the fleet with the ability to account for additional technology added within a powertrain bundle -
or more precisely, with the a priori methodological choice not to consider such technology
additions. This necessarily leads to an underestimation of efficiency, as noted above.
20 Fleet-Level Assessment, p. 48.
34
-------
In fact, Novation's process ensures that the powertrain efficiency of "future" powertrain
bundles are tied to the specific sub-technologies that are included in those bundles in the
MY2014 fleet. For example, under the Novation methodology, a naturally aspirated engine, high
ratio spread transmission bundle would never be modeled with the combination of Atkinson
cycle, engine friction reduction, cooled EGR, early torque converter lockup, and gearbox
efficiency improvements, simply because that combination does not exist in the MY2014 fleet.
However, there is no inherent reason why manufacturers cannot build such a package if they
choose, and so the restriction in Novation's modeling artificially leads to lower estimates of
potential powertrain efficiency improvement.
In like fashion, both the Association of Global Automakers and Toyota provided similar
comments relating to EPA's criticism of Novation's methodology. Global stated that the
criticism "misses the point of Novation's work, which was in part to assess EPA's contention
that CO2 targets can be met through advancements to the current internal combustion engines."
Toyota claims that Novation's study includes "powertrain efficiency distributions and
deployment scenarios [which] are mechanisms that account for technology advancement."
However, like the original reports authored by Novation, these comments confound an
assumed advancement along a statistical curve due solely to "learning and implementation
improvements"21 with the incorporation of specific advanced technologies into a vehicle
powertrain. EPA's analysis accounts for the effects on efficiency attributable to each sub-
technology, and assumes that manufacturers will adopt the technologies of their choice as
needed. In contrast, the Novation methodology simply assumes, a priori, that powertrain
efficiency in 2025 is limited to small incremental improvements over that which is available
today, regardless of available combinations of sub-technologies. In their comments, UCS, ICCT,
and ACEEE agree with EPA's assessment of the Novation methodology, with ACEEE
commenting that Novation "assumes a given technology can be no more efficient on average in
2025 than the best implementations of that technology in 2014. This is an arbitrary constraint
that clearly does not apply for all technologies."
In Appendix A of the TSD, this inappropriate confounding of advancement along a statistical
curve with the incorporation of specific advanced technologies was discussed, presenting the
example of vehicles with Atkinson cycle engines or engines with cylinder deactivation, which
would presumably be included primarily within a bundle of SI naturally aspirated engines,
coupled with a non-high ratio spread transmissions and without stop-start.
In their comments, Novation responded that they were "requested to consider" only vehicle
packages used in the FRM, and "not alternative powertrain technologies that EPA may now be
evaluating (Novation comments p. 7)." EPA acknowledges that Novation may have been
following the request of the contracting organizations, the Alliance and Global, in not explicitly
considering the effect of Atkinson engines or cylinder deactivation technologies. However, such
direction to Novation does not mean EPA should disregard the resulting limitation in Novation's
work product. It is indisputable that Atkinson engines and cylinder deactivation exist in the MY
2014 fleet, yet the Novation methodology does not account for these actual technologies, instead
lumping all powertrains into generic groups and mistakenly attributing the actual differences in
powertrain efficiency due to advanced technology as "learning and implementation
21 Fleet-Level Assessment, p. 78.
35
-------
improvements." This lack of accounting for real technology, whether or not included in the
FRM, underscores the flaw in Novation's methodology.
Furthermore, in their comments, Novation states "the LPM, on which most of Novation's
analysis was focused, describes powertrains by broad technology packages ... In the Proposed
Determination, EPA continues the practice of defining powertrains as broad technology
packages; hence, by criticizing Novation, EPA is calling into question its own approach." This
mischaracterizes the usage of the LPM and again highlights the flaws in Novation's approach to
package building. EPA's technology packages are combinations of specific technologies, where
the effect on CO2 emissions of each sub-technology is accounted for in the aggregate package. In
contrast, Novation's technology bundles include powertrains incorporating a range of different
technologies, and all powertrains in the bundle are assumed to have an equivalent level of
technology. For example, as noted in the TSD and above, the Novation analysis would class
Atkinson cycle engines or engines with cylinder deactivation along with other naturally aspirated
engines, with no differentiation.
Responding to criticism of a lack of technical rationale for using a CI (diesel) engine as a
"proxy" to represent a 27 bar SI (gasoline) engine, Novation states (p. 8) that in their approach,
"the key attributes that allow diesel engines to achieve higher efficiencies than current spark-
ignition engines... are the same benefits that EPA was claiming for the direct-injected, dilute, and
highly boosted engines that served as the foundation of the FRM." Although EPA agrees that the
use of CI engine efficiency to represent a 27 bar SI engine is directionally correct, the context of
these comments highlights the differences between the EPA analysis and the approach used by
Novation in their studies. Rather than rely on superficial similarities between engine
technologies to estimate engine efficiencies as Novation does, EPA evaluates the SI engines
themselves. Consequently, there is no need to rely on a proxy engine of a different type and
results in a more robust analysis.
Vehicle Load Penetration Rate Changes: In Appendix A of the TSD, EPA noted that the
Novation study did not examine the effect of potential changes in vehicle load reduction
penetration rates, even in circumstances where it is clearly appropriate to do so. In comments
referring to this discussion, Novation states that their study "assumed the same mass,
aerodynamic drag, and tire rolling resistance reductions as assumed by the agencies in the FRM."
However, this comment confounds the consideration of changes in vehicle load reduction
penetration rates with incorporation of vehicle load reduction as a technology at all.
Novation states that their objectives in the studies were "to evaluate the sustainability of the
FRM powertrain effectiveness assumptions, not the vehicle load assumptions." A comment by
the Association of Global Automakers made a similar point. However, the Novation report goes
beyond simply evaluating powertrain effectiveness assumptions. Novation also uses their
analysis to model "alternative technology pathway scenarios,"22 where they seek to quantify the
technology penetration mix required to meet the MY2021 and MY2025 standards in an
alternative compliance scenario where powertrain technology effectiveness is lower.
Novation further states in their report that the entities commissioning the report, the Alliance
and Global, specifically requested this analysis be done "given the levels of vehicle energy
22 Fleet-Level Assessment, p. 64.
36
-------
reductions forecasted by the agencies,"23 which narrowed the focus of the study to powertrain
efficiency requirements only, and disallowed consideration of changes due to additional vehicle
load reductions. In response, EPA notes that potential changes in vehicle load reduction
penetration rates can reduce the need for addition of other technologies - particularly HEV and
BEVs - in the alternative technology pathway scenarios proposed by Novation. The decision to
omit vehicle load penetration rates from consideration thus leads to projections of a greater need
for relatively costly powertrain technology additions.
As further support for its decision to hold road load reductions constant, Novation suggests
that EPA keep also kept its vehicle penetration rates constant across the fleet in its analysis,
saying "in both the TAR and Proposed Determination documents, EPA uses the same, generic,
assumptions for [reduction in mass, aerodynamic drag, and tire rolling resistance] as it did in the
FRM. Again, by criticizing Novation, EPA is calling into question its own assumptions."
However, it is not the case that EPA kept its vehicle penetration rates constant.
Novation's statement appears to confound discrete levels of reduction in vehicle load
parameters and overall penetration rates of these technologies into the fleet. For example, when
building future vehicle packages, EPA assumes, as a modeling convenience, discrete levels of
reduction in mass, aerodynamic, and rolling resistance loading. These levels have remained the
same since the FRM. However, EPA's OMEGA model assumes that manufacturers will choose
the most cost-effective technologies throughout their fleet to comply on a fleet-wide basis. When
technology cost or effectiveness numbers change based on stakeholder input, the penetration
rates of specific technologies, including vehicle load reduction technologies, can also change.
Thus, although EPA has kept the definition of vehicle load reduction levels constant, that does
not mean EPA has kept vehicle load reduction penetration rates constant across the fleet, as
Novation erroneously states. This is another example of a significant difference between the
Novation Analytics analysis and the EPA analyses; EPA applies technology in packages with
increasing content where individual technologies have been modeled with high fidelity. Changes
in road load result in significantly different engine and transmission operation, which affects the
overall effectiveness of the entire technology package. EPA believes that this process best
reflects how manufacturers design and develop vehicles to optimize the efficiency of the vehicle
as a system.
Plausibility Checks: In their comments, Novation states that "plausibility checks show
individual vehicle simulations from the FRM that had cycle average efficiencies that were higher
than the peak engine efficiency of the best engine maps used in the FRM, which is an impossible
outcome." EPA agrees that average cycle efficiencies exceeding peak engine efficiencies is
impossible, but more importantly EPA has examined the average cycle efficiencies of the
packages used in the TSD and found no such cases. In fact, the vast majority of the technology
packages applied in the PD central analysis for 2025 have average cycle efficiencies no more
than 84 percent of the peak engine efficiency, and no applied technology packages have average
cycle efficiencies more than 92 percent of the peak engine efficiency.
In the TSD Appendix A, EPA referenced an article by the Union of Concerned Scientists
which claimed that one current production vehicle, a Honda Fit, would be deemed implausible
by the Novation methodology. In their comments, Novation disagrees, stating "Novation would
73 Fleet-Level Assessment, p. 64.
37
-------
not deem the Honda Fit implausible. The MY2016 Fit is within the best 1 percent of Si-based
powertrains, having a combined efficiency of 25.5 percent; yet, it is 12 percent below the stated
plausibility limits established by Novation Analytics." EPA acknowledges that Novation's
calculation of the Flonda Fit powertrain efficiency is correct.
In the TSD Appendix A, EPA gave an example of the overly restrictive assumptions Novation
uses, specifically using current engine technology to determine the limit of on-cycle-to-peak
engine efficiency ratio ("Plausibility Test 2"). The example refers to two engine maps, a
MY2013 Chevrolet Malibu 2.5L 14 GDI map and a 27-bar BMEP cooled EGR turbo GDI map
(Figure 1.1 in the TSD Appendix A, and reproduced as Figure 2-6 in this RTC document). The
27-bar BMEP cooled EGR turbo GDI map has an enlarged area of high efficiency in the lower
left (indicated by the arrow in Figure 2-6b). Since this is the area where engines tend to run over
the cycle, the figure shows an example of how the application of engine technology can result in
a better match between engine operation and peak efficiency. This reduces CO2 emissions,
precisely by increasing the on-cycle-to-peak engine efficiency ratio. A comment received from
UCS agrees, stating in addition that "[IJowering this ratio is precisely the objective of much of
the research on conventional spark-ignition engines."
1000 2000 3000 4000 5000 6000
bUU 1UUU IbUU AJUU &JJ JUUU JbUU
RPM RPM
(a) MY2013 Chevrolet Malibu 2.5L 14 GDI (b) 27-Bar BMEP Cooled EGR Turbo GDI
Figure 2-6 Two Engine BSFC Maps, Reproduced in Technology Effectiveness - Phase II: Vehicle-Level
Assessment
NOTE: These maps are cited during the development of "Plausibility Test 2," The left-hand map is overlaid with
areas of typical 011-cycle engine operation. Original sources are given in the Novation report.
EPAJJDDS plotted on 11-Jul-2014
Avg. load (bmep) and
speed decrease with
increasing displacement
However, because the Novation report develops their plausibility limit for on-cycle-to-peak
engine efficiency ratio based on a few MY2013-2014 vehicles, there is no room left below their
arbitrary limit for the potential improvement in the efficiency matching shown in Figure 2-6. Put
another way, the plausibility limit developed by Novation implicitly assumes that any potential
improvements in engine technology which increase the efficiency of the engine while operating,
relative to peak engine efficiency, are implausible.
38
-------
Novation, in their comments, disagrees with the EPA assessment that they have no room for
potential improvement due to efficiency matching, saying, "Novation assumed future
improvements to on-cycle-to-peak engine efficiency ratios of 19 percent on the city cycle, 10
percent on the highway cycle, yielding 15 percent combined." However, the assumed future
improvements cited by Novation are specifically tied to the implementation of two technologies:
engine stop-start and higher ratio spread transmissions.24 Neither of these technologies alters the
area of peak efficiency on the engine map, and thus, there is no accounting for potential
improvement in the efficiency matching as stated in the TSD. A comment received from UCS
agrees with this assessment, pointing out that the limit that "the ratio of test cycle efficiency to
peak efficiency should not exceed 0.78 is violated by engines already now in prototype."
Novation in their comments further disagrees with EPA's example of the overly restrictive
nature of their plausibility checks by stating that EPA "relies on an illustrative example of an
engine map that is not from an actual, tested engine." In response, EPA agrees that this is an
illustrative example, all the more so since the 27 bar map shown, and included in Novation's
original report, contains the overlaid arrow indicating how the area of greater efficiency is
expended to the lower left, to better match the operating range of the engine during vehicle cycle
operation. This is specifically illustrative of potential improvement in efficiency matching, and
also illustrates UCS' comment that lowering the on-cycle-to-peak engine efficiency ratio is
precisely the objective of much of the research on conventional spark-ignition engines.
Novation continues in their comments, "[f]urthermore, the technology assumed from this map
was not included in the TAR or the Proposed Determination." Although this statement is
something of a non sequitur, the 27 bar Ricardo map shown in Figure 2-6, and the underlying
technology package, was indeed used to create 24 bar turbo downsized maps for the Draft TAR
and the PD.
In their comments on the Proposed Determination, the Alliance reiterated their
recommendation that EPA adopt the "plausibility checks" developed by the Alliance's
contractor, Novation Analytics, and stated that EPA did not provide any reason for rejecting the
methodology recommended. EPA has considered the proposed plausibility limits developed by
Novation and explained the reasons for not adopting them, discussing these points at length in
Appendix A of the TSD. EPA has also considered Novation's additional comments on the issue,
and does not find them persuasive, as discussed above. EPA thus is not adopting the
recommended plausibility checks.
In their comments, both the Alliance and Novation furthermore stated that EPA did not
propose alternative numerical limits to the plausibility checks developed by Novation. These
comments presuppose that the development of a numerical "plausibility" limit is necessary. EPA
discussed at length in Appendix A of the TSD the shortcomings of the proposed plausibility limit
calculation, and as an alternative noted that "calculation of powertrain efficiency can serve as a
gross QC check on estimated technology effectiveness by quickly identifying the highest
efficiency packages for further review (as shown in Appendix B [of the TSD])." Correctly and
effectively using powertrain efficiency analysis as a QC check does not require the adoption of
arbitrary limits, and EPA declines to do so.
24 Novation Analytics, Technology Effectiveness - Phase II: Vehicle-Level Assessment, version 1.0, prepared for the
Alliance of Automobile Manufacturers & Association of Global Automakers, September 20, 2016, p. 28.
39
-------
Other Considerations: In Appendix A of the TSD, EPA disagreed with the assertion by the
Alliance and Novation that the Lumped Parameter Model [LPM] is "not based on the
fundamental factors determining vehicle CO2 and fuel consumption." EPA further explained the
usage of the LPM, in particular the LPM "exemplar" vehicles, each of which has different engine
sizes and road loads. The differences between exemplar vehicles does indeed account for the
"fundamental factors determining vehicle CO2" that Novation refers to, as the power-to-weight
ratios of the exemplar vehicles vary, altering the relative areas of the engine map where each
vehicle operates. In their comments on the TSD, Novation asserts that the LPM "provides the
processing speed required to support the OMEGA model." Although this is another non sequitur,
EPA does not disagree with this assessment.
In their comments, Novation further states that "there was a lack of information published by
EPA," and that EPA "has been resistant to providing support for these studies." They
furthermore cite an email from Michael R. Olechiw (Director, Light-duty Vehicles and Small
Engines Center, US EPA) to Greg Pannone (President, Novation Analytics). This email was
regarding details of the 2010 FRM. The purpose of the ongoing inquiries from Novation
Analytics is now clear, as Novation has explained the scope of their work as being limited to a
study of the 2010 FRM. If EPA appeared dismissive in its response, it was because the MTE is
intended to consider the MY2022-2025 standards, and it was therefore difficult to understand
how the 2010 FRM (which established standards for MY2012-2016) results were relevant. EPA
did offer to discuss the email with Novation; however, the offer was declined.
Additionally, in their comments, both the Alliance of Automobile Manufacturers and the
Association of Global Automakers state that the methodology Novation used was shared with
EPA in 2014 and "EPA never raised concerns or provided feedback indicating that they found
the methodology insufficient or lacking in any way." EPA in fact conducted numerous meetings
with multiple stakeholders as part of the process of gathering information to inform the Midterm
Evaluation process. Rather than respond to each stakeholder individually and in real time, the
sum total of information gathered from stakeholders was synthesized into the Draft TAR. When
the Alliance indicated that they disagreed with some points in the Draft TAR and, in support,
formally submitted reports from Novation as part of the comment period, EPA replied to those
comments in the Proposed Determination.
Novation additionally states that "[t]he methodology used by Novation... has been
independently reported by other research.... Consequently, to suggest that this approach is
without merit is to suggest that these other authors and peer reviewers were also incorrect."
Indeed, the use of tractive work and powertrain efficiency metrics is well-known, and can be a
useful modeling technique when applied in an appropriate way. However, EPA does not agree
that it was applied appropriately throughout the Novation study. Novation interprets EPA's
statements over-broadly, implying that EPA's criticism is of the techniques themselves rather
than the application thereof. EPA specifically rejects the implication that statements in Appendix
A of the TSD apply to researchers who apply these techniques appropriately.
2.2.4 Vehicle Classifications
Summary of Comments on the Draft TAR addressed in the Proposed Determination
In the 2012 FRM and Draft TAR analyses, vehicles were classified into 19 vehicle types,
which were based on six size-based categories for estimating effectiveness, and several cost
40
-------
categories defined by the various engine and valvetrain configurations most prevalent in the
baseline fleet. While overall this method of grouping placed similar vehicles together,
stakeholder comments on the Draft TAR highlighted some examples where dissimilar vehicles
were assigned the same cost and effectiveness benefits.
In response to these comments, as described in TSD Chapter 2.3.1.4, EPA refined the vehicle
classification approach for the Proposed Determination in several ways. First, we classified
vehicles according to the attributes of vehicle road load power and engine power-to-vehicle
weight ratio for the purpose of assigning the most representative estimates for technology
effectiveness (see Chapter 2.3.3.2 of the TSD). Second, we implemented a classification by
vehicle curb weights, together with engine configuration and the capability for heavy towing, as
attributes used for assigning technology costs. Third, we expanded the number of vehicle types
from 19 to 29.
Compared to the Draft TAR, the 29 vehicle types used for the Proposed Determination each
contained a narrower range of values or the vehicle characteristics that have the greatest
influence on technology effectiveness and cost: power-to-weight ratio, road load power, curb
weight, and original engine configuration. Overall, consistent with the public comments, this
updated classification approach provided greater resolution than the 19 vehicle types used in the
Draft TAR, and advanced the goal of applying the most representative cost and effectiveness
estimates for technologies applied to the MY2015 fleet.
Summary of Comments and Responses on the Proposed Determination
The Union of Concerned Scientists (UCS) expressed support for the changes in classifications
to a power and road load basis, characterizing this change as responsive to "one of the strongest
industry concerns" expressed in comments on the Draft TAR. UCS further stated that the
updated classifications strengthen EPA's analysis by "narrowing the error bars and more
accurately representing the real vehicle fleet." The American Council for an Energy-Efficient
Economy (ACEEE) also positively highlighted this change when describing the additional
analysis EPA performed for the Proposed Determination.
2.2.5 Performance Neutrality
Summary of Comments on the Draft TAR addressed in the Proposed Determination
EPA's assessments for the 2012 FRM and Draft TAR were based on the application of
technology packages while holding the underlying acceleration performance constant. To
achieve this, ALPHA modeling runs were used to generate technology effectiveness values while
maintaining a set of acceleration metrics within a reasonable window by adjusting engine
displacement.
The Alliance of Automobile Manufacturers, in comments on the Draft TAR stated, "In
practice, manufacturers have a limited number of engine displacements to choose from and will
likely select the size of engine that maintains or improves performance." For the Proposed
Determination, EPA continued to apply the constant performance criterion, and did not attempt
to model discrete engine sizes or fleet-wide performance improvements that are made available,
at least in part, by the efficiency technologies adopted to comply with the standards. As
discussed in TSD Chapter 2.3.1.2, even if our model produces a greater variation in technology
packages than exists today, this does not require that manufacturers actually produce a greater
41
-------
variety of component sizes than exist currently in order for our overall results to be valid. In
actual vehicle design, manufacturers will design discretely sized components, and for each
vehicle choose the available size closest to the optimal for the given load and performance
requirements. For example, in some cases, the chosen engine will be slightly smaller than
optimal (and thus have lower fuel consumption), and in some cases the chosen engine will be
slightly larger than optimal (and thus have higher fuel consumption).
Other comments on the Draft TAR criticized the use of acceleration time as the main metric
used to represent performance neutrality, and stated that top gear gradeability is another key
metric that was omitted in the analysis. For the Proposed Determination, EPA did not
incorporate gradeability as an additional performance metric. As discussed in TSD Chapter
2.3.4.2.2, maintaining top gear at 75 mph up a grade, as the commenters recommend, may not be
appropriate for advanced eight-speed transmissions, where EPA testing has indicated downshifts
regularly occur and are likely to be less noticeable to the driver (due in part to the smaller step
changes in speed between each gear of a higher-gear-number transmission, as discussed in TSD
Chapter 2.2.3.10 at p. 2-59).
Summary and Response to Comments on the Proposed Determination
In comments on the Proposed Determination, the Alliance of Automobile Manufacturers
noted that within the ALPHA files generated for the Proposed Determination, the "Truck" class
file shows 0-60 mph acceleration times averaging 15.9 seconds. A discussion of "performance
ballast" was not included in the TSD, and EPA thanks the Alliance for pointing this out. To
model the acceleration performance of the truck class, an additional performance weight was
added to the ETW to simulate hauling a load. The value of the "performance ballast" (3000 kg)
was noted in column J of the ALPHA files referred to by the Alliance in their comment.
The performance ballast was not used during the standard FTP and HWFET emissions cycles,
but only during the performance cycle used to calculate acceleration times. The acceleration
performance times noted by the Alliance reflect the presence of the added mass. The approach of
ensuring performance neutrality while hauling 3000 kg generally results in more conservative
technology effectiveness than would be obtained by using simply ETW or curb weight during the
acceleration cycles.
The Alliance also suggested in comments that EPA incorporate a gradeability metric into the
performance calculation, such that the vehicle "maintain[s] top gear at 75 mph while climbing a
given grade." The subject of gradeability was discussed in the TSD Chapter 2.3.4.2.2, where
EPA stated that "EPA does not believe this metric is appropriate for advanced eight-speed
transmissions;" however, the Alliance disagreed with EPA's assessment. In response to the
Alliance's comment, EPA has identified from publically available sources 14 examples from
MY2012 to MY2017 where vehicles were refreshed, maintaining the same chassis and engines
but incorporating transmissions with wider ratio spreads.25 Twelve vehicles maintained curb
weight within 1 percent; of the remaining two, one increased curb weight by 4 percent and one
decreased curb weight by 4 percent; the average curb weights for the 12 vehicles remained
unchanged. For each vehicle, the top gear ratio (including final drive and accounting, in a single
instance, for tire size difference) was calculated. In all 14 vehicles, the top gear ratio was
reduced, by 12 percent on average. Although top gear ratio and vehicle weight are not the only
25 See docketed spreadsheet "Comparison of transmission top gear for selected vehicles.xlsx."
42
-------
components in the calculation of top gear gradeability (other quantities such as change in wide-
open throttle torque at speed and transmission efficiency play a small role), they are the largest
factors, and directionally indicative of manufacturers' choices when implementing transmissions
with wider ratio spreads. EPA stands by its analysis that top gear gradeability is not an
appropriate metric when judging performance of advanced eight-speed transmissions.
Further, in considering the Alliance's comment, EPA used its ALPHA tool to estimate
gradeability of the vehicle technology packages used to calibrate the lumped parameter model.
For the Draft TAR, the calibration packages were constructed with a consistent final drive ratio,
but for the Proposed Determination, a dynamic final drive resizing algorithm was adopted which
increased the final drive ratio as engines were downsized. This algorithm resulted in top gear
ratios that more closely match industry trends; the resulting gear ratios are included in the
ALPHA calibration files available on the EPA website.26 The results were that packages
modeled with the TRX21 transmission, when properly matched for acceleration performance,
outperformed the baseline vehicle in gradeability in most cases. In the two cases where it did not,
gradeability was reduced by 0.5 percent of grade. Packages modeled with the TRX22
transmission and advanced engines also generally outperformed the baseline vehicle in
gradeability. In many cases, this was due to higher available torque at engine rpm equivalent to
75 mph. For three of the six vehicle classes, all advanced packages showed improved
gradeability. For two of the classes, gradeability results were mixed, with gradeability improving
up to 3 percent of grade for most packages and decreasing less than 0.5 percent of grade for
some packages. For the final HPW class, gradeability increased around 2 percent of grade for
some packages and decreased less than 1 percent of grade for the remainder. Packages modeled
with the base GDI engine and TRX22 transmission had in some cases further reduced
gradeability. Thus, the gradeability generally improves when the acceleration performance is
held constant, and EPA continues to believe that matching acceleration performance as a metric
is the appropriate way to ensure performance neutrality.
2.3 Estimated Costs (Technology Costs, Total Costs, Learning)
This chapter reviews key comments on the general topic of costs. Comments relating to cost
as it applies to the EPA analyses range across a wide variety of topics, including estimates of
technology costs for specific technologies, projections of compliance costs for individual
manufacturers and across the fleet, modeling of specific types of costs such as direct and indirect
costs, the impact of manufacturer learning, and other related topics. Many of our responses to
comments related to cost are therefore distributed among the other chapters of this RTC
document that deal more specifically with various assessment and modeling topics and our
assessment of specific technologies.
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Numerous comments on the Draft TAR presented general arguments that EPA's estimated
costs are too optimistic (primarily from industry commenters) or are too conservative (primarily
from NGO commenters). In many cases, these comments can be described as general comments
because they express an overall viewpoint on costs, though commenters included varying
degrees of supporting evidence for their position, or evidence that was not readily applicable or
26 ALPHA v2.1 Calibration Sample, available from https://www.epa.gov/regulations-emissions-vehicles-and-
engines/advanced-light-duty-powertrain-and-hybrid-analysis-alpha.
43
-------
not relevant to assessing the cost, effectiveness, and implementation feasibility in MYs 2022-
2025. EPA responded to such comments in the context of the specific cost-related issues that
commenters raised.
In commenting on the Draft TAR, multiple comments from NGOs (American Council for an
Energy-Efficient Economy (ACEEE), Union of Concerned Scientists (UCS), and Environmental
Defense Fund (EDF)) supported EPA's use of Indirect Cost Multipliers (ICMs) rather than retail
price equivalents (RPEs) as a means of estimating indirect costs.
We also received some comments on our modeling of cost reductions through manufacturer
learning. Ford argued that product cadence does not allow for cost reductions from learning to be
realized since new products are constantly being developed. In the Proposed Determination, we
noted that the learning effects we estimated should be taken as occurring at the level of the
supplier, not that of the automaker. Since we have not estimated efficiency improvements to
individual technologies during the time frame of the analysis, we do not believe that such
redesign to improve the "current best technology" to the "next best technology" is necessary to
achieve the reductions we expect for the costs we have estimated. More discussion of these
comments and responses and the use of ICMs may be found in TSD Chapter 2.3.2 at p. 2-214
and TSD Chapter 2.3.2.2.1 beginning at p. 2-223.
Summary and Response to Comments on the Proposed Determination
Many of the comments on the Proposed Determination related to cost were largely similar to
those received on the Draft TAR, in that they expressed general arguments that the costs were
either too optimistic or too conservative, supported by varying degrees of supporting evidence or
information. Many of these comments may be described as general comments and some express
viewpoints that are shared by multiple commenters, so a response directed to one commenter
may also apply to other commenters. Many of our responses to comments related to cost are
distributed among the other chapters of this RTC document that deal more specifically with
various assessment and modeling topics and our assessment of specific technologies.
Consumers Union stated its assessment that the standards are cost effective. Consumer
Federation of America was also supportive of EPA's technology cost estimates, presenting
arguments supporting its position that industry tends to overestimate the cost of complying with
regulatory standards.
The Alliance and some individual OEMs stated that differences in compliance cost estimates
between the EPA and NHTSA analyses in the Draft TAR suggest that the agencies disagree
about how OEMs can comply. EPA disagrees, and believes that this is a misinterpretation of the
results. In the Draft TAR, EPA, NHTSA and CARB concluded that the standards can be met
predominantly with advanced gasoline technologies and very low penetration of electric
vehicles, although the modeling supporting those conclusions projected slightly different cost
impacts. This is to be expected given the differences in modeling tools and inputs between the
two agencies, as clearly explained in the Draft TAR (for example, in Draft TAR Chapter 2.3
which describes the agencies' approach to independent GHG and CAFE analyses).
Subaru commented that our estimated costs are surprisingly low and argues that a limited-line
OEM like Subaru has a much narrower path to compliance than EPA assumes. Subaru also
argues that, even if ICEs achieve a brake thermal efficiency (BTE) of 50 percent, Subaru would
44
-------
need significant electrification to comply. However, because the EPA analysis has not specified
that Subaru would have to achieve any specific level of BTE in its engines in order to comply,
the significance of this specific example is unclear, and the comment provides no additional
information to evaluate the claim or compare it to the assumptions in the EPA analysis to
understand why the projections differ. Further, our analysis shows no increased electrification for
Subaru beyond that necessary for the ZEV program (see Table C.21 of the Proposed
Determination). In addition, EPA does not agree that the limited-line nature of Subaru's fleet
impacts compliance pathways or that its customer demands make compliance more difficult than
for other OEMs. Subaru's fleet consists largely of vehicles placed on the truck curve, which has
considerably higher CO2 targets for each footprint value than does the car curve. Subaru also
questions the $0 cost of 60 percent of Subaru's earned GHG credits. However, this comment
appears to be directed at the NHTSA analysis, since EPA did not make any such valuation of
earned credits.
ICCT commented that EPA could reduce per vehicle costs by "several hundred dollars" by
removing technology availability restrictions and including or expanding upon available
technologies. However, we have chosen to remain consistent with the 2012 FRM's determination
of technology availability restrictions (technology penetration caps), recognizing that this likely
makes our analysis more conservative. See also the response in Chapter 2.5.3 of this RTC
document where we address a similar ICCT comment with regard to engine technologies.
Ford reiterated that increased cadence of technology implementation limits opportunity for
learning to reduce costs. Ford also argued that consumers do not value fuel economy highly
enough to recover what they expect to be the additional costs of complying with the standards.
We disagree with the latter position and discuss this more in Chapter 3 of this Response to
Comments document. As for the impact of cadence on learning, as we discussed in the Proposed
Determination, our learning impacts are assumed to occur at the supplier level rather than the
OEM level (see the Proposed Determination TSD Section 2.3.2 at page 2-2015). As such, a
possible transition from naturally aspirated to 18-bar turbocharging to 24-bar turbocharging for
the OEM is not expected to impact the suppliers' learning on the turbochargers themselves. Ford
also argues that EPA should assess the comprehensive cost increments related to both Tier 3 and
GHG compliance. We believe we have done this by analyzing the Tier 3 requirements in that
rulemaking and the GHG requirements via the 2012 FRM and again in the Draft TAR and the
Proposed Determination.
Honda states that it believes "the targets for 2025 are correct, and consistent with long-term
environmental goals," but also states that "compliance costs are more than double the amount
estimated by the agencies," stating that the technology needed to comply is underestimated while
marketing challenges are another difficulty. Honda's comment on the Proposed Determination
did not contain detailed written remarks and primarily directed EPA to the comments by Global
Automakers, and stated that Honda generally supports those comments.27 EPA has responded to
the comments from Global in the chapters of this RTC document that correspond to the specific
detailed topics that Global raises, and EPA's response to Honda's comments are therefore
represented by those responses.
27 At the time of its comment, Honda also provided to the Docket a redacted copy of a presentation delivered to EPA
on November 9, 2016.
45
-------
Mercedes-Benz also commented that EPA overestimates its ability to comply and
underestimates the cost of complying, is concerned that the cost of compliance will be higher for
the Mercedes fleet than for competing OEMs with a more diverse fleet (due largely due to
expectation of a greater than projected reliance on electrification), and requests that EPA
consider additional flexibilities to promote a level playing field. Comments that relate to a higher
need for electrification than EPA projects are addressed in Chapter 2.1 of this RTC document,
while comments related to credits and flexibilities are addressed in Chapter 3.9.
Mercedes-Benz also commented that they anticipate their costs to be higher because
designing, certifying, training, and stocking of parts for a low volume electrified product is not
cost-effective. In response, EPA notes that Mercedes-Benz has a long history of including
relatively low volume, somewhat "niche" vehicles with conventional powertrains (often high-
performance) in their product line, and does not see a compelling rationale as to why it expects
these issues to affect low volume electrified vehicles more strongly than they affect these low
volume conventional vehicles. Mercedes also expressed concerns with design cost amortization
for small volume manufacturers. However, Mercedes currently designs and sells a wide variety
of products, and these products are regularly equipped with a variety of features and technologies
that are leading edge. EPA believes that Mercedes, and other lower volume, high performance
and luxury vehicle manufacturers, will manage their cost amortization the same way they do for
all of the features that are not related to improving vehicle efficiency.
2.4 Lead Time
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Lead time is a significant component of technical feasibility, in that time is an inherent factor
in bringing any advanced technology from research to widespread production. In the TSD, EPA
discussed lead time in the context of specific technologies (such as Atkinson cycle engine
technology) as well as in the context of general technical feasibility of the MY2022-2025
standards.
Some commenters on the Draft TAR made general allusions to the perceived difficulty of
meeting the 2022-2025 standards with advanced technologies given the available lead time. In
contrast, several NGOs recognized the value and adequacy of the lead time already provided by
the standards. In response, EPA pointed out that the standards for MY2022-2025 were first
established in 2012, providing the auto manufacturers with up to 13 years of lead time for
product planning to meet these standards, representing multiple vehicle redesign cycles that
provide opportunities for technology introduction. EPA also cited ongoing evidence of the
increasingly rapid pace at which manufacturers are bringing advanced technologies into the fleet,
to the extent that technology adoption rates and the pace of innovation have accelerated even
beyond what EPA projected when initially setting these standards. EPA also pointed out that the
technologies considered in the Draft TAR and Proposed Determination analyses are either
currently in production or will be commercially produced in the next several years. More
discussion may be found in Section IV.B of the Proposed Determination document at p. 48-50,
and TSD Chapter 2.3.1.1 at p. 2-207.
Some comments received on the Draft TAR related specifically to the lead time required for
introduction of Atkinson cycle engine technology at the penetration rates projected in the Draft
TAR. EPA disagreed that introduction of Atkinson cycle technology would require greater lead
46
-------
time than afforded by the time frame of the 2022-2025 standards, pointing out (as discussed in
the TSD at p. 2-309 to 2-310) that the steps required to implement an Atkinson cycle engine are
relatively modest compared to implementing other engine technologies, and that many of the
building blocks are already available in the MY2016 fleet, including gasoline direct injection and
high levels of valve train authority (further suggesting that a major vehicle redesign cycle is not
necessarily required to introduce this technology). EPA also cited the fact that the technology has
been introduced and has undergone several revisions within the past 5 years, suggesting that lead
time requirements are not as great as the commenters suggest.
EPA also cited several examples of manufacturers that have either already implemented or
have indicated plans to implement forms of Atkinson cycle technology. EPA also noted that the
projected technology penetrations of the Draft TAR are meant to illustrate one of many possible
technology pathways for compliance, and that manufacturers are free to pursue paths that are less
reliant on Atkinson cycle technology if they choose to do so. EPA cited sensitivity analyses that
indicated that cost effective compliance paths using primarily other advanced engine
technologies exist even when Atkinson cycle technology is limited (i.e. arbitrarily constrained
for purpose of the sensitivity analysis) to far lower penetrations. EPA concluded that it is feasible
for this technology to be incorporated by any manufacturer and that there is sufficient lead time
until MYs 2022-2025 that this technology could represent a significant penetration rate of a
company's products, if it chooses to employ this technology. More discussion of these comments
can be found in Section IV.B of the Proposed Determination document at p. 48-50, and TSD
Chapter 2.3.4.1.8.3 at p. 2-309 to 2-310, as well as in Chapter 2) (Atkinson Cycle Engine) of this
RTC document.
Similarly, commenters positing the need for significant penetration of electrified vehicles as a
necessary compliance path further stated that there would be inadequate lead time to do so. EPA
reiterated that both the Draft TAR and Proposed Determination analyses continue to indicate that
relatively low penetrations of strong hybrids or electric vehicles would be needed to comply with
the standards, and that there will be adequate lead time for manufacturers to achieve these low
levels of penetration. More discussion may be found in Section IV.B of the Proposed
Determination document at p. 49.
Summary and Response to Comments on the Proposed Determination
In their comments on the Proposed Determination, the Alliance stated that "EPA's response to
technology lead-time and adoption is inadequate and misleading." In their comments, the
Alliance disagreed that the promulgation of the 2022-2025 MY standards in calendar year 2012
constitutes the start of the lead time available to meet said standards. The Alliance further
commented that many of the key technologies considered by EPA for the Proposed
Determination are not currently in production. The Alliance specifically identified the non-
hybrid application of an Atkinson cycle engine with cooled EGR, cylinder deactivation, higher
compression ratio and ability to run on regular gasoline as an example of a key technology. The
comments further contend that single examples of technology do not reflect the wide diversity of
products produced and sold in the U.S.
In response, EPA notes that the Alliance and many of its members have commented at length
about the extensive research and development and capital investments required to meet future
GHG standards, while at the same time commenting that the Final Determination could have
been delayed by 18 months with no meaningful impact on the industry or the environment.
47
-------
Assuming a typical 5-year vehicle design cycle followed by a 5-year production life, most if not
all of the vehicle program cycles that have completed design and development to date will have
turned over prior to MY2025, with only the most recently initiated design and development
programs likely to produce vehicles that will be part of the MY2025 fleet EPA believes it is
appropriate to consider the promulgation of the standards in the 2012 FRM as a legitimate
starting point, and has provided the auto industry with a considerable amount of lead time.
Comments from the Alliance stated that insufficient lead time was available to implement a
non-hybrid Atkinson cycle implementation with high compression ratio, cooled EGR, and
cylinder deactivation. Similarly, Global Automakers commented that lead time to introduce a
new technology typically includes substantial time for initial development and integration prior
to introduction, and specifically criticized the example of Mazda's introduction of Atkinson
cycle engines as not including the time Mazda would have required to develop and integrate the
engine into its vehicle lineup. In response, EPA notes that the first manufacturer to implement an
advanced technology typically faces the greatest burden of resolving fundamental uncertainties
and evaluating viable approaches for implementing the technology, following a line of inquiry
that is also likely to be the subject of other researchers in the field at about the same time.
Subsequent applications of this technology cannot be expected to entail the same degree of
investment and risk in their development and integration, particularly now that production
Atkinson cycle engines are available for study and experimentation. EPA has provided additional
response regarding the lead time for this technology in Chapter 2.5.1 of this RTC document.
Global Automakers also commented that EPA was "prioritizing a single set of engine
technologies" by modeling the Atkinson 2 package as having significant projected penetration,
and requested an explanation of why EPA had "chosen to emphasize one technology over
another." In response, we note that EPA models technologies on a performance basis, and judges
their feasibility with respect to its assessment of the state of technology development. As stated
above and in the Proposed Determination, it is our assessment that Atkinson cycle technology is
a feasible option for compliance with the MY2022-2025 standards, in part due to the observation
that the steps required to implement it have become relatively modest with respect to the
building blocks already present in much of the MY2016 fleet. Projected technology penetrations
are a result of how a given technology competes with other available technologies for inclusion
in the cost-minimizing compliance fleet, and is therefore a result of the fleet compliance analysis
as a whole and is not the result of any specific prioritization or emphasis. In addition, for the
Proposed Determination EPA conducted a sensitivity analysis by artificially constraining
Atkinson 2 technology to 10 percent penetration. This sensitivity demonstrated that cost-
effective compliance paths using primarily other advanced gasoline engine technologies continue
to exist even under this scenario, at only modestly increased costs (see Section C. 1.2.1.4 of the
Proposed Determination Appendix at p. A-144 and p. A-147). EPA further discusses comments
relating Atkinson 2 technology in Chapter 2.5.1 of this RTC document.
General Motors commented that the Novation Analytics analysis performed for the Alliance
indicates that meeting the MY2025 standards would require the rate of improvement in engine
efficiency to increase to a rate double that of historical rates, and characterized this pace of
innovation as "unrealistic and unsubstantiated." GM further points to EPA's powertrain
efficiency projections in the TSD as being close to the necessary rate of improvement reported
by Novation. EPA disagrees that the rate of improvement reported in the TSD is unrealistic or
unsubstantiated. We note that the adoption of technology-forcing standards would be expected to
48
-------
lead to an increase in the rate of innovation. As we noted in Chapter 4.1.3 of the TSD at p. 4-8,
"in the absence of a forcing mechanism such as regulation ... manufacturers may prefer smaller,
incremental innovations," due in part to factors such as risk and uncertainty. Thus, historical
rates from a period without technology-forcing standards would not be expected to represent the
realistic potential pace of innovation under new standards. Figure 25 in the Novation report28
illustrates this, appearing to show a faster rate of improvement in powertrain efficiency from
2012-2014 (when the MY2012-2016 standards came into effect) than from 2005-2012 (a period
of flat regulatory stringency). In fact, the depicted rate of improvement appears to be similar to
or possibly greater than the 0.7 percent rate that Novation cites and which the commenter
considers unrealistic. Chapter 6 of the 2016 EPA Trends Report, which tracks technology
adoption across the industry as well as by individual manufacturers, also shows that the historical
fleet-average rate of technology adoption tends to mask the fact that individual manufacturers
often achieve high penetration rates of a technology very quickly (following the first application
of the technology by that manufacturer). The implication is that if some regulatory or market
force incentivized manufacturers to begin adopting technology at the same time, then the fleet
average would respond much more rapidly than the historical trend would suggest. Therefore,
EPA disagrees that historical rates of improvement are an accurate reflection of rates of
innovation that would apply to a time frame influenced by greater than historical stringency.
2.5 Individual Technologies
2.5.1 Atkinson Cycle Engine
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Appendix A.2.3.1 of the Proposed Determination and Chapter 2.3.4.1.8 of the TSD discuss
the Atkinson Cycle engine technology, particularly as applied to non-hybrid vehicle applications.
The TSD discussion includes an overview of the comments EPA received on this topic in the
Draft TAR and EPA's responses to those comments.
Comments received on EPA's Draft TAR assessment of Atkinson Cycle engines in non-
hybrid applications were primarily focused on the effectiveness estimates, asserting that EPA's
evaluation of the technology was overly optimistic. Specifically, the commenters stated that in
practice there are limitations of cooled exhaust gas recirculation (cEGR) for knock prevention,
that EPA's torque curve was incorrect, and that manufacturers do not have sufficient lead time to
adopt the technology at the penetration rates projected in EPA's compliance analysis. EPA's
responses in the Proposed Determination TSD Chapter 2.3.4.1.8.1 and TSD Appendix D
described the justification for the cEGR effectiveness benefits, and provide additional
clarification regarding an apparent misinterpretation by the commenters of materials published
and presented publicly by EPA. Those responses also indicate that the Atkinson Cycle
architecture, enhanced with cEGR and cylinder deactivation (DEAC) and with a higher
compression ratio, is already demonstrated both domestically (by both Mazda and VW) and in
Japan and Europe. EPA noted that engine modeling and initial hardware testing appear to show
synergies between the use of cEGR and DEAC with Atkinson Cycle engines. See TSD Chapter
2.3.4.1.4. TSD Appendix D also presented fuel differentiation maps documenting the minimal
effects of using different fuels (including a comparison of the technology's effectiveness when
28 Novation Analytics, "Technology Effectiveness - Phase I: Fleet-Level Assessment (Version 1.1)", October 19,
2015. Submitted to EPA Docket EPA-HQOAR-2015-0827.
49
-------
using Tier 2 and Tier 3 fuels). EPA also noted that commenters had failed to provide any data,
or description, of why DEAC could not be applied in conjunction with cEGR on Atkinson cycle
engines. Id. Chapter 2.3.4.1.8.3 of the PD TSD provided additional detail regarding the specific
design changes which manufacturers could use to adopt Atkinson engine technology (should
they decide to follow a compliance path that includes it).
With respect to the issue of lead time, EPA noted that many of the building blocks necessary
to operate an engine in Atkinson mode are already present in the 2016 fleet, including gasoline
direct injection (GDI), increased valve phasing authority, higher compression ratios, and (in
some instances) cEGR. EPA also explained that some of the potential packaging obstacles
mentioned in the comments, such as exhaust manifold design, should not be an impediment
because more conventional manifold designs (not requiring a revamping of vehicle architecture)
are both available and demonstrated in Atkinson Cycle applications. We responded that there is
sufficient lead time before MY2022 to adopt the technology and that it could be incorporated
without needing to be part of a major vehicle redesign. In addition, and as explained in the
Proposed Determination Appendix C. 1.2.1.4 and TSD Chapter 2.3.4.1.8.3, EPA conducted
sensitivity analyses constraining penetration of Atkinson-cycle engines and found that there are
other cost-effective compliance paths available which rely chiefly on advanced gasoline
technologies, and that a compliance path that includes lower penetrations of Atkinson cycle
engine technology need not result in high penetrations of electrification.
Summary of Comments on the Proposed Determination
EPA received several comments relating to our discussion of Atkinson Cycle engine
technologies in the Proposed Determination. In general, we received these same comments on
the Draft TAR, and we addressed them in the Proposed Determination Appendix and TSD as
stated above. In addition, we received the following new comments on the Proposed
Determination:
• Atkinson 2 (ATK2)29 penetration rate change from Draft TAR to Proposed
Determination (Toyota)
• Comments on several different aspects of Atkinson Cycle engine technologies
including technology benefits, lead time, and technology adoption rates (Alliance)
• Lack of a full-scale assessment of the research and development and manufacturing
costs that would be required to convert to this [Atkinson] engine technology (Global)
• ATK2 far less effective than EPA's predictions and not been sufficiently validated in
commercial production (FCA)
• ATK2 effectiveness with cEGR should be significantly higher than what was used by
EPA (ICCT)
Where appropriate, comments relating specifically to the issue of lead time for Atkinson cycle
technology were addressed in Chapter 0 of this Response to Comments document.
29 Atkinson 2 represents the application of Atkinson cycle engine operation in a conventional (i.e. non-hybrid)
powertrain architecture. See Appendix to the Proposed Determination A.2.3.1 at p. A-6.
50
-------
Response to Comments on the Proposed Determination
Toyota maintained that although the technology is deployed in approximately 7 percent of the
current fleet (consistent with EPA's estimates), the technology actually does not operate that
much of the time in Atkinson mode due to issues of control and power limitations without
providing any substantiating data. Toyota further stated (without supporting reference) that the
technology was insufficient to meet 2025 target levels. In response, EPA benchmarking of the
Mazda SKYACTIV-G using both chassis dynamometer and engine dynamometer testing
confirmed substantial use of Atkinson mode, particularly in areas of operation important for
compliance over the regulatory drive cycles. Some of this can be seen in the contour plot of
effective compression versus engine speed and torque in Figure 2.15 of the TSD which shows a
significant degree of Atkinson operation even at fairly high loads (e.g., peak effective
compression ratio of 11:1 vs. expansion ratio of 13:1) and illustrates how reduced effective
compression ratio is used during part-load operation of this engine to reduce the need for
throttling and reduce part-load pumping losses. Section 2.3.3.4 of the TSD also briefly
summarizes and cites a peer-reviewed EPA paper (SAE 2016-01-1007) that evaluated a Mazda
SKYACTIV-G engine using engine dynamometer hardware-in-the-loop (HIL) simulation of
vehicle operation. When the engine was evaluated during engine dynamometer testing using
HIL simulation of D-segment mid-size passenger car with a footprint of approximately 48 ft2, an
advanced 8-speed transmission, and moderate levels of road load reduction, the CO2 emissions
results were consistent with compliance with MY2025 GHG standards after application of AC
credit and with no further application of advanced engine technology beyond that of the 2014
Mazda SKYACTIV-G.30
The Alliance commented on several aspects of Atkinson engine technologies including
technology benefits, lead time, and technology adoption rates. The Alliance commented that
EPA had either disregarded its similar comments on the Draft TAR or had provided an
inadequate response. Contrary to the Alliance's assertions, each of these topics raised by the
Alliance in their comments on the draft TAR and again in their comments on the Proposed
Determination was carefully considered by EPA in the Proposed Determination analysis, and
each topic was fully addressed in responses to comments in the Proposed Determination
document and the TSD. See the Proposed Determination Appendix Section A.2.3.1 beginning at
p. A-7 and C.1.1.3.2 at A-132, and TSD Chapter 2.3.4.1.8.1 at p. 2-299 through 2-308.
In topic #22 of their comments, the Alliance confounds the ability to adopt Atkinson
technology with the effectiveness of the technology. The Alliance stated that EPA's response
"oversimplifies the requirements of introducing complex engine architectures to a large portion
of the new vehicle fleet." The Alliance then references a presentation made by Mazda regarding
the importance of the 4-2-1 exhaust system to the overall performance of the SkyActiv
technology and required packaging coordination with the SkyActiv Body. In response, EPA
recognizes the importance of the 4-2-1 exhaust system to the SkyActiv/Atkinson implementation
and also recognizes that the packaging of the exhaust system will require coordination with
ancillary vehicle systems. However, these types of packaging coordination activities are part of
the routine design and integration process for vehicle manufacturers, and these types of activities
30 Ellies, B., Schenk, C., and Dekraker, P., "Benchmarking and Hardware-in-the-Loop Operation of a 2014 MAZDA
SkyActiv 2.0L 13:1 Compression Ratio Engine," SAE Technical Paper 2016-01-1007, 2016, doi: 10.4271/2016-01
1007.
51
-------
are required independent of the technology pathway that each manufacturer has chosen. For
example, if a manufacturer chooses to implement a downsized turbo-charged engine in lieu of a
naturally aspirated Atkinson engine with 4-2-1 exhaust, the vehicle manufacturer will need to
adopt thermal management, exhaust manifold packaging, and NVH controls that will also require
coordination with the body, interior and electrical systems. Likewise, a manufacturer choosing
to adopt a cylinder deactivation solution will also require coordination with body and chassis
systems to avoid deteriorated NVH performance. EPA believes that the Atkinson technology is
not unique in its vehicle coordination requirements and that the Alliance has provided no
supporting data to support a different conclusion. On slide 17 of the same presentation
referenced by the Alliance, Mazda in fact discusses that the 4-2-1 technology is not new, and
while it did cause some challenges, Mazda was able to engineer robust solutions and package the
technology into very compact spaces, (e.g., Mazda2/Toyota Yaris, Mazda CX-3).
With regard to the benefits of Atkinson technology, the Alliance commented that EPA has not
yet demonstrated the projected benefit of an Atkinson cycle engine with cooled EGR and
cylinder deactivation and that EPA did not provide physical test results with a combination of
Atkinson, cooled EGR, and cylinder deactivation. In addition, the Alliance noted that there are
no current examples of Atkinson cycle engines being produced with cooled EGR and cylinder
deactivation. EPA's response to these comments remains unchanged from our response to
comments on the Draft TAR. Note the summary of our responses to those comments above and
the associated references. EPA provided physical engine dynamometer test results using a
combination of Atkinson Cycle and cooled EGR and results from engine testing conducted using
cooled EGR and Atkinson Cycle with physical deactivation of two out of four cylinders. Results
were presented in TSD Chapter 2.3.4.1.8.1 of the Proposed Determination and showed that
effectiveness used within the Lumped Parameter Model for this combination of technologies was
conservative relative to engine dynamometer test data. Data with cylinder deactivation was also
compared with published data from Mazda for one of their developmental engines using cylinder
deactivation.
The Alliance commented that many key technologies were not currently in production, e.g.
non-hybrid Atkinson cycle engines with higher compression ratio, cooled EGR, and cylinder
deactivation. This is similar to comments provided by the Alliance to the Draft TAR and EPA
responded in TSD Chapter 2.3.4.1.3 of the Proposed Determination (see in particular p. 2-290).
In response, EPA notes that Mazda presented data at the 2015 Vienna Motor Symposium from a
SKYACTIV-G engine with a cylinder deactivation system at an advanced stage of development.
The engine demonstrated effectiveness comparable to EPA estimates for applying cylinder
deactivation to ATK2 and comparable to EPA engine dynamometer testing of the SKYACTIV-G
with 2 cylinders disabled. Mazda has used cooled EGR with previous production applications of
their SKYACTIV-G engine, currently uses cooled EGR in the SKYACTIV Turbo engine in the
2017 Mazda CX9, and cooled EGR is currently used by Toyota and Hyundai in Atkinson Cycle
engines for both hybrid electric vehicle (HEV) and in non-HEV applications. At the 2017 North
American International Auto Show, Toyota announced that the base engine in the redesigned
2018 Toyota Camry would be Toyota's 2.5L 14 Dynamic Force Engine with a peak brake
thermal efficiency of 40%. The Toyota 2.5L 14 Dynamic Force Engine engine combines
Atkinson Cycle with cooled EGR and a dual PFI/GDI fuel injection system. In 2016, Toyota's
Camry model was the best-selling mid-size passenger car in the U.S. VW has already introduced
a 4-cylinder Miller Cycle engine, the EA211 TSI® evo, which combines cylinder deactivation,
52
-------
cooled EGR, early intake valve closing, and turbocharging. Miller Cycle is essentially a boosted
version of Atkinson Cycle.
Finally, with respect to the projected rates of adoption of Atkinson technologies, the Alliance
commented that EPA had not addressed the concern with auto manufacturers that had already
invested in other alternatives. The Alliance speculated that EPA was "unable or unwilling to
refute the Alliance's broader concerns directly." On the contrary, EPA believes that there is
sufficient time for companies to deploy the Atkinson technology should they wish to do so since
the technology can be implemented as a developmental extension of existing, current-production
engine hardware (e.g., GDI, dual camshaft phasing with high authority, increased intake port
tumble) in use in a large percentage of vehicles. As stated in the Proposed Determination, many
of the underlying technologies required to operate an engine in Atkinson mode are already in
production in many vehicles, including GDI, high authority valve trains, and higher compression
ratios. See also the discussion of lead time in Chapter 0 of this Response to Comments
document.
In addition, EPA's analysis shows that there are other cost-effective pathways to compliance
not heavily reliant on either ATK2 or electrification available for companies wishing to pursue
them, whether for reasons of commitment of resources or for marketing response. In an effort to
quantify the effects of different technology pathways, EPA conducted several sensitivity
analyses, including a scenario in which Atkinson penetration was capped at 10 percent, and a
scenario with reduced penetration of mass reduction. See Proposed Determination Appendix
C. 1.2.1.4. Overall projected cost did not vary significantly across these scenarios, ranging from
a low of $800 per vehicle across the fleet (primary analysis) to a high of $1,115 (still less than
projected in the 2012 FRM, a cost that the agencies found to be reasonable). Despite the artificial
constraints on certain technologies, including Atkinson, the overall cost of compliance remains
stable across these different pathways. EPA believes that these sensitivity analyses directly
respond to the Alliance's concern regarding vehicle manufacturers that have chosen a
compliance pathway that is different from that identified in the Proposed Determination. As
vehicle manufacturers choose among alternative pathways, EPA does not believe that their costs
will be significantly different from those projected in the Proposed Determination.
Global Automakers noted "the lack of a full-scale assessment of the research and
development and manufacturing costs that would be required to convert to this [Atkinson] engine
technology" and further suggested that more lead time would be needed to adopt it, pointing out
that Mazda had spent time developing and integrating the engine into its vehicle lineup which
EPA had not taken into consideration. These comments seem to be grounded in the assumption
that EPA has determined that Atkinson cycle technology is the only technology that will allow
the vehicle manufacturers to meet the future standards and that the EPA analysis and projected
technology penetrations are prescriptive. As previously stated in the Draft TAR and the
Proposed Determination, EPA believes that Atkinson cycle technology is one of several cost
effective powertrain alternatives available to vehicle manufacturers to meet the MY2025
standards. In EPA's analysis we have added technology in a cost effective manner to establish
future compliance (i.e., projected technology penetrations are a result of how a given technology
competes with other available technologies for inclusion in the cost-minimizing compliance
fleet). If a vehicle manufacturer is pursuing an alternative pathway for compliance, the EPA
analysis does not force a manufacturer into a different compliance solution, and the research,
development and manufacturing costs of conversion to an alternative pathway are not required.
53
-------
Honda maintained that larger engine displacement would be needed to deal with issues of
power loss and torque recovery. Use of a broad range of camshaft phasing authority and GDI
allows the degree of Atkinson Cycle to be varied across the speed and load range of the engine.
In Mazda's implementation of Atkinson Cycle with GDI as a replacement for PFI engines, peak
torque and rated power were improved relative to their PFI predecessors at the same cylinder
displacement. For example, when the Mazda 2.5L SKYACTIV-G engines replaced the previous
2.5L PFI MZR L5-VE engines in the Mazda3 and Mazda6, peak torque increased from 225 N-m
to 251 N-m and rated power increased from 125 kW to 138 kW. Peak torque for the 2.5L
SKYACTIV-G engine was also available at considerably lower engine speed than its PFI
predecessor (3250 rpm vs. 4500 rpm, respectively). As described in TSD 2.3.4.1.8.1, EPA
increased engine displacement by 5% in analyses used for the Proposed Determination for all
"advanced" ATK2 engine packages to which a 1-point increase in geometric CR and cEGR are
applied. This was done to reflect a reduction in peak BMEP and a resultant necessity for
increased engine displacement to maintain vehicle acceleration performance and was done to
maintain a conservative assessment of "advanced" ATK2 effectiveness. This adjustment resulted
in a decrease in LPM CO2 effectiveness for the proposed determination relative to the Draft TAR
of approximately 0.1 to 0.65%, with the range roughly coinciding with low and high power-to-
weight-ratio vehicles, respectively.
FCA commented that their own extensive internal analysis showed ATK2 to be far less
effective than EPA's predictions and also stated that ATK2 had not been sufficiently validated in
commercial production. The comments are identical to comments provided by FCA on the Draft
TAR and do not take into consideration EPA's response to the original comments, nor the engine
dynamometer and chassis dynamometer test data generated by EPA and summarized within the
Proposed Determination TSD showing that EPA's modeled effectiveness is conservative. FCA
did not provide engine dynamometer data, chassis dynamometer data, or modeling data to
support a lower effectiveness for ATK2.
ICCT comments cite a technology report on non-HEV Atkinson Cycle engines that estimates
effectiveness to be approximately double the effectiveness used by EPA for ATK2 with the
addition of cooled EGR in the proposed determination. ICCT also found that EPA's reported
effectiveness had reduced by approximately 3 to 5 percent for ATK2 with cooled EGR between
the Draft TAR and the Proposed Determination. ICCT based their estimate of 12.5 percent
effectiveness of ATK2 with cooled EGR in part on effectiveness derived from public data from
Mazda, Toyota, and Hyundai. In response, EPA notes that it is difficult to isolate individual
technology effectiveness as opposed to the effectiveness of a combination of technologies that
are part of an overall vehicle package. Atkinson Cycle effectiveness is also somewhat sensitive
to vehicle road load, transmission gear ratio spread and number of gears, all of which were
accounted for in EPA's ALPHA modeling. Based on ALPHA modeling results, we also would
not expect ATK2 to have identical effectiveness regardless of vehicle class. The effectiveness by
vehicle class shown by ICCT between the Draft TAR and the Proposed Determination directly
compares vehicle classes that are close to one another but that are not truly identical. Some
variation in effectiveness between the Draft TAR and Proposed Determination analyses may also
be due to differences in the combination of technologies applied to specific vehicle packages
rather than significant differences in ATK2 effectiveness between the two analyses. Engine
displacements for ATK2 with cooled EGR were increased by 5 percent between the Draft TAR
and the Proposed Determination in part in response to comments received from vehicle
54
-------
manufacturers regarding maintaining equivalent vehicle performance during operation on 87
AKI in-use regular-grade gasolines. EPA estimates the impact of this change on CO2
effectiveness to be significantly less than 1 percent, or considerably less than the 3 to 5 percent
differences reported by ICCT.
2.5.2 Turbo Downsizing
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Appendix A.2.3.2 of the Proposed Determination and Chapters 2.3.3.3.8 and 2.3.4.1.9 of the
TSD discussed engine turbo downsizing (TDS) technology. The discussion EPA provided in the
TSD included an overview of the comments EPA received on this topic in the Draft TAR and
EPA's responses to those comments. Comments received on the EPA's Draft TAR assessment
of turbocharged/downsized engines were primarily focused on the following topics:
• Choice of the Ricardo EGRB research engine as representative of 2025
turbocharged/downsized engine technology instead of the 2010 Ford 1.6L EcoBoost or
2014 Ford 2.7L EcoBoost engines
• The impacts of octane and relative CO2 effectiveness during operation of Tier 2
certification gasoline vs. Tier 3 certification gasoline or in-use regular-grade 87AKI
E10 gasoline, including presentation of the Alliance data from mid-level (E20, E30)
ethanol blended gasoline at different octane levels to represent impacts from E0 or
E10 gasoline at different octane levels
• The relative benefits of cEGR
• Impacts of differences in crevice volume
• Discussion of displacement/vehicle-mass (D/M) as a market-acceptance metric for
engine downsizing
EPA's responses in the Proposed Determination TSD Chapter 2.3.4.1.9.1 described the
justification for using the Ricardo EGRB V6 engine as the basis for evaluating
turbocharged/downsized engine effectiveness, explained the unrepresentative nature of the fuels
used in the study cited by the Alliance, and summarized CO2 emissions from EPA chassis
dynamometer testing of vehicles with turbocharged/downsized engines and other engine types
using Tier 2 E0 96 RON gasoline and Tier 3 El0 87 RON gasoline. EPA responses to comments
regarding crevice volumes and cEGR use are also included in TSD Chapter 2.3.4.1.9.1. EPA's
responses in the Proposed Determination TSD Chapter 2.3.3.3.8 discussed D/M of
turbocharged/downsized engines, including D/M in current vehicle applications and current
market acceptability of vehicles with D/M of less than 0.9.
Summary of Comments on the Proposed Determination
The Alliance commented on several aspects of turbo downsized technology starting with a
broad statement that EPA's assessment of gasoline turbocharged direct injection engines was
optimistic. The Alliance supported their comments through several comparisons including a
comparison of EPA's 1.17 L GTDI TDS24 engine configuration and the turbo downsized, GDI
engine applied by NHTSA in its Draft TAR analysis, and a comparison of the same EPA engine
to the current Ford 2.7L Eco Boost engine.
55
-------
Toyota commented that they had not received a response to questions regarding the rationale
for why 27 bar BMEP turbocharging was not analyzed as part of the Draft TAR or Proposed
Determination.
Response to Comments on the Proposed Determination
Regarding the inclusion of 27-bar peak BMEP turbocharged/downsized engines in the FRM
and limiting peak BMEP to 24-bar in the Draft TAR and Proposed Determination, EPA found
only a small incremental effectiveness improvement between 24-bar and 27-bar BMEP engines
with both engines using cooled EGR. The boosting requirements for 27-bar peak BMEP also
necessitated a more complex and higher cost boosting system relative to 24-bar BMEP (i.e., use
of sequential turbocharging at 27-bar BMEP as opposed to VNT at 24-bar BMEP). Engines with
24-bar peak BMEP have more potential for friction reduction than would be possible at 27-bar
peak BMEP due to the cylinder pressures required for higher BMEP and resultant connecting rod
and main bearing loads. EPA still expects turbocharged/downsized engines exceeding 24-bar
peak BMEP in the MY2022-2025 timeframe, particularly in limited production high-
performance vehicles and in some cases with the addition of variable compression ratio. Such
engines are already at an advanced stage of development or entering production (e.g., 2017
Honda Civic Type R, 2018 Infiniti QX50).
The Alliance's comments regarding EPA data described future, advanced
turbocharged/downsized engines such as TDS24 as having 10-20 percent lower fuel flow than
either the Ford Ecoboost 2.7L engine or the 24-bar BMEP turbocharged GDI engine modeled by
a subcontractor to Argonne National Laboratories (ANL) for NHTSA's Draft TAR analysis. The
Alliance described those particular areas of operation where the fuel flow differences were
greatest as important regions of operation over the regulatory drive cycles and thus the EPA
results were characterized by the Alliance as overly optimistic and not plausible.
Differences in fuel consumption or CO2 emissions between these types of engines are not
entirely surprising considering the different levels of engine technologies applied. Please refer to
the discussion of engine technology differences between the Ford Ecoboost 2.7L engine and
TDS24 in the response to Alliance comments to the Draft TAR in Chapter 2.3.4.1.9.1 of the
TSD. The TSD also included comparisons of data from both the TDS24 and the Ford 2.7L
Ecoboost engines with publicly available data from advanced turbocharged/downsized engines
from MY2017 VW and Honda light-duty vehicles. The EPA analyses and comparisons using
the VW and Honda turbocharged/downsized engines were not considered within the Alliance
analysis despite the fact that these current production engines are somewhat closer to EPA
TDS24 with respect to engine technology than the 2015 Ford Ecoboost 2.7L.
In considering the Alliance comments regarding EPA's modeling of drive-cycle CO2
emissions from advanced turbocharged/downsized engines, we compared TDS24, the Alliance-
referenced ANL 24-bar BMEP configuration, the Ford Ecoboost 2.7L, and the three additional
turbocharged/downsized engines mentioned in the Proposed Determination (VW EA211-evo and
EA888-3B, and the Honda L15B7) using ALPHA simulations of a vehicle configuration similar
to what was used within ALPHA simulation results provided by the Alliance in their comments
to the Proposed Determination. As stated previously in the TSD, these three additional engines
reflect more modern applications of engine technology by automobile manufacturers than the
Ford Ecoboost 2.7L. While still lacking some of the more advanced features of TDS24 (e.g.,
VVL, dual high/low pressure loop cEGR, centrally-mounted high-pressure piezo fuel injection),
56
-------
all three engines achieve comparable or higher efficiency over regions of engine operation that
are important for compliance with CO2 emissions standards. Two of the engines (VW EA211-
evo and EA888-3B) used Miller Cycle. One of the engines (VW EA211-evo) also uses 2/4-
cylinder deactivation, a VNT turbocharger and a third generation of high-pressure (350-bar)
direct fuel injection. The engines are also part of a growing trend towards moderate to
significantly higher stroke to bore ratio (S/B), which improves combustion thermodynamics and
in-cylinder turbulence generation. The EA211-evo and Honda L15B7 have S/B of 1.15 and
1.22, respectively, while the older Ford Ecoboost 2.7L design uses a S/B of 1.0.
The engines were all scaled to provide approximately equivalent torque between 1000 and
3500 rpm to TDS24 and equivalent vehicle performance for a low power-to-weight ratio, high
road-load (LPW-HRL) vehicle configuration similar to a CUV. This vehicle configuration was
chosen to be approximately comparable to the vehicle configuration summarized by the Alliance
within Table 1 of their comments on the Proposed Determination. Table 2-4 appearing below
summarizes the vehicle, engine and transmission combinations analyzed. The vehicle
configurations with turbocharged/downsized engines were also compared to an exemplar vehicle
with a 2.38 L naturally aspirated GDI engine and 6-speed automatic transmission to reflect an
approximately MY2015 level of technology. The examples with turbocharged/downsized
engines were configured with advanced 8-speed automatic transmissions, a 5% reduction in
vehicle mass, and 10% reductions in rolling resistance and aerodynamic drag to reflect a
moderate level of road load reduction for compliance with 2022-2025 LD GHG standards.
An initial analysis looked at fuel usage for a region of operation identified by the Alliance in
its comments as important for compliance over the drive cycle. This region of operation within
the Alliance's analysis showed 10-20% fuel differences and represented operation of TDS24 at
below 2-bar BMEP or approximately 19 N-m of torque (see the Alliance comments on Proposed
Determination, Figure 3). Figure 2-7 and Figure 2-8 show the contribution of mass of fuel
consumed at different torque points (i.e., torque at all speeds) encountered over the combined
FTP and HwFET cycles as well as cumulative fuel usage over the combined cycles for 1.17L
TDS24 and a Ford 2.7L Ecoboost scaled for equivalent performance (1.24L 3-cylinder).
Operation over the regulatory cycles at less than 19 N-m of torque accounted for less than 4% of
fuel used by TSD24 and less than 7% of fuel used by the Ford Ecoboost 2.7, and thus the sub-19
N-m region of operation with 10-20% fuel consumption differences would account for a
difference of approximately 1% over the drive cycle.
57
-------
U Combined Cycle Fuel Mass Consumed @ Torque Value
— Combined Cycle Cumulative Fuel Consumed (%)
Operation at <19 N-m
accounts for <4% of
combined-cycle fuel
consumed
50 100
Engine Torque (N-m)
i i i
BMEP (bar)
Figure 2-7 EPA TDS24 1.17L 13 Combined Cycle Fuel Consumption Distribution vs. Engine Torque
U Combined Cycle Fuel Mass Consumed @ Torque Value
Combined Cycle Cumulative Fuel Consumed (%)
50 O
Operation at <19 N-m
ITl
accounts for <7% of
-combined-cycle fuel
consumed
50 100
Engine Torque (N-m)
BMEP (bar)
Figure 2-8 Ford Ecoboost 2.7L Scaled to 1.24L 13 Combined Cycle Fuel Consumption Distribution vs.
Engine Torque
58
-------
Figure 2-9
1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000
Speed ( RPM )
BSFC and Combined-Cycle Fuel Consumption "Heat Map" for the EPA TDS24 1.17L 13.
Figure 2-10
o
1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000
Speed ( RPM )
BSFC and Combined-Cycle Fuel Consumption "Heat Map" for the Ford Ecoboost 2.7L scaled to
a 1.24L 13.
59
-------
Figure 2-11 BSFC and Combined-Cycle Fuel Consumption "Heat Map" for the Honda L15B7 scaled to a
1.32L 13.
1200 1400 1600 1800 2000 2200
Speed { RPM)
2400
2600
3000
Uncertainty in fuel flow measurements increases at very light loads due to the resulting low
fuel flow rates. Relatively small differences in differences in fuel flow measured at light-load
(i.e., low-flow) conditions appear large on a percentage basis even if absolute differences are
small and even if such operating conditions do not represent a significant contribution to cycle-
integrated fuel consumption or CO2 emissions. The relative importance of different speed/load
conditions on fuel consumed over the drive cycles can be visualized using "heat maps" such as
those shown in Figure 2-9, Figure 2-10, and Figure 2-11 for TDS24, and the scaled Ford
Ecoboost 2.7L and Flonda LI 5B7 engine configurations, respectively. Please note that the heat
maps were plotted over the same 1000-3000 rpm and sub-120 N-m torque operational range
shown within Figure 2 and Figure 3 of the Alliance comments to the Proposed Determination
and do not reflect the full range of engine operation. The "heat maps" indicate that fuel
consumption differences at higher load conditions than those identified by the Alliance (e.g.,
from approximately 40 N-m to approximately 110 N-m torque) would be significantly more
important with respect to drive cycle fuel consumption and CO2 emissions than operation at less
than 19 N-m of torque for the specific examples shown here and within the Alliance comments
to the Proposed Determination.
Figure 2-12 and Figure 2-13 are EPA's recreation of Figure 2 and Figure 3, respectively, from
the Alliance comments to the Proposed Determination showing fuel consumption differences
between 1000 - 3000 rpm and below 120 N-m of torque. Figure 2-12 shows a comparisons of
60
-------
the EPA TDS24 and the Honda L15B7 to the ANL 24-bar BMEP engine and Figure 2-13
compares the TDS24 and the Honda L15B7 to a scaled version of the Ford 2.7L Ecoboost.
When properly scaling the engines for equivalent torque, EPA's analysis found somewhat
larger differences at light loads between TDS24 and the scaled Ford Ecoboost 2.7L and
significantly larger differences relative to the ANL 24-bar configurations. Comparing the scaled
Honda L15B7 engine to the scaled Ford Ecoboost 2.7L and the ANL 24-bar configuration
showed remarkably similar differences to those found with the comparison to TDS24 despite the
Honda engine's lack of some of the technologies used for TDS24 (e.g., cEGR, VVL and VNT).
120
100
1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000
Engine Speed (RPM)
Figure 2-12 Percentage Differences in Fuel Consumption Between The 1.17L TDS24 and the MY2015 Ford
Ecoboost 2.7L Engines (left) and between the MY2017 Honda L15B7 and the MY2015 Ford Ecoboost 2.7L
Engines.
Note: Technology used by TDS24 but not used in the MY2015 Ford Ecoboost 2.7L includes higher pressure, centrally-
mounted piezo fuel injection; variable valve lift; VNT turbocharging; cooled low and high pressure external EGR;
and higher cylinder pressure capability. Technology used by the MY2017 Honda L15B7 TDS24 but not used in the
MY2015 Ford Ecoboost 2.7L includes higher pressure solenoid fuel injection improved S/B ratio (~1.2) for reduced
thermal losses and increased intake port tumble for improved combustion phasing. Both the Ford Ecoboost 2.7L and
the Honda L15B7 have large ranges (> 50 °CAD) of both intake and exhaust cam phasing authority, but EPA
benchmarking of both engines shows a larger range cam phasing, particularly at light load, for the Honda L15B7.
61
-------
1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000
Engine Speed (RPM)
60
0
1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000
Engine Speed (RPM)
Figure 2-13 Percentage Differences in Fuel Consumption Between the 1.17L TDS24 and the ANL 24-Bar
BMEP Engine Configuration (left) and between the MY2017 Honda L15B7 and the ANL 24-Bar BMEP
Engine Configuration.
Note: Technology used by TDS24 but not used in the ANL 24-Bar BMEP engine configuration includes higher
pressure, centrally-mounted piezo fuel injection; wider control authority for camshaft phasing; VNT turbocharging;
integrated exhaust manifold, and an additional cooled low pressure external EGR loop; Technology used by the
MY2017 Honda L15B7 TDS24 but not used in the ANL 24-Bar BMEP engine configuration includes higher pressure
solenoid fuel injection, wider control authority for camshaft phasing, improved S/B ratio (~1.2) for reduced thermal
losses and increased intake port tumble for improved combustion phasing.
The CO2 emissions results for a LPW-HRL ALPHA vehicle simulations using six
turbocharged downsized engines and a naturally-aspirated GDI exemplar configuration are
summarized in Table 2-4.The EPA TDS24 showed a 3.3% improvement relative to the Honda
engine and approximately equivalent combined cycle CO2 emissions relative to the VW engines
when comparing ALPHA simulations without start-stop. The EPA TDS24 had a 14.4%
combined cycle CO2 reduction relative to the ANL 24-Bar BMEP engine and a 9.4% CO2
reduction relative to the Ford 2.7L Ecoboost. The EPA simulation results with start-stop active
showed approximately equivalent results for the EPA TDS24, Honda L15B7, VW EA211-evo,
and VW EA888-3B, with EPA TDS24 having 14.2% and 7.5% CO2 reductions relative to the
scaled ANL 24-bar and Ford 2.7L Ecoboost, respectively.
As was found with the fuel consumption difference maps, the differences in CO2 emissions
from ALPHA vehicle simulations between the scaled Honda L15B7 and either the scaled Ford
2.7L Ecoboost or the ANL 24-bar configurations were comparable to differences relative to EPA
TDS24. The Honda LI5B7 had all .4% combined cycle CO2 reduction relative to the ANL
24-Bar BMEP configuration and a 6.3% CO2 reduction relative to the Ford Ecoboost 2.7L when
compared without start/stop. The EPA simulation results with start-stop active showed that the
Honda L15B7 had CO2 reductions of 12.8% and 6.1% relative to the Ford Ecoboost 2.7L and
ANL 24-Bar BMEP configuration. Again, the Honda L15B7 engine to the scaled Ford Ecoboost
2.7L and the ANL 24-bar configuration showed remarkably similar simulation results to those
found with the EPA TDS24 despite the Honda engine's lack of some of the technologies used for
TDS24 (e.g., higher BMEP, cEGR, VVL and VNT). Some of this may be due to the improved
S/B and increased variation of intake cam phasing (although similar levels of authority) for the
Honda engine relative to TDS24. Both the ALPHA simulation results and the BSFC maps of the
62
-------
2017 VW and Honda engines show that CO2 emissions comparable to TDS24 can be achieved
with the most recently developed production turbocharged/downsized engines and with less
application of advanced gasoline SI engine technology than projected for TDS24. TheCCh
vehicle modeling results achieved with TDS24 turbocharged downsized engines are not only
plausible, but are conservative. With application of further technologies (e.g., cEGR, VVL and
VNT) to recently developed engines like the Honda L15B7 engine, CO2 effectiveness would
likely be improved relative to EPA TDS24.
Table 2-4 ALPHA Model Inputs and Results for a LPW-HRL Vehicle Type (e.g., CUV).
Road load reductions reflecting a 5% reduction and vehicle mass and 10% reductions in aerodynamic drag
and rolling resistance were applied to the turbocharged/downsized vehicle configurations.
Engine (basis)
Base 2.5L 14
2017 VW EA211-fevo
2017 VW EA888-3B
2015 Ford 2.7L
E co boost
2017 Honda
L15B7 (CRV)
ANL 24-Bar BMEP
Turbo GDI
EPA TDS24
Engine Technology
NA, GDI,
DCP
GDI Turbo (17-bar
BMEP), IEM, DCP,
Miller Cycle, cEGR,
DEAC (2/4), 350-bar
solenoid Fl
GDI Turbo (20-bar
BMEP, IEM, DCP,
Miller Cycle
GDI Turbo (24-bar
BMEP), IEM, DCP
GDI Turbo (20-bar
BMEP), IEM, DCP
(>50° intake cam
auth.)
GDI Turbo (24-bar
BMEP), DCP,
DWL, HP cEGR
GDI Turbo (24-
bar BMEP), IEM,
DCP (50° intake
cam auth.),
CVVL, HP/LP
cEGR, 350-bar
piezo Fl
Displacement After
Scaling (L)
2.38
1.65
1.49
1.24
1.32
1.16
1.17
# Cylinders After
Scaling
4
4
4
3
3
3
3
ETW
3855
3677
3677
3677
3677
3677
3677
Road Load Coefficients
A
34.95
29.364045
29.364045
029.364045
29.364045
29.364045
29.364045
B
0.0875
0.0875
0.0875
0.0875
0.0875
0.0875
0.0875
C
0.02526
0.022734
0.022734
0.022734
0.022734
0.022734
0.022734
HP @ 50 MPH
13.66
12.08
12.08
12.08
12.08
12.08
12.08
Transmission
TRX11
6-sp. Auto.
TRX22
8-Sp. Auto
TRX22
8-Sp. Auto
TRX22
8-Sp. Auto
TRX22
8-Sp. Auto
TRX22
8-Sp. Auto
TRX22
8-Sp. Auto
Gear Ratio Spread
3.785-0.616
5.501-0.632
5.501-0.632
5.501-0.632
5.501-0.632
5.501-0.632
5.501-0.632
Final Drive Ratio
3.73
3.51
3.68
3.95
3.86
4.11
4.02
C02 (Combined, w/o
start-stop)
281.4
198.9
197.5
217.2
203.5
229.8
196.7
TDS24 % C02
difference w/o start-
stop
-30.1%
-1.1%
-0.4%
-9.4%
-3.3%
-14.4%
—-
C02 (With Truck AC
Credit, w/o start-stop)
257.0
174.5
173.1
192.8
179.1
205.4
172.3
C02 (Combined, w/
start-stop)
274.7
195.8
195.0
211.6
198.7
227.9
195.6
TDS24 % C02
difference w/ start-
stop
-28.8%
-0.1%
0.3%
-7.5%
-1.6%
-14.2%
—-
C02 (With Truck AC
Credit, w/ start-stop)
250.3
171.4
170.6
187.2
174.3
203.5
171.2
63
-------
2.5.3 Other Engine Technologies (Cylinder Deactivation, Cam Phasing, Variable Valve
Lift)
Summary of Comments on the Draft TAR addressed in the Proposed Determination
TSD Chapter 2.3.4.1 included a discussion of valvetrain technologies. The TSD discussion
included an overview of the comments EPA received on this topic in the Draft TAR and EPA's
responses to those comments. Comments received on EPA's Draft TAR assessment of
valvetrain technologies were primarily focused on the following topics:
• Cylinder deactivation (DEAC) effectiveness, operational area, and appropriateness of
using cEGR with DEAC
• The effectiveness of intake cam phasing (ICP) and dual cam phasing (DCP)
• The effectiveness of discrete variable valve lift (DVVL), and continuously variable
valve lift (CVVL)
EPA's responses in the Proposed Determination TSD Chapter 2.3.4.1.3 described the
justification for fixed-cylinder DEAC effectiveness and summarized the current production GM
application of cylinder deactivation used by EPA as a source of data for the operating range and
degree of activation of DEAC. EPA's responses in the Proposed Determination TSD Chapter
2.3.4.1.3 also provided current production examples of light-duty vehicle applications using
cEGR and Atkinson Cycle and a combination of cEGR, DEAC and early intake valve closing.
EPA's responses regarding the effectiveness of other valvetrain technologies were summarized
in TSD Chapters 2.3.4.1.4 through 2.3.4.1.7 and are supported by peer-reviewed published data.
Summary of Comments on the Proposed Determination
EPA received several comments relating to our discussion of valvetrain technologies in the
Proposed Determination. With the following exceptions, we received these same comments on
the Draft TAR, and we addressed them in the TSD as stated above.
Toyota suggested that several other technologies which have not yet been commercialized,
such as the combination of Atkinson 2 with high compression ratio, cooled EGR, electric
boosting, dynamic cylinder deactivation, and variable compression ratio, will likely be part of the
ongoing conventional ICE improvements and questioned whether they would be sufficient to
meet the 2022-2025 model year standards.
ICCT commented that several key technologies were not modeled, including e-boost, variable
compression ratio, and dynamic cylinder deactivation, and provided a series of citations in which
these technologies were examined in the public domain. It emphasized that "the single most
important factor in the accuracy of cost and benefit for projections is the use of the latest, most
up to date technology data and developments", and indicated that compliance costs would be
overstated without considering the most recently available technology developments.
MEMA claimed that EPA did not respond to its initial comment to the Draft TAR that EPA
revisit light-duty diesel data and analysis in the Final Determination. It pointed to a white paper
published by MARTEC as a source of updated information regarding light-duty diesel
effectiveness and cost estimates.
64
-------
Similarly, MECA "believes that light-duty diesel powertrains provide a cost-effective, durable
approach for vehicle manufacturers to improve the average fuel economy of their fleets,
particularly in the larger power category that includes small pick-up trucks and SUVs."
UCS commented that while EPA estimated an effectiveness for fixed cylinder deactivation of
3.9 to 5.3 percent in the Draft TAR, its application in the OMEGA and Lumped Parameter
models "seems to fall consistently below this range." It also commented that EPA's exclusion of
dynamic cylinder deactivation from its effectiveness analysis is too conservative.
Response to Comments on the Proposed Determination
Regarding Toyota's comment questioning the ability of conventional ICE technologies to
meet 2022-2025 standards, EPA acknowledges that some manufacturers may choose increased
hybridization and electrification as a compliance strategy that best suits their current positioning
in the marketplace. Based on its extensive modeling of all current and emerging technologies
expected to be commercialized, EPA illustrated in the Draft TAR one technical pathway (of
many that exist) that it believes to be cost-effective. Ultimately it will be up to the manufacturers
to determine the compliance strategy that best complements their future vehicle lineup.
In response to ICCT's comments on new technologies not considered, EPA acknowledges
that technologies continue to emerge in the marketplace; however, both detailed modeling of the
physical systems, as well as a rigorous cost assessment (e.g. teardown analysis), are required to
consider these technologies with the same level of robustness as the other technologies reviewed
in the Draft TAR. EPA agrees that including additional emerging technologies, when supported
by a rigorous performance and cost assessment, could provide even greater flexibility and
potentially lower compliance costs than the array of technologies already represented, and
acknowledges that its assessment might be conservative absent their inclusion in its analysis.
In response to MECA's and MEMA's comments concerning diesel powertrains: EPA
reviewed the MARTEC report. The MARTEC report reviews light-duty diesel efficiency and
fuel consumption improvements relative to advanced gasoline engines, but it does not take into
consideration the CO2 emissions control effectiveness of light-duty diesels. The carbon content
of diesel fuel is higher than gasoline on both a volumetric and mass basis. On a volumetric basis,
diesel fuel combustion emits 10.18 kg of CO2 per gallon versus 8.887 kg of CO2 per gallon of
gasoline31, which results in a 14.5% CO2 penalty for diesels relative to spark ignition gasoline-
fueled vehicles at equivalent fuel economy. Thus while vehicles with diesel engines generally
have improved fuel economy and thermal efficiency relative to vehicles with advanced gasoline
engines, vehicles with advanced gasoline engines can have comparable CO2 emissions due to the
reduced carbon content of gasoline relative to diesel fuel.
EPA also carefully considered new light-duty diesel engine technology developments that
have occurred after the FRM and the NRC diesel analysis. Chapter 5.2.2.11 of the Draft TAR
and Chapter 2.2.2.11 of the TSD to the Proposed Determination discuss the basis used for
determining the effectiveness of advanced light-duty diesel engines. The Lumped Parameter
Model and OMEGA were updated to take into account the CO2 effectiveness of Tier 3 compliant
light-duty diesels using dual-mode PCCI/diffusional combustion, higher peak BMEP, and
31 Diesel fuel C02 emissions factor is from Title 40 CFR § 600.113. Gasoline C02 emissions factor is from 75 FR
25324, May 7, 2010.
65
-------
advanced boosting systems based on data developed as part of the U.S. DOE and Cummins
ATLAS Research and Development Program. As part of an analysis for the Draft TAR, EPA
also commissioned a detailed tear-down study of a Tier 2-compliant light-duty VW diesel
engine. Approximately halfway through the study, both EPA and California compliance actions
determined that the engine used for the tear-down study was noncompliant with Tier 2 emissions
standards. The investigation of light-duty diesel compliance with federal emissions regulations
was also underway throughout the development of the Proposed Determination and is still
ongoing. While we agree with MECA's comments that light-duty diesels can continue to
comply with future emissions standards, the ongoing investigation complicates our ability to
determine an accurate bill of materials and costs for a truly Tier 3 compliant diesel emissions
control system. Furthermore, EPA's OMEGA results and the relative lack of diesel technology
should not be interpreted as an indictment of diesel technology. EPA fully expects that
manufacturers will continue to produce diesels for the US market and will do so because they
have determined that diesel technology provides them a more cost effective path to compliance
than that estimated by our OMEGA runs and because of the operational advantages of light-duty
diesel engines in specific applications (e.g., sustained high-load conditions from operation at
high loaded vehicle weights and/or loaded trailer weights in heavy-light-duty trucks and
MDPVs).
In response to UCS's comments about cylinder deactivation, EPA notes that in Chapter
2.3.4.1.3 of the TSD, EPA addressed the effectiveness of fixed cylinder deactivation, its
appropriateness in the context of several independent estimates, and its exclusion of rolling
dynamic cylinder deactivation in the Proposed Determination. More importantly, the individual
incremental effectiveness of any given technology (when incorporated into a vehicle technology
package within the Lumped Parameter model) is entirely dependent upon the other technologies
simultaneously applied to the vehicle. Frequently, technologies will provide a lower incremental
effectiveness than they would if they were applied to a vehicle with no other advanced
technologies present, due to the synergistic effects of multiple technologies addressing the same
physical losses.
2.5.4 Transmissions
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Appendix A.2.4 of the Proposed Determination and Chapter 2.3.4.2 of the TSD discussed
transmission technology. The TSD discussion addressed the comments EPA received on this
topic in the Draft TAR.
Commenters questioned the way EPA had assessed and classified automated transmissions in
the Draft TAR. The Draft TAR mapped all types of automated transmissions to a consistent
TRX32 transmission level. Commenters claimed the TRX designations were unnecessarily
complicated and did not recognize unique efficiencies of different transmission technologies.
EPA's responses in the TSD Chapter 2.3.4.2.1 noted that the TRX binning system was created in
response to industry comments, which correctly pointed out that transmission choice is based on
market and functional objectives, which may not always be the same as the most cost-effective
transmission selected by the OMEGA model. The TRX designations allow EPA to maintain the
type of transmission technology found in the baseline fleet during OMEGA modeling - which
32 TRX is a shorthand term EPA uses to designate transmission technology levels.
66
-------
reflects the market and functional objectives - rather than allowing the OMEGA model to default
to a seemingly more cost-effective transmission solution. Put another way, the TRX designation
implicitly assumes that manufacturers will likely maintain the transmission type already in the
baseline fleet for a specific vehicle, consistent with stakeholder comments. EPA acknowledged
that different transmissions have unique effectiveness, but stated that the effectiveness gains
between TRX levels will be similar (see TSD Fig. 2-116), and did not consider the additional
CO2 benefit gained from changing transmission type in its analysis. The agency's ultimate
finding on this issue was that the TRX classification system provided a reasonable means of
assessing technology cost and effectiveness while maintaining maximum manufacturer
flexibility.
Commenters also questioned the CVT effectiveness value EPA used in the Draft TAR,
specifically disagreeing with EPA's expectation for potential future effectiveness increases in
CVTs. In response to these comments, EPA in the Proposed Determination assumed that CVTs
would be aligned with a more advanced transmission type (in the TRX classification scheme,
CVTs would be classified as TRX21), a conservative assumption because it results in much of
the potential transmission improvement being included within the baseline (unlike the approach
in the Draft TAR, which classified CVTs in the baseline fleet as less advanced (TRX11)). Thus,
the classification used in the TSD recognizes fewer efficiency improvements to meet potential
standards. See TSD Chapters 2.3.4.2.1 and 2.3.4.2.2.
A comment from the Alliance disagreed with EPA's Draft TAR assessment of the
effectiveness of TRX11 transmissions. No further information was provided, but EPA
documented its basis for TRX11 effectiveness values in the TSD Chapter 2.3.4.2.2, and stands
behind its documented analysis.
Commenters also questioned the relative values for front- and rear-wheel drive transmission
effectiveness used by EPA in the Draft TAR. Specifically, commenters stated that packaging
difficulties in front wheel drive transmissions tend to increase spin and churning losses. EPA's
response in the TSD Chapter 2.3.4.2.2 clarifies that additional losses associated with the
differential were included when modeling transmissions.
A comment on EPA's Draft TAR assessment of transmission efficiency stated that industry
progress on transmission efficiency should be appropriately quantified in the baseline fleet.
EPA's response in the TSD Chapter 2.3.4.2.2 described how we quantified transmission
efficiency using baseline fleet transmissions. In addition, as explained above, in response to
Draft TAR comments on CVT effectiveness, in the Proposed Determination EPA reclassified
CVTs within the baseline as the more advanced TRX21. TSD Chapter 2.3.4.2.1 provided
additional information on the assumptions EPA made in the assessment of transmission
technology in the baseline fleet.
Comments on EPA's Draft TAR assessment of transmission efficiency stated that EPA's
estimated effectiveness differences between current six- and eight-speed transmissions were
high. These comments included reference to modeling results by Ford, and an assessment of
simulation differences between the EPA and Ford simulations. Chapter 2.3.4.2.2 of the TSD and
the Proposed Determination Appendix A.2.4 provide a discussion of these differences (among
other things, Ford assumed a gradeability metric which EPA believes is both inappropriate for
advanced eight-speed transmissions, and not necessarily present in production vehicles
67
-------
ostensibly designed to meet that metric), and why EPA regards the simulation results as
corroborative.
Comments on EPA's Draft TAR assessment of transmission efficiency and transmission
modeling stated that top gear gradeability should be maintained as a performance metric when
implementing advanced transmissions. EPA's responses in the TSD Chapter 2.3.4.2.2 provide
additional discussion that manufacturers are not currently maintaining top gear gradeability due
to the inherent advantages of advanced transmissions. Further discussion of comments received
on gradeability in the TSD is in Chapter 2.2.5 of this Response to Comments document.
Commenters also stated that manufacturers expected only marginal improvements due to
HEG233 (i.e., the additional effectiveness gain from TRX21 to TRX22), presenting an example
from FCA. EPA's responses in the TSD Chapter 2.3.4.2.3 and in the Proposed Determination
Appendix A.2.4 provide additional detail regarding corroboration of the HEG2 effectiveness
values, a discussion of the portion of HEG2 technologies represented by the FCA example
(indicating, among other things, that the FCA example reflected a partial use of the technology
and so was not optimized, and that efficiency values quoted by transmission suppliers are
consistent with those obtained by EPA), and why this information is consistent with EPA's
assumptions.
Summary and Response to Comments on the Proposed Determination
EPA received comments from stakeholders on effectiveness values assigned to transmission
technology. Specifically, General Motors stated that "after conducting the in-depth technology
analysis of the Silverado pickup and Malibu midsize car ... EPA errors are apparent, the most
notable being overstated fuel economy improvements attributable to transmission and/or gearbox
improvements." EPA acknowledges receiving comments claimed as confidential business
information (CBI) under 40 CFR Part 2 from General Motors detailing this in-depth technology
analysis. EPA considered this CBI information and believes the improvements attributed by
General Motors to transmission improvements are conservative, and do not reflect the same
extent of technology development used by EPA in TRX21 and TRX22 transmission packages.
Consequently, EPA stands by its analysis.
Additionally, FCA commented that "EPA's Lumped Parameter Model for the 2012 Rule
predicted approximately 13% improvements in 8-speed transmission efficiency over baseline for
all vehicle types. In the Draft TAR, the Lumped Parameter Model predicted an 8-speed
efficiency improvement of 30% for all vehicle types. In the Proposed Determination, however,
EPA's Lumped Parameter Model predicted improvements ranging from 28% to 55%." EPA is
unable to determine the basis for the commenter's assertion of 28-55% improvement. In the TSD
Chapter 2.3.4.2.3 (Table 2.85), the effectiveness associated with progressing from one TRX level
to another is presented. However, the average value of efficiency improvement from TRX11 to
TRX22 (representing moving from a nominal current six-speed transmission to an advanced
eight-speed transmission) is 13 percent, and does not reach 55 percent.
33 High Efficiency Gearbox technology level 2.
68
-------
FCA furthermore referred to EPA's testing of 5- and 8-speed Dodge Chargers in 2015 and the
SAE paper written detailing this testing and the associated analysis.34 The average value of 13
percent noted above is somewhat higher than the testing and analysis performed by EPA in 2015
(and referenced in the TSD and in FCA's comments) might suggest. However, the testing and
analysis included a number of conservative assumptions as detailed in the paper; for example,
the ALPHA modeling used as a basis for transmission efficiency projections predicted about 1
percent lower CO2 reduction than realized by the tested 8-speed Charger. In addition, since 2015
and through the TSD, EPA has continued to update its transmission data and ALPHA model with
the latest data available, as (for example) including the effect of a fast transmission warmup
within the TRX22 effectiveness numbers. Consequently, EPA stands by its analysis used to
inform the TSD, and believes it represents the best estimation of transmission effectiveness
available.
In commenting on disparities in vehicle-level transmission effectiveness between that
predicted by EPA's testing and the Lumped Parameter Model, FCA also notes, correctly, that
part of the difference between EPA's analysis of potential effectiveness improvements associated
with advanced transmissions is associated with corresponding engine downsizing to maintain
performance neutrality, and suggests that EPA dismissed comments on the Draft TAR that
criticized this approach. EPA clearly explained its rationale regarding this approach to
maintaining performance neutrality in the TSD Chapter 2.3.1.2, where we reviewed the
philosophical basis for maintaining performance neutrality, which was strongly supported by the
National Academy of Sciences — that technology comparisons should be made on the basis of
equivalent acceleration performance, such that the cost-effectiveness values of competing
technologies can be fairly compared.
The Alliance of Automobile Manufacturers also commented that EPA dismissed the Ford
"Transmission Walk" simulation, included as an attachment and referenced by the Alliance in
their comments on the Draft TAR. The attachment referred to by the Alliance walks through a
series of simulations to identify discrepancies between the respective technology assumptions
made by EPA and Ford. Contrary to the Alliance's statement, EPA believes that the simulations
are generally effective at identifying the differences between EPA's transmission technology
assumptions and Ford's transmission technology assumptions. However, as stated in Chapter
2.3.4.2.2 of the TSD, EPA disagreed with the assumptions made by Ford and the Alliance, and
continues to stand by its analysis. As there explained, there were important differences between
the simulation methodology used by Ford and that used by EPA, including the use of different
engines (the Ford simulation used a 2.0L EcoBoost engine, while EPA used a naturally aspirated
GDI engine), and differing lockup strategies between transmissions. In addition, the Ford
simulation assumed no changes in engine displacement whereas EPA applied a performance-
neutral downsizing strategy in its simulation. EPA continues to believe that this analysis (and
these differences) account for the difference in effectiveness percentages, since "effectiveness
percentages reported for transmissions paired with unimproved engines would be reduced when
the same transmission is paired with a more advanced engine." Id. p. 2-330. Moreover, this
34 Moskalik, A., Hula, A., Barba, D., and Kargul, J., "Investigating the Effect of Advanced Automatic Transmissions
onFuel Consumption Using Vehicle Testing and Modeling," SAE Int. J. Engines 9(3):1916-1928, 2016,
doi: 10.4271/2016-01-1142.
69
-------
discussion shows that EPA did not 'dismiss' Ford's comments, but carefully considered them
and explained why there was a difference in effectiveness estimates.
The Association of Global Automakers, in their comments, stated (p. 28, Global comments)
that EPA, in evaluating transmission effectiveness numbers, had used dynamometer tests on two
Dodge Chargers with identical powertrains to inform those evaluations, but had failed to identify
whether other aspects of these vehicles, which might have had an impact on the testing results,
were identical. As noted in the reference given in the TSD,35 the two vehicles had the same
engines, rear end ratios, and tires. Further information on test vehicles and process is contained
within the referenced paper.
The Association of Global Automakers also commented on EPA's "continued reliance on
DCTs," questioned "the assumptions underlying the percentage of CVT penetration," and
commented that EPA is not accounting for "the significant consumer acceptance issues
associated with [CVTs] " In response to earlier comments, as noted in Chapter 2.3.4.2.1 of the
TSD, EPA has implemented a TRX classification scheme where all conventional ATs (as well as
DCTs and CVTs) in the baseline fleet were mapped to three different designations: Null, TRX11
and TRX21. Any future transmission technology improvements are represented by advancement
through the TRX transmission levels. EPA assumes that all transmission types can be
represented by these TRX levels, so that progression through the levels requires only refinement
of a particular transmission, not wholesale movement to another transmission type.
The TRX designation system was implemented in response to earlier industry comments.
Those comments pointed out that transmission choice is based on market and functional
objectives, and on manufacturers' own analyses of transmission types, as Global commented.
Thus, manufacturer's choice of transmission type (CVT, DCT, conventional AT, or AMT) may
not always be the same as the most cost-effective transmission selected by the OMEGA model.
The TRX designations allow EPA to maintain the type of transmission technology found in the
baseline fleet during OMEGA modelling - which reflects the market and functional objectives —
rather than allowing the OMEGA model to default to a more (seemingly) cost-effective
transmission solution. Put another way, the TRX designation implicitly assumes that
manufacturers will likely maintain the transmission type already in the baseline fleet for a
specific vehicle, consistent with stakeholder comments. This designation system was
implemented in response to precisely the same concerns about specific transmission technologies
raised by Global in their comments.
The Union of Concerned Scientists strongly supported EPA's current classification of
transmissions into bins but noted that EPA's decision to classify CVTs in the baseline fleet as
TRX21 is clearly a conservative approach (as noted in Chapter 2.3.4.2.1 of the TSD). EPA
agrees this is likely a conservative approach; however, comments from stakeholders on the Draft
TAR indicate that classifying CVTs in the baseline as TRX21 is more reflective of potential
transmission improvement than a TRX22 classification, and thus EPA implemented the current
classification of CVTs in the TSD.
35 Moskalik, A., Hula, A., Barba, D., and Kargul, J., "Investigating the Effect of Advanced Automatic Transmissions
onFuel Consumption Using Vehicle Testing and Modeling," SAEInt. J. Engines 9(3):1916-1928, 2016,
doi: 10.4271/2016-01-1142.
70
-------
The Association of Global Automakers also noted that in the TSD, EPA makes the statement
"technology packages and vehicle classes where DCTs are applicable have been re-evaluated to
reflect manufacturers' current choices" but does not provide further explanation. EPA's intention
in the Draft TAR and TSD was to emphasize that, in response to comments received, EPA had
elected to implement a TRX classification scheme, which implicitly assumes that manufacturers
will maintain the transmission type they currently choose to implement.
General Motors commented that EPA's effectiveness estimates fail to account for the impacts
of "additional pump loading to accumulators needed to enable engine stop-start, electrical losses
associated with electrical auxiliary pumps to provide line pressure while the engine is off, and ...
vibration damping technologies that allow early Torque Converter Clutch (TCC) lock-up,"
further stating that these reduce effectiveness by adding inertia to the input side of the
transmission. EPA interprets this comment as applying primarily to the effectiveness values
assumed for stop-start technology, and addresses this comment in Chapter 2.5.7 of this RTC
document.
An anonymous citizen commented that GHG standards "reduce the motivation to produce
manual transmission vehicles," even though "true manual transmission cars provide safety
advantages that make them more attractive to some consumers," citing for example that they
discourage texting while driving and that their mechanical nature means that they cannot be
'hacked.' The citizen continues that "manufacturers should be allowed less stringent fuel
standards for true manual transmission cars." EPA addresses this comment in Chapter 3.2 of this
RTC document where we examine factors affecting availability of manual transmissions in the
market.
2.5.5 Battery Technology / Cost
This chapter reviews comments that relate specifically to battery technology and cost.
Comments that relate more specifically to non-battery costs are discussed in Chapter 2.5.6 of this
Response to Comments (RTC) document. Comments related to PHEVs and BEVs as GHG-
reducing technologies are discussed in RTC Chapter 2.5.10. Discussion of comments that relate
to electrification but not to the abovementioned technology issues, such as electrified vehicle
penetration rates and similar aspects of the Proposed Determination, may be found in Chapters
2.1 and 2.3 of this RTC document.
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Chapter 2.2.4.5 of the TSD discussed the state of battery technology for electrified vehicles.
Chapter 2.3.4.3.7 of the TSD develops battery cost estimates for PHEVs, BEVs, and HEVs, and
also discussed many of the comments on the Draft TAR that relate to battery technology.
A number of comments received on the Draft TAR related to EPA's projection of battery
costs for BEVs. Comments from two major OEMs appeared to generally support aspects of the
projected battery costs, or considered them to appear conservative. Comments from BEV-
specific manufacturers described both the projected battery costs and battery sizes as
conservative when compared to recent industry trends and forecasts. For example, Tesla Motors
stated that they expect to achieve battery costs by 2020 that are far below the 2025 Draft TAR
assumptions, and also that the battery capacity assumed necessary to achieve 200 miles of range
was overstated (see TSD Chapter 2.3.4.3.7 at p. 2-356 to 2-358).
71
-------
Comments suggesting that the projected costs and perhaps also the projected sizing for battery
packs were conservative, as well as continued collection of information on new products and
announcements since the Draft TAR, contributed to EPA's decision to update the projected pack
costs and sizes for the Proposed Determination analysis in order to reflect the latest information
available. This resulted in generally smaller projected pack sizes that more closely align with
production examples and also slightly lower costs per kWh for some packs than assumed in the
Draft TAR. More discussion of the observations and perceived trends in battery cost projections
which contributed to the decision to update the battery analysis are summarized generally in
Section A.2.5 of the Proposed Determination Appendix (at A-12 to A-13) as well as in TSD
Chapter 2.3.4.3.7.1 at p. 2-369. The updates and their effect on battery sizing and costs are fully
discussed in TSD Chapter 2.3.4.3.7 (Cost of Batteries for xEVs).
Other comments related generally to specific assumptions used for battery pack topology,
configuration, and assumed production volumes. EPA addressed these comments with
clarifications and additional rationale in TSD 2.3.4.3.7.4 at p. 2-387 to 2-388. In response to
other observations found in the comments, EPA also clarified certain aspects of the battery
analysis, such as, that packs are constructed of cells specifically designed for the power and
energy requirements of the vehicle to which they are assigned, that economies of scale are taken
into account on that basis, and that indirect costs associated with research and development are
included (see TSD Chapter 2.3.4.3.7 at p. 2-356 to 2-357).
Summary of Comments on the Proposed Determination
EPA received several comments relating to our projection of battery costs in the Proposed
Determination. Faraday Future, Union of Concerned Scientists (UCS), and the International
Council for Clean Transportation (ICCT) described EPA's projection of electric vehicle costs as
being overstated; since these comments can be interpreted as relating to both battery and non-
battery costs, they are addressed in Chapter 2.5.10.2 (Battery Electric Vehicles).
Global Automakers made several criticisms of EPA's battery cost analysis, raising issues
related to the assumed annual production volume and the references that EPA cited in assessing
its battery cost projections, and also suggested that EPA failed to consider cost information that
had been provided by manufacturers.
Global stated that EPA's MY2025 battery costs are based on a volume exceeding 400,000 per
year, which it contends no individual manufacturer will reach. As stated above, EPA received
and responded to comments on the Draft TAR relating to the use of this volume as an input to
the battery cost model BatPaC. This comment on production volumes mirrors these previous
comments, which EPA addressed in TSD Chapter 2.3.4.3.7.4 at p. 2-388 to 2-389.
Response to Comments on the Proposed Determination
Global also stated that EPA cited only one study and the findings of one manufacturer to
support the judgment that its Draft TAR battery cost projections appeared conservative.
However, this comment refers to a passage in TSD Chapter 2.2.4.2, which is an overview section
that previewed highlights of the detailed sections that follow it, and accordingly is not where the
analysis supporting the judgment was described, nor where primary references were cited. TSD
Chapter 2.2.4.5.9 examined the Draft TAR battery cost projections with respect to the Nykvist
and Nilsson study and the General Motors announcement of battery cell costs for the Chevy Bolt,
72
-------
supplemented by several references for estimating pack costs from cell costs (both the Nykvist
and Chevy Bolt references have been widely cited across the industry). The conclusion of this
analysis was further supported by discussion of several manufacturer comments received on the
Draft TAR, as well as several additional references, which (as described above) were fully
described in Section A.2.5 of the Proposed Determination Appendix (at A-12 to A-13) and in
TSD Chapter 2.3.4.3.7.1 at p. 2-369. These passages clearly show that EPA did not rely on only
one reference, nor were the findings of only one manufacturer extended to the entire industry.
Global also stated that its comments on the Draft TAR, as well as confidential business
information (CBI) supplied to the agencies, included examples of actual costs that EPA did not
use in its Proposed Determination analysis. This information included projections of cost per ton
of CO2 reduced for BEVs and PHEVs versus non-electric technologies. However, cost per ton of
CO2 reduced is not a source of battery or non-battery cost information, and cannot be used as an
input to EPA models, which require battery costs to be specified with respect to total battery
capacity and power specifications, and component costs to be specified in terms of peak power
and direct manufacturing cost at a specific volume. Similarly, EPA generally cannot directly use
CBI as a basis for inputs to a publicly available model, and, to preserve confidentiality, can only
use such information on a limited internal basis. On this limited basis, battery costs currently
being paid by manufacturers for current-technology battery components are certainly
informative, but ultimately are of limited utility considering that the goal is to project battery
costs not for today's state of technology and demand situation but for a more optimized and
developed industry state in 2022-2025. The General Motors Chevy Bolt disclosure remains the
only publicly available and widely cited reference for battery cell costs being paid by a volume
manufacturer that are represented by the source as applying to a time frame close to 2022-2025.
EPA received no other comments or CBI that included information of this type that could be
used as modeling inputs or directly compared with the projected costs of the Draft TAR or
Proposed Determination.
2.5.6 Non-battery Technology / Cost
This chapter reviews comments that relate specifically to non-battery technology and cost.
Comments related more specifically to battery costs are discussed in Chapter 2.5.5 of this RTC,
while comments related to PHEVs and BEVs in general are discussed in RTC Chapter 2.5.10.
Discussion of comments that relate to electrification but not to the abovementioned technology
issues, such as electrified vehicle penetration rates and similar aspects of the Proposed
Determination, may be found in Chapters 2.1 and 2.3 of this RTC document.
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Chapter 2.2.4.3 of the TSD discussed the state of non-battery technology for electrified
vehicles. Chapter 2.3.4.3.6 of the TSD discussed comments on the Draft TAR that relate to non-
battery technologies and costs.
A number of comments received on the Draft TAR related to EPA's projection of non-battery
costs for BEVs. Tesla Motors commented that their projected non-battery component costs are
"lower by double-digit percentages in every category versus the 2020 U.S. DRIVE figures
considered in the TAR," and that they see "significant room for further cost reductions [within]
the regulatory timeline covered in the TAR (2022-2025)." While clarifying that the Draft TAR
non-battery costs were not derived from the U.S. DRIVE targets that were mentioned in the
73
-------
Draft TAR, EPA also noted that more information would be needed to supplement these
qualitative comments in order to effectively evaluate the EPA non-battery cost projections with
respect to Tesla's experience.
Other comments were received that relate more specifically to battery costs than to non-
battery costs, and these are reviewed in Chapter 2.5.5 of this RTC document, while comments
that relate more to BEV and PEV overall costs are reviewed in RTC Chapter 2.5.10. No
additional comments were received that included sufficiently specific data with which the non-
battery costs used in the Draft TAR could be effectively adjusted, either to represent larger or
smaller volumes, or more or less optimized development programs (as mentioned by some of the
comments).
Summary and Response to Comments on the Proposed Determination
In comments on the Proposed Determination, Faraday Future, Union of Concerned Scientists
(UCS), and the International Council for Clean Transportation (ICCT) repeated their position
that EPA's projection of electric vehicle costs is overstated. Since these comments can be
interpreted as relating to both battery and non-battery costs (and therefore to overall vehicle
cost), they are addressed in Chapter 2.5.10.2 (Battery Electric Vehicles). EPA did not receive
additional comment pertaining specifically to non-battery costs, except as addressed as part of
the discussion in the previously mentioned chapters.
2.5.7 Stop-Start
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Chapters 2.2.4.4.1 and 2.3.4.3.1 of the TSD discuss stop-start technology. Additional
discussion of stop-start technology in the context of the off-cycle credit program is found in
Section B.3.4.1 of the Proposed Determination Appendix.
Public comments on the Draft TAR did not directly address the cost or effectiveness values
EPA used for modeling stop-start technology in the Draft TAR technology assessment. One
comment suggested that the effectiveness of stop-start could be improved beyond the value
assumed in the Draft TAR when implemented with a dual energy storage system. EPA addressed
this comment in Chapter 2.3.4.3.1 of the TSD, pointing out that EPA had acknowledged the
possibility of dual systems but chose to model more standard configurations for which data is
more readily available. Some additional comment was received in the context of the off-cycle
credit program, relating primarily to the ability of the existing credit values for stop-start to
represent increased real-world idle time or potential benefits of 48-volt hybridization. These
comments were addressed in Section B.3.4.1 of the Proposed Determination Appendix, in part
by pointing out that system effectiveness is also an important factor in relating idle time to
achieved benefits, and that systems vary widely in how much of the idle time the engine is
actually turned off.
Summary and Response to Comments on the Proposed Determination
As part of its comments on advanced transmissions, General Motors commented that EPA
modeling of stop-start technology neglected to account for energy demands related to auxiliary
electrical pumps and hydraulic accumulator to maintain line pressure and restart the engine,
which were said to result in added inertia to the input side of the transmission and additional
74
-------
losses. In response, it is acknowledged that stop-start may be implemented in a number of ways,
some with hydraulic auxiliaries and others without. The choice of auxiliary support may depend
in part on specific integration issues, such as NVH attributes of specific vehicle architectures or
engines on which the technology is proposed for inclusion, and the ability of the base engine to
perform combustion-assisted restart. Although EPA's effectiveness values for start-stop
technology in the 2012 FRM were based on simulations, the values used in the Draft TAR and
Proposed Determination were based on actual vehicle performance, derived in part from 2-cycle
certification test data for the Ford Fusion Stop-Start option, and corroborated by similar data
from the Mazda3 iStop, meaning that losses due to auxiliary processes such as torque converter
lockup and pumping that are applicable to a stop-start implementation on a 4-cylinder gasoline
engine were taken into account when developing the effectiveness values. Potential
implementations and the associated losses are likely to vary among manufacturers depending on
the specifics of the hardware approach they choose to implement and the needs of the engine and
other components with which it is integrated. The comment did not detail the magnitude of the
auxiliary losses GM anticipates with its particular implementation or the degree to which it might
differ from the production stop-start systems on which EPA based its effectiveness values, and
EPA has no evidence that the losses for various implementations that may be found across the
future fleet would vary dramatically enough to call these values into question as to their
representativeness.
EPA did not receive additional comments on the Proposed Determination that concern
modeling of stop-start technology. MEMA submitted comments related to off-cycle credits for
48 V stop-start technology. Off-cycle credits are addressed in Chapter 3.9 of this RTC document.
2.5.8 Mild Hybrid (48V)
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Chapters 2.2.4.4.2 and 2.3.4.3.2 of the TSD discuss mild hybrid technology. Chapter 2.3.4.3.2
of the TSD discussed comments on the Draft TAR that relate to this topic and the EPA responses
to those comments.
Some of the public comments on the Draft TAR relating to mild hybrids were directed toward
the decline in projected penetration of mild hybrids as compared to the 2012 FRM analysis, to
which EPA responded by characterizing the difference as a result of interactions among various
modeling changes to many technologies across the analysis, and not the result of any a priori
assumption about the potential for this technology to enter the market. Other commenters
suggested that the cost and/or effectiveness values EPA assumed for this technology in the Draft
TAR technology assessment were more optimistic than their own respective projections,
although these comments were not accompanied by supporting data and therefore could not be
evaluated. EPA's efficiency estimates are based on demonstrated performance. In addition, EPA
pointed to how efficiencies can increase when 48volt (V) mild hybrid technology is used in
combination with other technologies, such as an electric supercharger. Several commenters in
fact recommended that EPA more fully recognize 48V hybridization as an enabling technology
by accounting for additional flexibilities and synergies that can accompany its introduction. EPA
acknowledged that these potential benefits can provide value, although this value is difficult to
quantify. A battery supplier commented that battery costs for 48V mild hybrid systems appeared
to be overstated compared to their cost projections and assumed learning rates said to be
applicable to their own products, to which EPA responded in part by observing that the relatively
75
-------
low current penetration of 48V systems in the U.S. and worldwide continues to lend significant
uncertainty to the proper learning rate that should be assumed, and that the rate assumed by EPA
represents an appropriate value. More detail on these comments and their responses can be found
in Chapter 2.3.4.3.2 of the TSD.
Summary and Response to Comments on the Proposed Determination
In comments on the Proposed Determination, ICCT again commented on the value of
investigating synergies between 48V hybridization and e-boost, which can lead to similar costs
but higher effectiveness. These comments mirror comments received on the Draft TAR, which
included comments relating to potential synergies. EPA addressed these Draft TAR comments in
Chapter 2.3.4.3.2 of the TSD at p. 2-337. EPA again acknowledges that 48 V hybridization may
enable synergies that can lead to improved efficiency of other systems and hence of the
powertrain as a whole. However, similar to the rationale presented in TSD Chapter 2.3.4.3.1 at p.
2-335 regarding a recommendation that EPA model a dual-battery stop-start implementation,
EPA must reasonably limit the number of variations of technologies considered, in recognition of
available data on cost and effectiveness. Both detailed modeling of the physical systems as well
as rigorous cost assessment are required to consider additional technologies with the same level
of robustness as the technologies EPA already considers. As EPA has noted several times,
because we expect that the industry will continue to innovate and develop additional and
increasingly effective technologies, we are not able to consider every possible technology
combination that manufacturers may ultimately find cost-effective to include in their future
compliance paths. For example, as mentioned in Section IV.C of the Proposed Determination at
p. 54, EPA has not considered several technologies that are known to be under active
development, such as electric boosting, dynamic cylinder deactivation, and variable compression
ratio. While including such technologies might reduce projected compliance costs if they prove
to be more cost-effective than other technologies currently in the analysis, the lack of inclusion
of some of these technologies lends a conservative feature to the analysis supporting the
Determination.
The Alliance also commented that EPA asserted that no change is needed to give 48V mild
hybrids more off-cycle credit than stop-start. Comments related to off-cycle credits are addressed
in Chapter 3.9 of this RTC document.
2.5.9 Strong Hybrid
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Chapters 2.2.4.4.3 and 2.3.4.3.3 of the TSD discuss strong hybrid technology. Chapter
2.3.4.3.3 of the TSD includes discussion of public comments on the Draft TAR that relate to this
topic and the EPA responses to those comments.
Public comments on the Draft TAR relating to strong hybrids were primarily directed toward
the decision to model strong hybrids without reference to specific architecture (P2 or power
split), and the potential for differences in cost and effectiveness of the two architectures. EPA
responded to these comments by further describing and clarifying the rationale for this decision,
noting in addition that the cost and effectiveness of the two architectures appear to be converging
and that opinions continue to vary about their relative attributes. Another commenter agreed with
the effectiveness estimates for strong hybrids but described the cost estimates as more optimistic
76
-------
than its own projections, although supporting evidence was not provided. More detail on these
comments and their responses can be found in Chapter 2.3.4.3.3 of the TSD.
Summary and Response to Comments on the Proposed Determination
Comments received on the Proposed Determination did not raise additional issues related to
modeling of strong hybrid technology other than those EPA has already addressed through its
responses to comments received on the Draft TAR.
2.5.10 Plug-in Vehicles
This chapter reviews comments related to battery electric vehicle (BEV) and plug-in hybrid
electric vehicle (PHEV) technologies. Comments that relate more specifically to battery- and
non-battery technologies that are found on these vehicles are reviewed in Chapters 2.5.5 and
2.5.6 of this RTC document, respectively. Discussion of comments that relate to electrification
but not to the abovementioned technology issues, such as electrified vehicle penetration rates and
similar aspects of the Proposed Determination, may be found in Chapters 2.1 and 2.3 of this RTC
document.
Chapters 2.2.4.4.4 and 2.3.4.4.5 of the TSD review the state of technology for plug-in
vehicles (PEVs), which include plug-in hybrid electric vehicles (PHEVs) and battery electric
vehicles (BEVs). Chapters 2.3.4.3.4 and 2.3.4.3.5 of the TSD summarize the cost and
effectiveness assumptions for these technologies and also include discussion of many of the
related public comments on the Draft TAR, and EPA responses to those comments.
2.5.10.1 Plug-in Hybrid Electric Vehicles (PHE Vs)
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Few public comments on the Draft TAR concerned PHEVs specifically, as distinguished from
broader issues common to plug-in vehicles in general. Some comments that were peripherally
related to PHEVs were received in the context of credits, incentives, and flexibilities, which are
discussed in Chapter 3.9 of this RTC document. Discussion of comments that relate to other
aspects of electrification such as electrified vehicle penetration rates and similar aspects of the
Proposed Determination may be found in Chapters 2.1 and 2.3 of this RTC document.
One comment was received relating to emissions on cap removal from the pressurized fuel
tank that is commonly associated with PHEVs. EPA responded that, while it is well understood
that the dual-powertrain aspect of PHEVs can present challenges for control of cold-start,
evaporative, and cap removal emissions, such emissions are not directly within the scope of the
Midterm Evaluation and are more properly addressed in the scope of other emission control
programs that relate to evaporative emissions (see TSD Chapter 2.3.4.3.4). Additional responses
to Draft TAR comments relating to criteria pollutants and evaporative emissions were presented
in TSD Chapter 2.3.3.3.8 at p. 2-269 (see also Chapter 2.6 of this RTC document for a review of
comments on this topic).
Summary and Response to Comments on the Proposed Determination
Comments on the Proposed Determination did not raise new issues related to PHEVs
specifically. Comments from MECA repeated their comment on the Draft TAR, stating that "an
increase in PHEV sales ... may lead to an unintended increase in the VOC inventory" due to puff
77
-------
losses from pressurized fuel tanks. This comment relates more closely to the topic of evaporative
emissions and criteria pollutants, which is discussed in Chapter 2.6 of this RTC document.
2.5.10.2 Battery Electric Vehicles (BEVs)
A number of public comments on the Draft TAR concerned battery electric vehicles. The
comments described in this section relate specifically to battery electric vehicles as a GHG-
reducing technology, rather than to the battery or non-battery technologies these vehicles may
include. Comments related to the latter topics are described in detail in Chapters 2.5.5 and 2.5.6
of this RTC document. Discussion of comments that relate to other aspects of electrification such
as electrified vehicle penetration rates and similar aspects of the Proposed Determination may be
found in Chapters 2.1 and 2.3 of this RTC document.
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Comments related to BEVs were focused on several aspects of BEV modeling, including
driving range, projected fleet penetrations, aspects of cost such as overall cost as well as learning
and warranty rates assigned to BEVs, accounting for upstream emissions in compliance
projections, and power and acceleration levels.
One OEM commenter suggested that the 200-mile range of BEV200, the longest-range BEV
in the analysis, may not be sufficient to compete with conventional vehicles on driving range
over the long term, and that a longer range should be considered, which would increase projected
costs. EPA acknowledged in the Draft TAR and in its response that despite the fact that some
BEVs in today's market offer a range in excess of 200 miles, other near-term product
announcements continue to target an approximate 200-mile range, making it uncertain at best
that BEV200 will be as unrepresentative of future BEVs as the commenter suggests, or that
BEV200 will fail to compete adequately with conventional vehicles to achieve the modest
penetration rates projected in the Draft TAR analysis (see TSD Chapter 2.2.4.4.5 at p. 2-101 and
Chapter 2.3.4.3.5 at p. 2-344). In estimating the number of ZEV program vehicles to include in
the OMEGA analysis fleet, EPA also noted that it believes that the sales-weighted average
approach that was used is the most appropriate and fair way to make this estimation with
publicly available information, and this method would include the effect of longer-range vehicles
that are present in the fleet (see TSD Chapter 2.3.4.3.5 at p. 2-344).
Other commenters suggested that BEV penetration rates projected in the Draft TAR analysis
were too low. In some cases, this conclusion reflected the commenters' assertion that the
projected effectiveness of advanced gasoline technologies was overly optimistic (an assertion
with which EPA disagreed, and continues not to accept, see e.g. Section II.B of the Proposed
Determination document at p. 24, among other places). In contrast, other commenters posited
greater penetration rates on the expectation that BEV market share would grow rapidly for other
reasons, such as better cost-effectiveness than assumed, or the groundbreaking effect of near-
term product introductions, or better performance and convenience relative to conventional
vehicles. EPA noted that the projected BEV penetrations of the Draft TAR are not directly
chosen or selected, but rather result from the combined influence of many quantitative variables
representing cost and effectiveness of these technologies as well as others that compete with
them for inclusion in the projected compliant fleet. Similarly, market penetration that may result
from other influences beyond cost effectiveness, such as relative utility, brand appeal,
performance, or other factors are less tangible and quantifiable by their nature and present
78
-------
difficulties with their representation in a model that is driven primarily by cost effectiveness.
More discussion may be found in TSD Chapter 2.3.4.3.5 at p. 2-343 to 2-344.
With respect to comments expressed by a BEV manufacturer and others that overall BEV
costs and warranty cost reserves may be overstated, EPA noted that manufacturers that are
dedicated expressly to BEVs may experience different learning effects and cost structures than
other manufacturers, and that an accurate accounting of electrification costs during the time
frame of the rule should represent costs as they are likely to be experienced across the full
spectrum of manufacturers, even those that may utilize BEVs as a relatively small portion of
their compliance path. This is consistent with the relatively modest levels of BEV penetration
that the Draft TAR analysis projected, and the observation that significant uncertainty remains as
to warranty reserves or other aspects of indirect cost that will be representative of the industry as
a whole as it evolves. While EPA generally agreed that BEV costs appear to be continuing on a
downward trajectory, quantifying that trajectory in a manner sufficient to inform the applicability
of the non-battery cost estimates would require more detailed information than the comments
provide, such as detailed cost breakdowns and the assumptions that underlie them. EPA believes
that its current non-battery cost estimates continue to represent a reasonably conservative
assessment within the context of the modeling as a whole. For complete discussion of these
comments and responses, see TSD Chapters 2.3.4.3.5 at p. 2-343, TSD 2.3.4.3.6 at p. 2-346 to 2-
348, and TSD 2.3.4.3.7 at p. 2-357.
EPA also received some comments on the assumed production volumes for electrified
vehicles as being higher than anticipated by penetration levels projected in the Draft TAR. EPA
addressed comments related to the effect of assumed volumes for battery production in TSD
Chapter 2.3.4.3.7.4 at p. 2-387 to 2-388, where the rationale for the chosen volume is clarified
and expanded (see also Chapter 2.5.5 of this RTC document). With respect to assumed BEV
production volumes and penetration projections, much of the same rationale applies, as was
discussed in more detail in TSD Chapter 2.3.4.3.6 at p. 2-347 to 2-348.
In response to Draft TAR comments on the effect of phasing in of accounting for upstream
emissions of BEVs and PHEVs in the compliance analysis, for the Proposed Determination
analysis, EPA included upstream emissions for BEV operation and the electricity portion of
PHEV operation in the compliance determinations for all manufacturers by MY2025 (TSD
Chapter 2.3.4.3.5 at p. 2-344).
One Draft TAR commenter stated that the power levels assumed for plug-in vehicles (both
PHEVs and BEVs) were lower than that manufacturer typically provides in its vehicle line. EPA
responded to this comment in TSD Chapter 2.3.4.3.7 at p. 2-358, in part by acknowledging that
although different manufacturers may have differing targets for performance, this is also true for
many other vehicle attributes, and while it would be difficult to extend the analysis to represent a
specific manufacturer's performance levels, variations in vehicle performance are now modeled
more effectively in aggregate due to modifications in the vehicle classifications. EPA also
described its method for assigning motor sizing and 0-60 acceleration times for plug-in vehicles
in TSD Chapter 2.2.4.4.6, and its outlook on future trends in 0-60 acceleration times in TSD
Chapter 2.3.4.3.7.4 at p. 2-359.
Comments were also received on the subject of incentives for BEVs, including the incentive
multiplier for MYs 2017 through 2021, and the 0 g/mi accounting for tailpipe emissions for MYs
2017-2025 (subject to sales thresholds for MYs 2022-2025). Public comments received on these
79
-------
incentives and multipliers were addressed in Section B.3.4.2 of the Proposed Determination
Appendix, and are reviewed again in Chapter 3.9 of this RTC document.
Summary and Response to Comments on the Proposed Determination
Several comments on the Proposed Determination related to BEVs. Some of these comments
repeated points raised previously in comments on the Draft TAR and addressed by EPA in the
Proposed Determination, as summarized above.
Comments from Faraday Future were strongly in favor of the Proposed Determination, and
presented a number of arguments regarding the potential for rapidly growing BEV market
penetration independent of regulatory action, due in part to falling costs, consumer acceptance,
and other influences. The comments also argued that this projected increased penetration of
BEVs could justify amending the standards to make them more stringent. These comments
largely repeated the comments on the Draft TAR and cited several references that EPA had
incorporated into the Proposed Determination. EPA addressed these comments in the Proposed
Determination and TSD, as described above.
ICCT repeated its contention that EPA may be overestimating BEV costs, and this concern
was also raised in comments from UCS. EPA addressed this and similar comments which had
been submitted on the Draft TAR in TSD Chapter 2.3.4.3.6 at p. 2-346 to 2-347. UCS also noted
that any overestimation of BEV costs would probably not have a significant impact on projected
compliance costs for the time frame of the 2017-2025 rule, but felt it necessary to flag the issue
on the basis that overestimating BEV costs could have a greater impact on development of future
rulemakings. EPA agrees, and anticipates continuing to refine the characterization of BEV costs
as appropriate to the consideration of potential light-duty rulemakings applicable to the post-
2025 time frame.
2.5.11 Fuels / Octane
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Chapter 2.3.1.3 of the TSD discussed the new Tier 3 fuel and the properties of this fuel,
including aromatics, ethanol, and octane. The TSD discussion considered comments received on
the Draft TAR on this topic, and contains the EPA response to those comments. Further
discussion of the impact of Tier 3 fuel specific to engine technologies can be found in Chapters
2.3.4.1.8 and 2.3.4.1.9 of the TSD, as well as Appendix D to that document.
Comments received on the Draft TAR claimed that EPA's assessment of certain engine
technologies that are sensitive to fuel octane did not account for the lower octane of Tier 3 fuel
and the anticipated resulting degradation in efficiency. EPA's response in the Proposed
Determination TSD Chapter 2.3.1.3 described the test fuel properties for both Tier 2 and Tier 3
fuels before discussing the interaction of octane with both test cycles and real world driving.
The response explains that the assessment of technology effectiveness was not premised on a
requirement for high octane fuel for normal operation. In fact, two key engine technologies
tested for our analysis in support of the GHG standards, downsized turbocharged engines
(including Ford Ecoboost engines with up to 24-bar BMEP) and non-hybrid Atkinson cycle
engines, are currently produced by several manufacturers without a high octane fuel usage
recommendation from any manufacturer when operating under normal conditions. Current EPA
guidance provides an assurance at certification that vehicles not labeled as requiring premium are
80
-------
not octane sensitive under normal operating conditions, including all EPA test cycles.
Additionally, EPA's response in Chapters 2.3.4.1.8 and 2.3.4.1.9 of the TSD, as well as the
discussion summarized in Table 4.2 (in TSD Appendix D), provided test results supporting the
determination that two main engine technologies (turbocharged/downsized and Atkinson) show a
small reduction in CO2 when using Tier 3 regular grade octane E10 fuel.
Summary and Response to Comments on the Proposed Determination
In comments on the Proposed Determination, the Alliance commented that as of MY2020, all
testing for GHG compliance is required to use gasoline blended with ethanol, but that EPA's
modeling in the Proposed Determination, which applies to the full time frame of the rule, reflects
non-blended fuel. The Alliance suggests that EPA's response that the difference is not
consequential here to be incongruous with its continued modeling practice, and recommends that
all modeling should use the fuel mandated for the regulatory time period being evaluated.
EPA's response in Chapters 2.3.4.1.8 and 2.3.4.1.9 of the TSD, as well as the discussion
summarized in Table 4.2 (in TSD Appendix D), provided test results supporting the
determination that two main engine technologies (downsized turbocharged and Atkinson) as well
as several other engine technologies show a small reduction and not a penalizing increase in CO2
when using Tier 3 regular grade octane E10 fuel. This is because the change to the Tier 3 E10
fuels included a substantial reduction in aromatics resulting in lower carbon content of the fuel
and therefore producing lower CO2 emissions even in cases with an increase in fuel
consumption. Based on these test results, EPA concluded that modeling performed using the Tier
2 high octane EO fuel did not provide a GHG advantage over the Tier 3 regular grade octane E10
and that the Alliance concern that all modeling should use the E10 fuel required after MY2020
does not result in an increase in GHG stringency when manufacturers start using the E10 test fuel
in MY2020.
Several ethanol industry trade groups commented that the higher octane fuel represented by
mid-level ethanol fuel blends (20 to 40 percent ethanol) can improve engine efficiency and may
even be necessary to meet the stringent GHG emission standards. Specifically, they commented
that higher octane fuel would enable higher compression ratio engines which typically result in
engine efficiency improvements. Further, these organizations commented that EPA did not
consider fuels and vehicles together in our projection of potential technology paths for
compliance with the standards.
EPA appreciates the comments from the ethanol trade organizations regarding the potential
for engine efficiency improvements from high octane fuels. EPA disagrees with the conclusion
that high octane fuel is necessary to obtain the engine efficiency levels necessary to meet the
standards. This is evident in the agency's analysis of two primary engine technology options that
could be used to meet the standards: Atkinson cycle engines and turbocharged downsized (TDS)
engines. EPA's response in Chapters 2.3.4.1.8 and 2.3.4.1.9 of the TSD, as well as the
discussion summarized in Table 4.2 (in TSD Appendix D), provided test results supporting the
conclusion that TDS and Atkinson technologies do not require premium fuel to achieve the
standards. This is largely because both of these technologies if properly implemented are able to
optimize engine efficiency over a varied range of operation by either adjusting the effective
compression ratio in the case of the Atkinson engines or by adjusting the amount of cylinder
charge in the case of turbocharged downsized engines, particularly as it applies to real world and
test cycle operational constraints.
81
-------
EPA does not disagree that high octane fuel may provide some additional performance
improvements beyond what is required to meet the standards, such as increased engine
performance levels when operating on high octane, as explained in some manufacturers' owner's
manuals (EPA discussed this and related points in TSD Chapter 2.3.1.3 atp. 2-211). High
octane fuel may also help address some shortfalls in FE and GHG emissions in real world
situations where some technologies may be particularly sensitive, such as high load under towing
conditions or high ambient temperature operation in summer seasons. Under these and other
extreme conditions, some manufacturers are already recommending use of premium high octane
fuel in currently produced vehicles. Note that EPA does not preclude vehicle manufacturers
from using a high octane fuel for certification of GHG emissions if they properly label vehicles
as "premium required." In addition, premium gasoline is widely available in the US market. The
Tier 3 certification fuels include a high octane E10 fuel for "premium required." in addition to a
high octane E85 fuel for flex fuel or dedicated alternative fuel vehicles. Manufacturers also have
a pathway to certify a vehicle using a mid-grade ethanol fuel (presumably with high octane
properties) if they determine this to be a pathway for compliance with GHG emission standards.
As discussed in the TSD Chapter 2.2.2.14 at p. 2-44, EPA looks forward to reviewing the results
of the Department of Energy research project "Co-Optima" for potential future options for
engines and fuels, including results from high octane mid-level ethanol fuel blends.
Several comments from the ethanol trade organizations highlighted the issue of updating the
R-factor term used in fuel economy calculations, on the basis that the underlying data used to
determine this adjustment was based on outdated 1975 test fuel. As discussed in TSD Chapter
2.3.1.3, consistent with its historical practice, when test fuel properties are updated EPA
determines appropriate test procedure adjustments, which include potential updates to R-factor,
in order to maintain the same level of stringency of the GHG and fuel economy standards
between the Tier 2 EO and Tier 3 E10 fuels. This work is currently under regulatory development
in consultation with vehicle manufacturers, industry and other stakeholders.
2.5.12 Mass Reduction
Summary of Comments on the Draft TAR addressed in the Proposed Determination
The analysis performed for the 2012 FRM assumed that all vehicles in the baseline start with
the same opportunity for mass reduction. For the Draft TAR analysis, EPA revised this
assumption by characterizing differences in the incremental cost and feasibility of additional
mass reduction between vehicles in the baseline. We received comments on the Draft TAR
which, while generally supportive of the direction of the change, expressed concern that we had
not properly accounted for the amount of mass reduction already implemented in the baseline
fleet. One specific example provided by the commenters involved the value of 200 lbs. assumed
in the Draft TAR to account for the mass of AWD/4WD systems. EPA maintained this
assumption for the Proposed Determination, noting in Chapter 2.3.4.6 of the TSD that the
vehicles examined are typical of the AWD/4WD vehicles within the fast-growing crossover
utility segment, and that while this weight may under-represent some of the largest 4WD
vehicles, it may also over-represent some of the smallest AWD vehicles.
FCA commented that the effectiveness estimates made by EPA for mass reduction were not
accurate due to the lack of consideration of Equivalent Test Weight (ETW) class bins and their
effect on fuel economy testing. That is, the test method used for fuel economy certification uses a
nominal ETW value rather than the precise actual weight of the vehicle. FCA recommended that
82
-------
EPA adjust its modeling so that mass reduction benefits are only reflected via the resulting
change to ETW class bin. As discussed in Chapter 2.3.4.6 of the TSD, EPA did not agree with
this recommendation. The average mass reduction projected in the Proposed Determination is
approximately 9 percent, which would move many vehicles in the fleet down by one or two
ETW bins. EPA stated its belief that the approach of allowing mass reduction in continuous
increments (actually 0.5 percent increments in the OMEGA analysis) does not cause a systemic
underestimation of costs, because cases where manufacturers may be getting less benefit from
mass reduction than projected in our analysis would be offset by other cases where
manufacturers may be getting more benefit.
Summary and Response to Comments on the Proposed Determination
The Alliance reiterated (the Alliance comments, p. 31) that OMEGA modeling for mass
reduction is inconsistent with the mandated test method and is therefore improper. Specifically,
the issue raised previously by the Alliance and FCA in comments to the Draft TAR that the
regulation assigns discrete bins whereas the OMEGA modeling uses a continuous function. The
Alliance also finds EPA's explanation that the modeling does not result in systematic over- or
under-estimation of costs to be without support. In response, EPA disagrees that the use of a
continuous function rather than discrete binning necessarily results in significant distortion of
costs. First, a portion of the effectiveness benefit of mass reduction is due to the influence of
vehicle mass on coast down coefficients, and this benefit is accounted for in EPA's analysis
method consistent with the certification process regardless of whether discrete or continuous test
bins are used. Second, EPA maintains the position presented in the Proposed Determination
regarding the use of continuous as opposed to discrete bins - the continuous method is a
simplified approach that, although not structurally identical to a binned approach, is likely to
generate a positive error for some individual vehicles and a negative error for others, which on
average across the fleet would be expected to cancel each other out, resulting in no systematic
net error in the aggregated result.
In contrast, ICCT commented that EPA "systematically underestimates" mass reduction
potential in the fleet, and provides 15 examples of models that have reduced mass by 4 to 15
percent, citing many co-benefits to consumers. UCS also characterized the projected levels of
mass reduction as conservative. EPA agrees with the ICCT comment that a number of vehicle
models in the Proposed Determination analysis have achieved more mass reduction than the 9
percent average penetration projected, but this does not necessarily indicate that potential for
mass reduction to contribute to compliance pathways has been systematically underestimated, or
is overly conservative as suggested by UCS. Although EPA believes that it is feasible for many
vehicles to achieve up to 20 percent mass reduction, the projection of only 9 percent average
mass reduction in the cost minimizing compliance pathway indicates that when this technology
competes with the full array of other available GHG reducing technologies, few vehicles are
projected to apply the maximum level of 20 percent.
Global Automakers commented on the increase in average projected mass reduction from 7
percent to 9 percent. Global stated that this was a significant change in mass reduction and
evidence of a wide variation in EPA's results. Regarding the Global Automakers comment, EPA
noted in the Proposed Determination Table IV.5 and TSD Appendix C.1.1.3.1 at p. A-132 that
the reference point for presenting mass reduction technology penetration values was revised from
using the baseline value (as in the Draft TAR) to using the null technology package. The
83
-------
increase from 7 to 9 percent is therefore primarily the result of reporting the estimated mass
reduction already present in the baseline. The 2 percent difference simply reflects differences in
how mass reduction in the baseline is presented in the technology penetration tables between the
Draft TAR and the Proposed Determination.
Honda provided a post-Draft TAR presentation to EPA in November, 2016. While the
presentation was not in response to the Proposed Determination, Honda provided a redacted
version of the presentation to the Docket when submitting its public comment (which directed
EPA to the comments from Global Automakers for which Honda indicated support). In the
presentation material, Honda indicates that EPA underestimates the amount of mass reduction
already contained in the Honda baseline fleet, and that the resultant projected cost of compliance
for Honda is therefore not correct. EPA has considered this information but does not agree that
the incremental costs for additional mass reduction beyond what exists in the Honda baseline
vehicles have been underestimated, nor does EPA agree that the feasibility of additional mass
reduction has been overestimated. Honda is correct in its interpretation of how the baseline mass
reduction was established; however, they do not consider that the curves used to calculate the
cost of mass reduction were directly informed by the detailed engineering study based on a
teardown analysis a 2011 Honda Accord. To the extent that the Accord already incorporated
lightweighting materials and a mass-efficient design in MY2011, then the resulting cost curve,
based on reducing the mass of the Honda Accord, would reflect higher costs.
FCA provided several comments relating to mass reduction in general and the levels of mass
reduction EPA projected for its fleet, as support for their contention that the technology
pathways projected for FCA are unrealistic.
The FCA commenter noted that, between the Draft TAR and Proposed Determination, the
projected percent mass reduction for FCA's fleet increased from 5.9 percent to 11.6 percent, and
stated that this level of mass reduction would "require complete product redesigns in less than
eight years." With regard to the increase in projected mass reduction, EPA addresses a similar
comment with respect to the mass reduction projections for Mercedes-Benz in Chapter 2.8 of this
RTC document. That response also applies to the FCA projections. In addition, the response to
the Global Automakers comment on the increase in average projected mass reduction from 7
percent to 9 percent, provided earlier in this chapter, should also be considered when comparing
mass reduction projections between the Draft TAR and Proposed Determination.
With regard to the need for complete redesigns, EPA notes (in addition to the responses
above, which also carry implications for the interpretation of the new projections) that the mass
reduction projections are only part of one potential compliance pathway. As discussed in
Appendix C of the Proposed Determination at p. A-144, EPA performed a sensitivity analysis in
which no additional mass reduction is applied beyond that included in the projected baseline
fleet. As discussed at p. A-146 of that document, this analysis showed that on the fleet level, the
incremental cost per vehicle result is not heavily dependent on mass reduction, and that cost-
effective compliance pathways continue to exist.
FCA stated that "even small percentage weight reduction targets are challenging for
automakers to realize given the offsetting customer demand for additional content/features and
regulatory safety requirements." In response, EPA has acknowledged that content and safety
requirements can compete with fuel economy for the benefits of mass reduction, and that in the
past manufacturers have used mass reduction to offset the effect of added features or safety
84
-------
measures. For example, see the TSD at p. 2-146 and p. 2-156. Consistent with the approach used
by EPA for other technologies to estimate technology costs and effectiveness while holding
vehicle utility and performance constant, EPA has also made it clear that the cost curves used in
its analysis are entirely applied towards a reduction in vehicle curb weight, as opposed to
offsetting mass increases from the addition of content and features. See Draft TAR at p. 5-368.
EPA has also acknowledged that other regulatory requirements may add weight, and in
composing the baseline for the Draft TAR, accounted for new finalized safety regulations that
were deemed to have an impact on mass (as discussed for example in Chapter 5.3.4.6.2 of the
Draft TAR). Given the conclusion of the mass reduction sensitivity discussed above, EPA
believes that accounting for the effect of future safety requirements would be unlikely to change
the conclusions of our analysis.
FCA also stated, "It is generally accepted that... relative to cost, the percent mass reduction
exhibits linear behavior up to 10% and beyond 10% exhibits exponential behavior," and cites a
September 2016 study by the Center for Automotive Research as support. EPA notes that the
commenter does not specify whether the cost reference is to cost per unit mass removed, or to
total cost, making it difficult to evaluate this comment because the cost unit affects the form of
the curve. But in general terms, the cost curves EPA uses in its analysis (see Draft TAR Figures
5.126 and 5.127 at p. 5-380 for examples) are consistent with the general cost curve presented in
a previous study by the same author of the cited study, which was reproduced in the Draft TAR
as Figure 5.116 at p. 5-368. They generally show increasing costs for increasing percentages of
mass reduction, ranging from a cost savings at lower levels of mass reduction, transitioning to
increasingly higher costs at increasing levels of mass reduction. EPA does not find clear support
in the cited study for the statement that industry and academia generally accept that the cost of
mass reduction transitions from linear to exponential at 10 percent. Rather, the commenter may
be suggesting that a certain degree of mass reduction is achievable at a reasonable cost, but that
the fleet average level projected for FCA lies beyond an exponential transition and therefore is
unreasonable. EPA does not find this argument persuasive. The EPA cost curves exhibit the
expected behavior of assigning increasing rates of cost to increasing levels of mass reduction.
The form, shape, and magnitude of these curves were established by means of rigorous and
detailed empirical studies as described in Chapter 5.3.4.6 of the Draft TAR.
American Iron and Steel Institute (AISI) stated that EPA had not addressed its Draft TAR
comments on the cost of insurance and repairs for lightweighted vehicles and on the impact of
lightweighting on business models and infrastructure. In their Draft TAR comments, AISI stated
that collision repair costs "will have an impact on the yearly vehicle insurance costs, with an
initial cost estimate of $396 in 2010 dollars" and attributed that to the Draft TAR. However,
AISI misunderstood that $396 value which was instead an estimate of the average value of
collision insurance, not an estimate of incremental cost increases resulting from the standards.
Importantly, EPA does consider the increased cost of insurance in the payback analysis—
increased costs meant to address the possible increase in collision-related repairs on the higher
cost vehicles under the standards. As described in Chapter 3.11.2 of the TSD (at p. 3-53), EPA
considered the impact of the standards on auto insurance expense by considering the cost of
insurance as being proportional to new vehicle price. To the extent that vehicles that incorporate
advanced technologies experience an increase in average vehicle price, projected insurance costs
are increased in proportion. Similarly, as described in TSD Chapter 2.3.2.3.2 at p. 2-230, EPA's
analysis accounts for the costs of repairs covered by manufacturers' warranties. The indirect cost
85
-------
multipliers (ICMs) applied in EPA's analyses include a component representing manufacturers'
warranty costs. Also as described in that section, sufficient information is not available to
quantify the frequency and cost of different types of repairs. The projected cost of mass
reduction technology does not include a factor to represent potential differences in repair costs of
various materials, in part because there is little data available to reliably characterize the
existence or magnitude of repair cost differences for each specific material that might be applied
in the future. It not clear that the implementation of lightweight materials will necessarily result
in higher repair costs, and simply drawing a conclusion on the direction any impact of a
materials change on future overall repair costs may be unwarranted. For example, in the
particular case of the vehicle body, repairing a panel made from aluminum will involve different
(and possibly costlier) techniques than are used for a steel panel, but the necessity for repair may
be reduced due to improved corrosion and dent resistance. Furthermore, the future cost of repairs
is difficult to predict in part because, as new materials gain increased usage in the fleet, repair
facilities and expertise will have to gain experience and capacity for performing the repairs, a
process which is likely to impose higher costs in the short term but which will gradually fall over
time.
The AISI comments on the impact of lightweighting on business models and infrastructure
appear to relate primarily to a concern that automakers should remain free to choose materials
that appropriately support the full spectrum of considerations relevant to their business model,
and not be directed towards higher-cost materials that may have unintended environmental and
other consequences. The implication is that the assessment of potential lightweighting materials
that EPA presented in the Draft TAR and Proposed Determination may act as guidance to
automakers as to which materials to use, and that if the assessment is not accurate as to relative
material costs and capabilities, automakers may fail to select the most appropriate materials.
However, the EPA assessment is not intended as prescriptive guidance to manufacturers
regarding which particular materials to use, and it seems unlikely that a manufacturer would
simply defer to the conclusions of the assessment without considering the specific needs and
goals of its business model and market segment. AISI also suggests that the assessment is biased
toward replacing steel with lower-density materials such as aluminum and other materials. The
cost curves that EPA uses to estimate the cost of lightweighting are derived from assessments of
the costs of achieving various levels of mass reduction by means of various types of materials,
including steel, aluminum, and other materials. EPA's projected cost-minimizing compliance
pathway includes varying levels of penetration of mass reduction technology for different
vehicles. The assessments that form the basis of the cost curves suggest that steel is likely to play
a strong role in achieving target mass reductions at the lower end of the curve, while lower
density materials are likely to play a strong role in achieving target mass reductions at the higher
end of the curve. In the Draft TAR, EPA extensively cited examples of successful applications
of advanced high strength steels, which clearly demonstrate the capabilities and potential cost-
effectiveness of this family of materials. In all of the technology assessments performed for the
light-duty GHG standards, EPA has assessed technologies on a performance basis, and
accordingly, with respect to mass reduction technologies, does not recommend one family of
materials over any other.
AISI also requested a response to its comment suggestion that EPA provide off-cycle credits
based on the GHG impact of the use of lightweight materials. This comment relates more
86
-------
specifically to off-cycle credits. Comments relating to off-cycle credits are addressed in Chapter
3.9 of this RTC document.
An anonymous citizen commented that the drive to reduce mass to improve fuel economy has
caused manufacturers to eliminate the full size spare tire (as an alternative to applying mass
reduction technology) and suggested that the fuel economy benefit would not be realized if the
owner chooses to add a full size spare. EPA recognizes that many manufacturers have replaced
the full size spare with a smaller temporary spare, and in some cases have even replaced the
temporary spare with a tire repair kit, which includes tire sealant and an inflator. While this does
reduce vehicle mass, it is not clear that fuel economy regulations are the only driver of this trend.
Removing the spare reduces manufacturing cost, and also frees up interior space, which can help
with component packaging and improve cargo capacity. Some vehicles without a spare are
equipped with so-called run-flat tires, which are designed to maintain enough pressure after
being punctured that the car may be driven to a service location for tire repair instead of relying
on a spare. While removing the spare may therefore provide a limited opportunity to improve
fuel economy, these other factors are also likely to be contributing to this trend.
2.5.13 Aerodynamics
Summary of Comments on the Draft TAR addressed in the Proposed Determination
For the FRM and Draft TAR technical assessments, EPA assumed constant cost and
feasibility assumptions for incremental improvements in aerodynamic performance for every
vehicle within a class. Several stakeholders submitted comments that EPA's Draft TAR
assessment did not account for the drag reduction that some vehicles have already adopted, and
recommended that the aero levels in the baseline fleet reflect appropriate drag reduction achieved
by each vehicle. Because these comments were made in reference to technologies represented in
the baseline fleet, they are also reviewed in Chapter 2.7.1 of this RTC document.
EPA agreed with the commenters that it is appropriate to account for aerodynamic drag
reductions already present in the baseline fleet in order to avoid overestimating the amount of
additional improvement that can be achieved at a given cost. In response to these comments, for
the Proposed Determination, EPA largely accepted commenters' suggestions to account for
aerodynamic improvements already present in the baseline fleet, using MY2015 certification
data for purposes of that accounting. EPA used an analysis of coastdown coefficients to estimate
the levels of aerodynamic drag reduction already present in MY2015 vehicles. The vehicles were
then binned into one of three aerodynamic technology levels according to the potential for future
improvement. The three levels correspond to the two aerodynamic technology levels modeled
throughout the EPA analysis (Aerol and Aero2), and a zero-technology level representing no
technology added. The assignment of each baseline vehicle to one of three aerodynamic
technology levels (rather than specific drag coefficient or frontal area values) is therefore
consistent with the structure of the modeling. The Aerol and Aero2 levels and the accounting
for drag reducing technology in the baseline fleet were described in Chapter 2.3.4.4 of the TSD
and is also reviewed in Section A.2.1 of the Proposed Determination Appendix at p. A-2 and A-
3.
Summary and Response to Comments on the Proposed Determination
87
-------
Mercedes-Benz expressed appreciation for the changes made to the 2015 baseline fleet
between the Draft TAR and the Proposed Determination to more accurately reflect the baseline
level of aerodynamic technology already present the fleet. EPA notes that these changes were
made as a result of the public comments on the Draft TAR, and reflected the suggestions in those
comments (as well as reflecting EPA's announced intention to use final MY2015 certification
data in the analysis as it became available). With respect to accounting for aerodynamic
improvements, for the Proposed Determination, EPA agreed with Draft TAR commenters that
the approach in the Draft TAR, whereby every vehicle in the baseline fleet had the potential for
20 percent improvement in aerodynamic drag, failed to account for improvements already made
in certain baseline vehicles, and thus overstated potential feasible improvements. In estimating
the amount of aerodynamic drag already present in the 2015 baseline fleet, EPA largely accepted
commenters' suggestions to account for aerodynamic improvement already achieved for each
vehicle in the 2015 fleet, using 2015 certification data for purposes of that calculation. See TSD
Table 2.139 showing aerodynamic drag area statistics for each of the 17 size classes of vehicle.
In comments that were raised in the context of OMEGA outputs, Toyota stated that in the
Proposed Determination, EPA did not account for the additional weight of the Aero2 package for
the Tundra/Tacoma platform. This comment referred back to a Toyota comment on the Draft
TAR that Toyota felt EPA had only partially addressed, which was, "OMEGA shows that many
vehicles, such as the Tundra and Tacoma, will have both AER02 and a net weight reduction.
Although it is possible to use AER02 on these trucks, the expectation is that the weight of the
vehicle would increase, not decrease as shown by OMEGA. This is because the additional
components associated with AER02 performance improvements will increase the weight of the
vehicle."
For technologies that can ordinarily be expected to add significant weight to the vehicle, EPA
does account for the additional weight and its impact on the ability to achieve a net weight
reduction by application of mass reduction technology. For example, as described in Section
2.3.4.3.7.1 of the TSD, the weight of batteries and other components of electrified vehicles is
explicitly modeled and accounted for because it is clear that their weight is significant and may
or may not allow a net vehicle weight reduction to be achieved within the limit of allowable
application of mass reduction technology. In fact, for these vehicles, OMEGA tracks both a
percent net weight reduction achieved (WRnet) and a percent of mass reduction technology
applied (WRtech). In many cases, WRtech is greater than WRnet, indicating that the weight of the
added battery and electric drivetrain (after conventional powertrain components are removed)
resulted in less net weight reduction than the applied level of mass reduction technology
otherwise would have achieved. In some extreme cases, there is a net weight increase, indicating
that the added mass was so great that it precluded achieving any net weight reduction within the
limits on allowable application of mass reduction technology. In the case of other technologies
(such as Aero2), the added weight is relatively much smaller and would not be expected to
preclude the possibility of achieving any net weight reduction on the vehicle. Toyota did not
specify in its comments on either the Draft TAR or Proposed Determination the magnitude of the
weight increase that it expects Aero2 to add to a Tundra-type vehicle, nor provided details as to
why it expects the weight increase to be so significant that it would preclude the possibility of
achieving any net weight reduction on a vehicle of this type (as OMEGA showed). In any case,
the relatively small mass addition that is likely to be associated with Aero2 is unlikely to
88
-------
significantly impact the ability to combine Aero2 with mass reduction technology and thereby
achieve a net weight reduction of the vehicle similar to what the OMEGA output indicated.
An anonymous citizen expressed concern over reductions in vehicle interior volume. While
this comment may refer to general downsizing of vehicles for fuel economy, it may also refer to
the potential for a manufacturer to reduce aerodynamic losses by reducing the frontal cross
sectional area of the vehicle rather than or in addition to improving its coefficient of drag. EPA
has not observed a discernible trend toward reduction of frontal area, as most manufacturers wish
to maintain the utility of interior volume and have focused on reducing drag coefficient by means
of modifications to the exterior surface and shape of the vehicle. EPA expects that manufacturers
will continue to improve the efficiency of their designs to be able to maintain interior volume
while reducing aerodynamic drag. This can be accomplished in part by making more efficient
use of available space within the envelope of the vehicle. For example, in the discussion of
aerodynamics in the Proposed Determination we cited the redesign of the 2015 Acura TLX sedan
as having reduced frontal area by 1.5 percent without sacrificing interior space (see TSD Chapter
2.2.5.2 at p. 2-137) due to the design approach described in the cited presentation materials from
Acura. It should also be noted that the footprint-based standards that are the subject of this Final
Determination are designed to improve the fuel economy of vehicle of all sizes, without
compelling manufacturers to manufacture smaller vehicles in order to comply with the standards.
2.5.14 Tire Rolling Resistance
Summary of Comments on the Draft TAR addressed in the Proposed Determination
For the FRM and Draft TAR technical assessments, EPA assumed constant cost and
feasibility assumptions for incremental improvements in tire rolling resistance. Commenters on
the Draft TAR noted that "low rolling resistance tires are increasingly specified by OEMs in new
vehicles," yet were not accounted for in EPA's baseline fleet. Because these comments were
made in reference to technologies represented in the baseline fleet, they are also reviewed in
Chapter 2.7.1 of this RTC document.
EPA agreed with the Draft TAR commenters that it is appropriate to account for tire rolling
resistance reductions already present in the baseline fleet in order to avoid overestimating the
amount of additional improvement that can be achieved at a given cost across the fleet. In
response to these comments, for the Proposed Determination, EPA largely accepted commenters'
suggestions to account for rolling resistance improvement already present in the baseline fleet,
using MY2015 certification data for purposes of that accounting. EPA used an analysis of
coastdown coefficients to estimate the levels of tire rolling resistance reduction already present
in MY2015 vehicles, and assigned one of three tire rolling resistance levels to each vehicle in the
baseline fleet. The three levels correspond to the two tire rolling resistance technology levels
modeled throughout the EPA analysis (LRRT1 and LRRT2), and a zero-technology level
representing no technology added. The assignment of each baseline vehicle to one of three
rolling resistance levels (rather than a specific rolling resistance value) is therefore consistent
with the structure of the modeling analysis, and while it is not intended to exactly represent the
specific level of technology present in any specific individual vehicle, EPA believes that it is
effective at representing existing technology and remaining potential in aggregate across the fleet
as required for the purposes of the analysis. The LRRT1 and LRRT2 levels and the accounting
for rolling resistance technology in the baseline fleet were described in Chapter 2.3.4.5 of the
89
-------
TSD at p. 2-410 and is also reviewed in Section A.2.1 of the Proposed Determination Appendix
at p. A-2 and A-3.
Summary Comments and Responses on the Proposed Determination
Some OEMs, through confidential comments provided to EPA, questioned the feasibility of
reducing tire rolling resistance. These comments included information describing the rolling
resistance characteristics said to be present on specific vehicles or across their respective fleets to
support their position that many vehicles in production already apply significant amounts of
rolling resistance technology and that this places the potential for further reductions in doubt. In
response, EPA has acknowledged that existing vehicles include varying levels of rolling
resistance reduction and has attempted to account for this by assigning levels of existing rolling
resistance technology to vehicles in the baseline fleet, as described above and in Chapter 2.3.4.5
of the TSD at p. 2-410. As stated previously in this discussion, EPA believes that on average and
for the purposes of the overall analysis, this method is effective at representing the amount of
technology already applied in aggregate across the fleet, and the remaining potential yet to be
applied.
2.5.15 Low Drag Brakes
Summary of Comments on the Draft TAR addressed in the Proposed Determination
The state of low-drag brake technology was discussed in TSD Chapter 2.2.8.4. TSD Chapter
2.3.4.7.4 discussed cost and effectiveness assumptions and addressed public comments that were
received on this technology.
In comments on the Draft TAR, Toyota commented on several aspects of EPA's low-drag
brake assessment. With respect to the Draft TAR analysis, Toyota commented on the
conclusions regarding the Direct Manufacturing Costs (DMC) and stated that in order to
"calculate such a detailed cost, it must be fixed with a special brake system of that of a specific
supplier." In response, EPA noted that the DMC for this technology is not derived from a
specific supplier's design, but rather is an aggregate cost representing all of the changes that can
be made to the brake system to reduce drag, including caliper seal and return rate and rotor and
lining changes. This is a more reasonable basis for a cost estimate than a specific supplier quote
because it represents individual component costs that should be present in any specific supplier
or OEM design. Toyota also commented on EPA's summary of available zero drag brake
systems. In response to these comments, EPA updated the description of the zero drag brake
technology in TSD Chapter 2.2.8.4, while noting that zero-drag brakes were not included in the
Draft TAR or Proposed Determination analyses. It should be noted that this makes the overall
technology assessments more conservative, because this represents one more technology option
that was not considered in the analysis but which manufacturers have at their option to comply
with the standards.
Commenters on the Proposed Determination did not present additional comments on low drag
brake technology.
90
-------
2.5.16 Air Conditioning
2.5.16.1 A/C Efficiency Credits
This chapter reviews comments relating to A/C efficiency credit mechanisms. The A/C
efficiency credit program includes several provisions for earning credits, including a pre-defined
credit menu (i.e. the list of credits for air conditioning system efficiency in 40 CFR 86.1868-12
for which the regulation provides a default value), which manufacturers can access by
completing testing requirements; an option for engineering analysis to replace part of the testing
requirement; and an alternative pathway to credits through the off-cycle credit program.
Because of its overlap with the off-cycle program, comments relating to A/C efficiency are in
some cases better addressed in the context of the off-cycle program. Comments on the off-cycle
program are reviewed in Chapter 3.9 of this RTC document.
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Chapter 2.2.9.1 of the TSD discussed the air conditioning (A/C) efficiency credit program and
addressed comments on the Draft TAR that relate to this topic.
Some comments received on the Draft TAR concerned the process for applying for A/C
credits under the off-cycle pathway. These comments primarily requested that EPA simplify and
standardize the off-cycle application process, in the specific context of application for A/C
efficiency credits. Commenters cited the need to provide greater certainty to manufacturers that
credits would be approved before making investments, and to reduce application burden. EPA's
response to requests for simplification and standardization of the off-cycle application process
are found in Section B.3.4.1 of the Proposed Determination Appendix at p. A-103 to A-106.
Other comments noted the importance of continuing to incentivize further innovation in A/C
efficiency technologies as new technologies emerge that are not in the credit menu, or when
manufacturers begin to reach the regulatory caps on menu credits. These commenters suggested
that EPA should consider adding new A/C efficiency technologies to the credit menu and/or
update the credit values, particularly those that would otherwise qualify for credits through an
off-cycle application as a non-menu A/C efficiency technology, or through a demonstration that
credit beyond the menu default value is warranted. EPA acknowledged that the credit menu has
been well received as a way to incentivize A/C improvements, but stated that it continues to
believe that expanding the credit menu would be inconsistent with the goal of transitioning the
A/C program toward a performance basis, and declined to alter the rule by expanding the list of
default A/C efficiency technologies, or by changing the default values. Non-menu
improvements would continue to be eligible for additional credit pending demonstration of
performance and approval by the agency. The full discussion is found in TSD Chapter 2.2.9.1.1
at p. 2-189.
Several commenters questioned the applicability of a cap on non-menu A/C efficiency
technologies claimed through the off-cycle process. EPA clarified that although the specific
regulatory caps specified under 40 CFR 86.1868-12(b)(2) apply to menu-based A/C credits and
are not part of the off-cycle regulation (which is defined at 40 CFR 86.1869-12), EPA has
discretion through its authority in the off-cycle approval process to take into account factors
deemed relevant, including consideration of synergies or interactions among applied
technologies, which could potentially be addressed by application of some form of cap or other
91
-------
applicable limit on total A/C credits, if warranted. More discussion on this topic is found in TSD
Chapter 2.2.9.1.1 at p. 2-190.
Some commenters expressed uncertainty about the capabilities of the AC 17 Test. In response
to suggestions that the test could not capture all possible usage conditions, EPA acknowledged
that no single test procedure is likely to be capable of such performance, which was well
understood at the time of its development. EPA also noted that the test represents an industry
best effort at identifying a test that would greatly improve upon the range of usage conditions
represented by the Idle Test, and that industry evaluation of the test shows that it achieves this
objective (as well as the ability to resolve small differences in CO2 effectiveness when carefully
conducted). More discussion is found in TSD Chapter 2.2.9.1.3 at p. 2-196.
Other commenters expressed uncertainty about the use of the AC 17 Test as part of the process
for qualifying for and quantifying A/C efficiency menu credits beginning in MY2020. In
response to a suggestion that the need to identify a baseline vehicle for A-to-B comparisons
makes the test unworkable, EPA disagreed, pointing out that the regulation provides for
engineering judgment when identifying a suitable baseline vehicle, and also that the engineering
analysis provision under 40 CFR 86.1868-12(g)(2) provides an alternative to locating and
performing an AC 17 test on a baseline vehicle if such a vehicle cannot be identified. In response
to comments that the process by which manufacturers may pursue the engineering analysis
option should be streamlined and made clearer, EPA pointed out that we are continuing to
coordinate with the SAE Cooperative Research Program (CRP) on standards for bench testing of
hardware-based technologies, which would contribute to this goal. EPA also continues to believe
that dialogue between EPA and industry stakeholders in the A/C credit program has been in the
past, and will continue to be, an effective means toward identifying and developing practical
solutions to this issue as well as other issues similar to those raised by some of the commenters.
More discussion is found in TSD Chapter 2.2.9.1.3 at p. 2-196 to 2-197.
Commenters also suggested that other aspects of the credit application process should be
streamlined. These comments included suggestions such as: (a) that EPA should consider joint
applications by OEMs for the same A/C efficiency technology; and (b) that EPA should consider
allowing suppliers to directly petition for credits and allow the approved credits to be applicable
to OEMs that later adopt the technology. EPA responded by pointing out that concerns with
system integration would likely make it very challenging if not impossible for a supplier, for
example, to be able to demonstrate (through a hypothetical supplier-sponsored credit application)
that a given A/C technology, as represented perhaps by a stock part number, would necessarily
always result in the same or similar level of GHG effectiveness regardless of the vehicle on
which it is installed. EPA expressed similar concerns with regard to the possibility of joint OEM
applications, and stated that it remains unclear that joint applications would be practical or
desirable as a means to streamline the process. More discussion can be found in TSD Chapter
2.2.9.1.3 at p. 2-198.
Summary of Comments and Responses on the Proposed Determination
EPA received comments relating to A/C efficiency from Motor and Equipment Manufacturers
Association (MEMA), Mercedes-Benz, and Toyota. Additional comments relating more closely
to the off-cycle credit program or to program harmonization were also received from these and
other commenters and are addressed in Chapters 3.9 and 3.10, respectively.
92
-------
MEMA recommended that EPA reconsider its stated intent in the Proposed Determination to
cap A/C credits earned through the off-cycle petition process, and stated appreciation that EPA
expects to consider the applicability of a cap on a case by case basis. MEMA also encouraged
EPA to provide deference in this matter as the industry deploys such technologies. EPA
addressed the application of a cap on A/C credits earned through the off-cycle process in TSD
Chapter 2.2.9.1.1 at p. 2-190. EPA also stated its interest in seeing that the A/C credit program is
able to operate as it was designed (TSD p. 2-197), and recognized that continued collaboration
and dialogue between EPA and the industry has been an effective mechanism toward pursuing
this outcome (TSD p. 2-197 and 2-199). Accordingly, EPA acknowledges that many
considerations are relevant in judging and promoting the successful operation of the A/C credit
program and expects to take such considerations into account where warranted.
Toyota praised EPA's collaboration with industry on issues related to A/C efficiency credits
and mentioned looking forward to EPA's continued support in this matter. Toyota also stated
that neither the requirement for AC 17 testing nor the off-cycle provisions for non-menu A/C
credits have been effective at transitioning the program to a performance basis. Toyota
recommended that EPA therefore provide relief by expanding the A/C credit menu until this
issue is resolved, and that relying on the off-cycle path to expand A/C credit opportunities is
unrealistic. In response, EPA also looks forward to continued collaboration with industry to
ensure that the A/C credit program continues to function as designed. As stated in the Proposed
Determination, EPA continues to believe that expanding the A/C credit menu with new pre-
approved A/C technologies is not necessary at this time. EPA also believes that the identification
of non-menu technologies that might be proposed for addition to the menu, as well as the
development of appropriate credit values for each, is best achieved by consideration of off-cycle
applications as these technologies are identified by stakeholders wishing to receive credit. EPA
also notes that the June 2016 petition by the Alliance and Global Automakers regarding certain
harmonization issues of the CAFE and GHG programs includes consideration of certain A/C
efficiency credit issues and improvement of the off-cycle credit approval process, as further
discussed in Chapter 3.9 of this RTC document.
2.5.16.2 A/C Refrigerant Credits
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Chapter 2.2.9.2 of the TSD discussed air conditioning leakage reduction and alternative
refrigerants for mobile air conditioning systems and addressed comments on the Draft TAR that
relate to this topic.
Regarding air conditioning refrigerants, EPA did not receive comments on our conclusions in
the Draft TAR that auto manufacturers are continuing to improve the leak-tightness of their A/C
systems and that many are transitioning to the use of low-GWP alternative refrigerants in a
number of vehicle models. These conclusions reinforce our earlier projections that a complete
transition to alternative refrigerants by MY2021 will in fact become reality (See TSD Chapter
2.9.2).
EPA received comments on the whether the supply of alternative refrigerants would be
sufficient for the projected transition. Some commenters indicated that supply is still a concern,
while others, including two producers of HFO-1234yf, commented that there will be sufficient
supply. Moreover, some automotive manufacturers are developing systems that can safely use
93
-------
other substitutes, including CO2, for which there is not a supply concern for the refrigerant.
Based on all of the information before the agency, EPA concluded in the TSD (Chapter 2.9.2.2),
that production plans for alternative refrigerants are in place to make available sufficient supply
no later than MY2021 to meet current and projected demand domestically as well as abroad.
Summary of Comments on the Proposed Determination
Honeywell supported EPA's conclusion in the Proposed Determination that the A/C crediting
system is appropriate and leading to significant GHG reductions through lower-leak A/C systems
and industry adoption of lower-GWP refrigerants. Honeywell states that lower warranty costs
can result, that the transition to HFO-1234yf systems has not been very difficult or costly for
manufacturers (and these systems in fact can cost less than existing systems), and that this new
refrigerant is easy to service in the field. Honeywell credits industry adoption of 1234yf and the
EPA A/C credit system for creating sufficient demand to enable Honeywell to build a major
production plant in the U.S., creating over 400 jobs and generating ancillary economic
development.
Response to Comments on the Proposed Determination
EPA acknowledges Honeywell's comments about lower-leak A/C systems and lower-GWP
refrigerant, including HFO-1234yf.
2.6 Criteria Emissions / Tier 3
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Chapter 2.3.3.3.8 of the TSD discussed the Tier 3 emission program and how it was
developed in full consideration of the GHG programs. The TSD discussion contains the agency
response to comments received on this topic in the Draft TAR.
Comments on this part of the Draft TAR asserted that the ALPHA model failed to account for
CO2 and degradation of fuel economy (FE) associated with Tier 3 emission control systems, and
the impact of more stringent Tier 3 evaporative emission regulations in the MTE analysis.
EPA's response to these comments (starting at page 2-269) highlighted that the two programs
were purposely coordinated to allow the development of the emission control hardware for both
criteria emissions and GHGs to be complementary. Several technologies used to reduce GHG
emissions can also reduce criteria emissions and vice-versa. For example, downsized engines
being used to comply with GHG emission requirements also generally have lower engine out
criteria pollutant emissions simply from the reduction in cylinders and displacement, resulting in
less opportunity for release of cold start emissions which consist of hydrocarbons and other
emissions such as PM. As a result, EPA determined that it was not appropriate to assume that
criteria emission control strategies implemented for Tier 3 automatically result in CO2 increases,
and therefore it was not necessary to change the modeling assumptions.
Additionally, EPA's response acknowledges that evaporative emission challenges exist today;
however, Tier 3 does not increase the purge requirements demanded from the existing Tier 2
program. Certainly some new technologies may reduce the opportunity to purge the canister;
however, these challenges have been addressed today in many vehicle applications. Technology
approaches as described in the TSD Chapter 2.3.3.3.8 at p. 2-269 have been effective at
addressing the purge challenge.
94
-------
We concluded in the response that the best available GDI technology has already achieved
criteria pollutant standards associated with the future Tier 3 requirements. Accordingly, EPA did
not accept the implication in the comment that there is a tradeoff between control of criteria
pollutant emissions and CO2 control.
Summary of Comments and Responses on the Proposed Determination
MECA commented that EPA should align Tier 3 PM limits with CARB LEVIII standards.
This request is out of the scope of this determination; however, EPA's position and underlying
data supporting the Tier 3 limits can be found in the Tier 3 final rule Technical Support
Document.
MECA commented that the increase in PHEV sales MY2022+ may lead to unintended
increases in VOC emissions due to sealed tanks and also issues with Tier 3 evaporative emission
provisions and that EPA should commit to expedited follow-up rulemaking. EPA is assessing
any potential issues with changes to evaporative emission control systems due to current and
future PHEVs and will determine if test procedures appropriately protect for any unintended
increases in VOC emissions. EPA and CARB have historically worked closely with
manufacturers to modify test procedures in cases where historical practices no longer match the
real world operating conditions of new technologies including PHEVs. If the agency determines
that test procedures are no longer protective of unintended VOC emissions during refueling,
operation or diurnals, EPA will investigate the extent of the issue and may propose potential
solutions.
The Alliance commented that while EPA discussed options for reducing criteria and PM
emissions that do not degrade GHG performance in the TSD, EPA did not address costs or
implementation issues of these options. In the TSD, EPA highlighted that the Tier 3 emission
standards were developed in full consideration of the GHG phase 2 standards, particularly
because both programs have almost identical implementation schedules and dates (MY2017 to
MY2025). The Tier 3 costs were generally determined with the expectation that phase 2 GHG
technologies would be introduced in vehicles during the same time frame as the implementation
of the Tier 3 emission standards. The Tier 3 cost analysis incorporated any additional cost
increases (or reductions in some cases) for criteria emission controls related to new GHG
technologies such as downsized turbocharged and Atkinson engines. The example in the TSD
for PM control involving dual injection strategies was provided only as an example of a potential
solution used in the past by a limited number of vehicles however most modern GDI injection
systems and controls do not require this approach and therefore it was not a strategy included in
the Tier 3 assessment or costs.
2.7 Baseline Fleet
This section reviews comments and responses on the data sources and approach used for
creation of the baseline fleet.
Comments relating specifically to the representation of advanced technologies in the baseline
fleet are addressed separately in Section 2.7.1 below. Comments related to EPA's inclusion of
ZEV-program vehicles in the baseline are reviewed in Section 2.7.2 below.
Summary of Comments on the Draft TAR addressed in the Proposed Determination
95
-------
TSD Chapter 1.1 discussed how EPA created the baseline fleet. Public comments on the
Draft TAR that are related to the baseline fleet were discussed in Section A.2.1 of the Proposed
Determination Appendix at p. A-2, and in TSD Chapter 1.1.2 beginning at p. 1-2.
Some commenters on the Draft TAR highlighted the importance of using the latest MY fleet
data to create the baseline fleet. For the Draft TAR, EPA had used MY2014 certification data,
which at the time was the most recent final complete data. See Draft TAR at p. 4-9. For the
Proposed Determination analysis, consistent with EPA's stated preference to use the most recent
model year data for which there is a complete set of certified data, EPA updated the baseline
fleet to a MY2015 basis using the MY2015 GHG compliance data set, which had been
completed since the Draft TAR analysis. More discussion of these comments may be found in
TSD Chapter 1.1.2 at p. 1-2. EPA also made adjustments to better represent the degree to which
low rolling resistance tires, aerodynamic technologies, and mass reduction had been
implemented in the 2015 baseline fleet, again consistent with commenters' suggestions. See
Proposed Determination Appendix A, Sections A.2.1.2 and A.2.1.3.
Other comments addressing the data used for the baseline included: concerns about the
presence of vehicles that are no longer in production, a recommendation that the baseline be
created using a multi-year average, and various perceived inconsistencies with contributing data
sources such as IHS-Polk and EIA's AEO2015 data. With respect to vehicles that are no longer
in production, EPA clarified that its projection of the future fleet is performed in a way that
respects the discontinuation and succession of models within a vehicle class, a normal aspect of a
manufacturer's product development. Further discussion of these comments may be found in
TSD Chapter 1.1.2 at p. 1-2 to 1-3, and 1-16.
Summary and Response to Comments on the Proposed Determination
The Alliance commented, "In addition to using projections instead of actuals, EPA
compounds the problem by cherry-picking projections from the Trends Report. For example,
EPA relies on the projection of car market share increasing from 57.4% in MY 2015 to 62.1% in
MY 2016. The 2016 Baseline Study predicts a car market share decrease to 55.7% in MY
2016." In response, EPA does not agree that it is cherry-picking projections from the Trends
report. EPA has not applied any of the data from any Trends report in the creation of the
MY2015 baseline used in the Proposed Determination. The process and data sources for creating
the MY2015 baseline are explained in Chapter 1.1.3 of the TSD for the Proposed Determination.
EPA used the Energy Information Administration (EIA) Annual Energy Outlook (AEO) 2016 for
determination of the car vs. truck market share. Specifically, EPA used a customized version of
EIA's 2016 annual energy outlook for car vs. truck market share for all years except for the
baseline year which uses actual volumes. See TSD Chapter 1.1.3.1.1. EPA has consistently
used versions of EIA's annual energy outlook in past analyses of the baseline and future fleet
forecasts for vehicle GHG standards assessments, and commenters have not questioned the
appropriateness of its use.
Global Automakers commented, "First, although it was appropriate for EPA to update their
baseline fleet to reflect the MY 2015 vehicles, by rushing forward with the determination, the
agency has foreclosed its ability to use the most up-to-date information in the Midterm
Evaluation. Much of the MY 2016 fleet is now complete, and this information could be used to
better inform EPA's data as well." In response, the MY2016 sales end at the end of calendar
year 2016, and most manufacturers do not report their final GHG data to EPA until three months
96
-------
after the end of the year as stated in Chapter 1.1.3 of the Proposed Determination TSD. MY2015
is the most recent year for which final certification data is available, and EPA has used that
information in the Proposed and Final Determinations (as recommended by commenters to the
Draft TAR).
Additional comments related to the creation of the baseline fleet were concerned specifically
with how existing technologies in the base fleet were represented. These comments are reviewed
in the following section.
2.7.1 Technologies in Baseline
This chapter reviews comments and responses on the topic of technologies represented in the
baseline fleet.
Summary of Comments on the Draft TAR addressed in the Proposed Determination
TSD Chapter 1.1 discussed how EPA created the baseline fleet. Public comments on the
Draft TAR that are related to the baseline fleet are discussed in Section A.2.1 of the Proposed
Determination Appendix at p. A-2, and in TSD Chapter 1.1.2 beginning at p. 1-2.
Some comments relating to the baseline fleet focused on how EPA accounted for technologies
already implemented in the current fleet, such as low rolling resistance tires, aerodynamic drag
reduction, and mass reduction. That is, that the Draft TAR analysis may have overestimated the
degree to which reductions in rolling resistance and aerodynamic drag could be implemented at
the estimated costs, on the grounds that some vehicles in the existing fleet have already
implemented some of these technologies to varying degrees. For example, one commenter noted
that EPA had acknowledged that "low rolling resistance tires are increasingly specified by OEMs
in new vehicles," yet had not apparently accounted for this existing penetration of this
technology in the baseline fleet. Similarly, some OEM commenters pointed out that aerodynamic
improvements have been implemented in new vehicle designs over the past several years, and
felt that these improvements were not adequately reflected in the Draft TAR aerodynamic
technology baseline. The logic of these comments is that if the technologies are not reflected in
the baseline, then the projected efficiency improvements would either not be feasible, or would
be feasible only at greater cost, because more advanced technologies would be needed since the
lower cost technologies would already be in baseline vehicles.
EPA agreed with many of these comments and made the types of adjustments recommended
by the commenters as to how these technologies are represented in the baseline fleet in the
Proposed Determination analysis.
Specifically, EPA updated its assessment of tire rolling resistance and aerodynamic drag
reduction technologies by accounting for their estimated presence in the baseline fleet and
modifying the permissible application of these technologies accordingly, as the commenters
suggested. The accounting for rolling resistance reduction was described in Chapter 2.3.4.5 of
the TSD at p. 2-410 and was also reviewed briefly in Section A.2.1 of the Proposed
Determination Appendix at p. A-2 and A-3. This accounting is also reviewed in Chapter 2.5.14
of this Response to Comments (RTC) document. The accounting for aerodynamic drag reduction
was described in Chapter 2.3.4.4 of the TSD and reviewed briefly in Section A.2.1 of the
Proposed Determination Appendix at p. A-2 and A-3. This accounting is also reviewed in
Chapter 2.5.13 of this RTC document.
97
-------
Because EPA also updated its baseline fleet basis to MY2015, accounting of mass reduction
present in each vehicle in the MY2015 baseline fleet was also updated. EPA performs this
accounting by comparing the MY2008 version of each model to the baseline MY version (here,
MY2015) according to the sales weighted average curb weights of the various trim levels, after
adjusting for size, additional safety requirements, and drive type, as reviewed in Chapter
2.3.4.6.1 oftheTSDatp. 2-411 to 2-412.
Summary and Response to Comments on the Proposed Determination
In comments on the Proposed Determination, Global Automakers stated, "Several errors were
identified in the Proposed Determination. One example occurs in EPA's classification of Civic
models which, due to the agency's methodology, resulted in a gross overestimation of hybrids
(HEVs) in Honda's modeled reference fleet. EPA acknowledged the existence of this error in the
Proposed Determination's TSD." In response, EPA notes that this refers to an error in the Draft
TAR's MY2014 baseline that does not exist in the MY2015 baseline used for the Proposed
Determination. The error resulted from a few Honda vehicles having been mislabeled. The error
affected only Honda vehicles, and only in the Draft TAR analysis. The commenter's reference
to "several errors" seems to point to errors described in the comments submitted by Honda on the
Draft TAR, which as noted EPA has since corrected. The commenters do not specifically point
to new errors introduced as part of the Proposed Determination. EPA made significant effort to
prevent errors in the projection process for the MY2015 baseline used for the Proposed
Determination, and has not identified any such errors in that projection data.
In commenting on the OMEGA analysis, some commenters did refer to errors in specific data.
Global Automakers stated, "Some of the Proposed Determination's OMEGA data files contain
what appear to be mistakes. Take EPA's modeling of electronic power steering (EPS), for
example. In the Proposed Determination, EPA's baseline fleet does not include EPS for any
Honda or Acura models. This is wrong and contradicts the agency's own MY 2014 baseline
analysis, which applied approximately 80% penetration to the fleet. Errors in the latest OMEGA
files are not limited to EPS. Other incorrectly applied technologies included aero designation (on
Civic HF) and VVT application (on multiple Honda and Acura models)."
With regard to EPS, EPA obtained EPS information from the MY2015 VOLPE baseline file
that was published with the Draft TAR in July of 2016. Because no commenters on the Draft
TAR questioned the EPS information in that file, the information was subsequently carried over
into the Proposed Determination analysis. EPA agrees that some of the vehicles in the baseline
were incorrectly coded with respect to the presence of EPS. However, EPA believes that the
effect of this error on projected costs is minimal and would certainly not change the conclusions
of the analysis. There exists considerable technology available to compensate for the lost
effectiveness improvement potential of EPS had EPS been included in the baseline. In fact, our
central case run (ICM, AEO reference case fuel prices) shows that EPS penetrates the fleet at a
rate above 90 percent in the reference case, with slightly more penetration in our control case.36
36 EPS was not included in our detailed technology penetration rate tables presented in Section C. 1.1.3 of the
Proposed Determination. The technology penetration rates of all technologies, including EPS, can be found in the
detailed "OMEGA-core Runs" files included in the docket and on the OMEGA webpage at
https://www.epa.gov/sites/production/files/2016-l l/omega-pd2016-omegacore-runs.zip. Specifically, the file
named "Tables_TechPens_20161118_icm_aeoR.xlsx" (and similarly named files for each sensitivity) provides
the detailed technology penetrations for all technologies.
98
-------
Therefore, less than 10 percent of the reference case fleet is impacted by this EPS error. Based
on the results of our other sensitivity analyses, EPA estimates that slightly more Atkinson-2,
stop-start, and transmission improvements would occur with only minor cost impacts. Therefore,
while we regret the error, we do not believe it has any effect on the conclusions of our
assessment.
Further, with respect to aerodynamic technology on the Honda Civic HF, EPA believes that
the representation of aerodynamic technology is appropriate based on EPA's estimation of
aerodynamic drag performance (as represented by estimated drag area or CdA37) for the vehicle.
As described in Chapter 2.5.13 of this RTC document and also in the TSD as cited there, for
each vehicle in the baseline, EPA accounts for the potential for additional aerodynamic
improvement by placing vehicles in three bins (no aero technology, Aerol, and Aero2) by a
methodology that analyzes the road load coefficients that are supplied by manufacturers for
vehicle certification. In the case of the Honda Civic HF, the particular road load coefficients
reported for this vehicle places it slightly short of the cutoff value for the Aerol technology bin.
The manufacturer-supplied road load coefficients for other Civic models indicated somewhat
lower C-coefficients than the Civic HF, and as a result the methodology placed those variants in
the Aerol bin. As stated previously in Chapter 2.5.13 of this RTC document where the
methodology is reviewed, while this method is not intended to exactly identify the specific drag
reducing technologies present in any individual vehicle, EPA believes that it is effective at
representing existing technology and remaining potential in aggregate across the fleet as required
for the purposes of the analysis.
Regarding representation of VVT technology, EPA believes that VVT on Honda and Acura
models is represented correctly per Honda's certification information and information from
Ward's Automotive. Honda certified that all of its gasoline engines had VVT. EPA uses Ward's
Automotive engine data as a quality control measure, and this data corroborates that all Honda
engines are equipped with VVT.
Most of the data used for creation of the MY2015 baseline fleet was taken directly from
manufacturer certification data, with the exception of data on valve actuation, EPS, and curb
weight. Valve actuation type (OHV, SOHC, DOHC) came from Ward's Automotive and was
verified by using the valves per cylinder information given in the manufacturer certification
information versus the valves per cylinder information given by Ward's Automotive. EPS, as
stated earlier, was not verified. Curb weight was verified by reference to publicly available
information and certification data to ensure it was accurate, and in some cases was corrected.
Given that the vast majority of the data in the MY2015 baseline comes directly from what
manufacturers have certified their vehicles as containing, EPA believes that the data is largely
accurate, aside from the EPS data that EPA acknowledges was incorrectly coded.
2.7.2 The ZEV Program in the OMEGA Analysis Fleet
Summary of Comments on the Draft TAR addressed in the Proposed Determination
In the Proposed Determination (and the Draft TAR), EPA recognized that state Zero Emission
Vehicle (ZEV) regulations would result in a significant number of electrified vehicles to be
present in future vehicle fleets. EPA reasoned that because these ZEVs are already required by
37 CdA is the product of a vehicle's coefficient of drag (Cd) and frontal area (A).
99
-------
separate laws in California and nine other states, these vehicles will be part of the reference fleet
by virtue of those requirements. The federal standards thus would not be imposing additional
requirements or costs to these vehicles, nor would the federal standards result in benefits which
would not otherwise occur. To avoid double counting, EPA thus considered these ZEV vehicles
to be part of the reference fleet, and projected the number of electrified vehicles thus included.
The Alliance of Automobile Manufacturers and others commented that including electrified
vehicles that result from compliance with the ZEV program as part of our reference fleet analysis
was unfairly counting their benefits without estimating their costs. EPA noted that this comment
is mistaken because the ZEV program vehicles are included in both the reference and control
cases; therefore, EPA is neither calculating the benefits nor costs of the ZEV program as part of
the MY2022-2025 standards assessment. EPA also noted that the methodology is consistent
with OMB Circular A-4, which states that in developing a baseline for purposes of analyzing the
potential effects of a proposed rule,"[t]his baseline should be the best assessment of the way the
world would look absent the proposed action."38 Other commenters, including the
Environmental Defense Fund and the Union of Concerned Scientists, supported the inclusion of
ZEV program vehicles in the reference fleet, pointing out that this approach ensures that the
costs of the ZEV program, which are not imposed by the 2022-2025 standards but rather by state
law, are not included as costs of the national rule. Further, in the Proposed Determination TSD at
page 1-33, EPA noted that any ZEV vehicles sold in California and other states would help a
manufacturer in meeting the EPA GHG standards. While the fleet-average GHG emissions
standards establish minimum standards, they do not limit the ability of manufacturers to achieve
further reductions, and any manufacturer that does will generate credits that can be used or sold.
ZEVs sold in California and other states will help a manufacturer to meet (or exceed) the EPA
GHG standards. More detail on these responses may be found in TSD Chapter 1.2.1.1 at p. 1-32.
Summary of Comments on the Proposed Determination
In comments on the Proposed Determination, several organizations provided comments
relating to the inclusion of ZEV vehicles in the baseline and reference cases.
Comments from American Council for an Energy-Efficient Economy (ACEEE), Union of
Concerned Scientists (UCS), Center for Biological Diversity (CBD), Natural Resources Defense
Council (NRDC), and the International Council on Clean Transportation (ICCT) explicitly
supported the way EPA accounted for ZEV vehicles. ACEEE echoed EPA's position that
inclusion of ZEVs in the reference fleet attributes neither benefits nor costs of ZEVs to the
federal program. CBD presented arguments supporting EPA's rationale for not including the
costs, noting that California's program, which has also been independently adopted by several
other states, is not part of the national program. NRDC states that the ZEV programs should be
treated similarly to the federal Tier 3 and California LEV III standard costs, and that the ZEV
programs are not GHG control programs despite their having an effect of lowering the cost of
compliance with the GHG program.
Comments from Global Automakers, the Alliance, Toyota, and NADA were largely similar to
comments received on the Draft TAR in that they continued to disagree with the inclusion of
ZEV vehicles without accounting for their cost. The Alliance stated that EPA dismissed its
38 Office of Management and Budget Circular No. A-4, "Regulatory Analysis," at page 15, available at
https://www.whitehouse.gov/omb/memoranda_m03-21.
100
-------
comments on this topic, and that EPA misunderstood its primary point that having the ZEV
mandate vehicles in the reference case lowers the need for other technologies. Global also
commented on inclusion of ZEV vehicles in the reference case and stated that the EPA response
was insufficient. NADA stated that it supports the position of the Alliance and Global. Toyota
echoed similar concerns and pointed out that the ZEV program has a history of revising its
targets, suggesting that actual ZEV penetration levels are uncertain.
The Global Automakers also commented that the ZEV program requires ZEV program
vehicle sales to increase to 15 percent or more by 2025. Global Automakers also argued that
EPA must consider the cumulative costs and burdens of the GHG regulations, the CAFE
requirements and the ZEV mandate; yet has not done so; and that EPA had included the benefits
of the ZEV mandate but not the costs of the mandate. Global Automakers argued that EPA had
created a reference case with lower GHG emissions than would otherwise exist, at zero cost,
thereby misleadingly suggesting that manufacturers can use only lower-cost technologies to meet
the federal standards.
Response to Comments on the Proposed Determination
EPA understands the comments from industry about this issue. As we see it, there are two
primary issues presented in these comments: (1) the impact of the ZEV program on the cost per
vehicle estimates; and (2) the impact of the ZEV program on the costs and benefits of the
MY2022-2025 GHG standards. On the first issue, EPA has been clear that the presence of the
ZEVs in the analysis fleet serves to reduce the cost per vehicle estimated by the analysis (see the
Proposed Determination TSD at page 1-33). EPA has not quantified the small impact on cost per
vehicle that occurs due to the inclusion of ZEV program vehicles. Given the ZEV program sales
of roughly 420,000 vehicles projected for MY2025, and the inclusion of upstream emissions in
the compliance CO2 level of BEVs and PHEVs, we believe that the cost per vehicle would be
only marginally higher were the ZEV program vehicles not included. However, we have not
modeled the program in this way, since we are following the OMB guidelines for establishing an
appropriate reference case reflecting "the way the world would look" absent the MY2022-2025
standards, as noted above. Since the ZEV regulations are independent of the national GHG
emissions standards, these standards would be in effect even if the MY2022-2025 standards did
not exist. The ZEV vehicles will in fact be part of the compliance determination for each
manufacturer's fleet, and thus we have modeled them in this manner.
On the second issue, EPA has not calculated the costs or benefits of the ZEV program for
California and the other states adopting the ZEV program because those costs and benefits are
not imposed or realized by the GHG standards.
Regarding the comment from Global Automakers that ZEV sales must increase to 15 percent
by 2025, the Draft TAR (Chapter 4.1.4.2.1) included a description of the CARB ZEV
requirements. The ZEV regulation establishes a minimum number of ZEV credits that must be
met, not an actual percentage of new car sales. Each ZEV earns a number of credits based on its
electric range. When adopted in 2012, ARB estimated it would take approximately 15 percent of
sales to be ZEVs to meet the credit requirement. However, since then, ZEV technology has
advanced more rapidly than anticipated and longer range BEVs and PHEVs are already on the
market, with even more announced for future release. In the original 15 percent calculation,
BEVs were assumed to have a real world range of 70 miles, and accordingly earn 1.5 credits per
vehicle. However, current BEVs, like many of the Tesla models and the GM Bolt EV, already
101
-------
have a range exceeding 200 miles and accordingly, earn more than 3 credits per vehicle. The
Proposed Determination analysis, like the Draft TAR, included BEVs with a 200 mile range.
Most importantly, CARB helped develop the ZEV program sales projections used in both the
Draft TAR and the Proposed Determination so the percentage of the fleet made up by ZEV
program vehicles is consistent with CARB's expectations at that time.
2.8 OMEGA
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Comments received on the Draft TAR pertaining to EPA's OMEGA model were primarily
oriented toward its modeling inputs and outputs rather than OMEGA itself or its algorithms.
While not pertaining directly to OMEGA, EPA noted that several OEM commenters
suggested that they plan to expand their use of off-cycle credits in the coming years, as
evidenced in part by their suggestions that EPA should consider removing the 10 g/mi cap on the
use of off-cycle menu technologies. Comments regarding the 10 g/mi cap, which are an off-cycle
topic, are discussed further in Chapter 3.9 of this RTC document. With regard to OMEGA, EPA
recognized that the commenters' emphasis on the importance of off-cycle credits (specifically,
those in the off-cycle technology menu) suggested that the OMEGA model could be improved
by incorporating additional representation of the potential for off-cycle credits to contribute to
potential compliance paths. In response, EPA built into OMEGA the ability to model use of two
additional levels of off-cycle credits that had not been modeled in previous analyses. More
discussion of this improvement may be found in the Proposed Determination Appendix A.2.7 at
p. A-15 and C.l.1.1.3.
In addition to those comments addressing off-cycle technologies and credits, we also received
comments on the Draft TAR suggesting that certain technologies—namely Atkinson 2 (advanced
Atkinson, ATK2+cooled EGR), the most advanced transmissions (TRX22), and mass
reduction—could not be adopted at the levels projected by OMEGA in the MY2022-2025
timeframe. As such, these commenters believed that the EPA analysis was relying too heavily on
technologies that in their opinion could not penetrate into the fleet by MY2025, in large part due
to lead time requirements and considerations related to global vehicle manufacturing and
platform sharing. EPA responded generally to these comments in Chapter 2.3.1.1 of the TSD,
and addressed lead time issues for specific technologies such as Atkinson 2 in the corresponding
technology sections of the TSD. To further address those concerns, for the Proposed
Determination we conducted several sensitivity runs that limited (i.e., artificially constrained)
penetration of advanced transmissions, mass reduction, and Atkinson 2. These sensitivity cases
showed that, while incremental costs did increase in some cases (e.g., the transmission-limited
pathway and the Atkinson-limited pathway), they decreased in others (e.g., the mass reduction-
limited pathway). Even where sensitivity cases showed incremental cost increases, compliance
was still feasible and at reasonable cost. These sensitivities and their implications were described
in brief in Section A.2.8 of the Proposed Determination Appendix (at p. A-16), with a full
description in Sections C.l.2.1.4 and C. 1.2.1.5 of the same document (at p. A-144 to A-146).
Summary of Comments on the Proposed Determination
In comments on the Proposed Determination, CARB commented, "Modeling results and
technology assumptions used in the analysis are robust." Other comments relating to OMEGA
102
-------
were, like the Draft TAR comments, primarily concerned with specific inputs and outputs to the
model, although some comment was received on the use and design of the model.
The Alliance and some individual OEMs made several comments regarding the modeling of
future compliance using the OMEGA model and commented on some of the underlying inputs to
the model. Specifically, the Alliance commented on the creation of a 2021 MY reference case
that complies with the 2021 MY standards and the performance of the 2012 to 2016 MY fleet
with respect to the EPA-projected performance in the 2010 FRM. In addition, the Alliance
commented on the number of changes made to individual vehicle models being projected by
OMEGA.
The Alliance also expressed concern with some aspects of the design of the OMEGA model
and what it describes as its "optimized" output, stating that OMEGA "generally seems to deliver
a level of fleet optimization that manufacturers cannot replicate." They argued that OMEGA
includes 7,200 vehicle technology combinations in the projected MY2025 fleet while the
MY2015 baseline fleet has only 2,600, and believe that this large number of technology
combinations is "not realistic from a technology investment or vehicle sales perspective." That
is, they believe that practical considerations not accounted for by the OMEGA model would
prevent manufacturers from building a compliant fleet as modeled, and the fleet they could
actually build would therefore be costlier than suggested by the "ideal" perspective of OMEGA.
Global Automakers comments that the "dramatic changes in the overall modeling results
[between the Draft TAR and Proposed Determination] raises significant questions about the
efficacy of EPA's modeling approach," and that it "further demonstrates that between the release
of the Draft TAR and the release of the Proposed Determination, there was either (a) dramatic
methodology/modeling revisions, (b) corrections to — or incorporation of new — data mistakes,
or (c) all of the above." They also state that "one would expect results that coalesce around
certain conclusions, not ones that wildly diverge."
Global Automakers also commented on the projected MY2025 penetration of Atkinson 2
having fallen from 44 percent in the Draft TAR to 27 percent in the Proposed Determination.
They state that it is "unclear from the Proposed Determination and the TSD how EPA arrived at
these changes." Citing EPA's description of the change in the Proposed Determination (as being
the result of refinements in effectiveness modeling that more appropriately reflect the relative
improvements allocated to advanced engines and transmissions in powertrain packages), they
state that industry should examine whether these refinements are reasonable. They also describe
the projected penetration rate of Atkinson 2 to be unreasonable given the time available.
Mercedes-Benz commented that EPA used, in the Draft TAR, both the Indirect Cost
Multiplier (ICM) and Retail Price Equivalent (RPE) approaches to estimate indirect costs, and
that these two approaches had produced "very different technology penetration rates for our
fleets." They described these package shifts as significant and therefore "call into question the
validity of the models used to make these assessments," and Recommended that they be
"investigated and corrected." They also said that the Proposed Determination did not discuss
why with the use of RPE vs. ICM has such an impact on projections for some manufacturers.
They also believe that results for the RPE case were not included in the Proposed Determination,
and that they would have indicated a higher level of technology and cost if they had been.
Mercedes also commented that the "large variation in agency modeling projections has made it
103
-------
difficult to understand how the model works and create comprehensive feedback to share with
the agency in such abbreviated time frames."
Mercedes-Benz also stated the belief that EPA's conclusions regarding the need for only low
levels of electrification may apply to the average U.S. fleet but do not apply to relatively low
volume, luxury-oriented manufacturers such as Mercedes-Benz due to limited opportunity to
spread its compliance obligations across their fleets. They argue that because they have already
added more fuel- efficient technologies than many other OEMs, they will need to add more
technology on its vehicles than the average U.S. fleet to avoid having to purchase credits from
(and therefore subsidize) competing automakers.
Mercedes-Benz also commented that EPA included too many model variants in its projected
compliance paths, stating that "it is not practical to implement so many unique configurations of
vehicles, especially for a low volume manufacturer."
Comments from Toyota criticized EPA's use of MY2021 as the basis for estimating MY2025
compliance costs, arguing that this "takes 2021 model year compliance as a given" and masks
what they expect to be a more challenging and costly path to MY2025 compliance than assumed
in the 2012FRM.
Toyota referred back to their comments on the Draft TAR and stated that some of their
feedback had not been fully addressed in the Proposed Determination, relating to what it
described as an error in allowing EPS to be applied to the LX570/Land Cruiser and the failure to
account for added weight of the AER02 package.
Toyota also asked why the Proposed Determination analysis revised the mix of passenger cars
and light-duty trucks to a lower truck share (53/47 percent car/truck) than the share represented
in the Draft TAR (52/48 percent car/truck) when sales trends seem to suggest the opposite trend.
General Motors also commented specifically on the differences in transmission penetration
rates between the Draft TAR and Proposed Determination, both for their fleet and for the fleet as
a whole.
FCA presented a chart showing that between the Draft TAR and Proposed Determination, the
projected percent mass reduction for FCA's fleet increased from 5.9 percent to 11.6 percent.
While the main responses to the issues raised in FCA's mass reduction-related comments are
provided in Chapter 2.5.12 of this RTC document (under the topic of mass reduction), the topic
of changes in projected mass reduction between the two analyses is addressed below in the
context of a similar comment by Mercedes-Benz.
Response to Comments on the Proposed Determination
Mercedes-Benz questions the different technology penetration rates generated when using the
ICM approach to estimating indirect costs relative to using the RPE approach, and mistakenly
argues that RPE results were not included in the Proposed Determination. RPE results were
included with the Proposed Determination as a sensitivity analysis. Different results with ICM vs
RPE should be expected since different inputs are being used. Importantly, EPA considers the
ICM approach to be the more appropriate approach and, as we explained in the Draft TAR, this
position is supported by many stakeholders (see Draft TAR, Chapter 5.3.2.2). Importantly, the
OMEGA runs, central and sensitivity, that we have conducted are meant to reflect possible
pathways toward compliance given different sets of inputs and constraints placed on some of the
104
-------
more important inputs. The outputs of those runs are meant to inform as to the possible
technology penetrations, costs, and environmental impacts as opposed to predicting a future
outcome. EPA makes no claim that the OMEGA results reflect the exact future fleet that any
given manufacturer will choose, but if the fleet were as indicated by OMEGA, we believe that
the manufacturer's GHG performance and costs would be consistent with those shown in the
outputs.
The Mercedes-Benz comments also claim that large differences exist between the Draft TAR
and the Proposed Determination in terms of technology penetration rates, and include a chart
showing the differences (Figure 5, Mercedes-Benz comments). In reviewing the Mercedes
comments, EPA sees considerable consistency in many of the more important technologies. The
MHEV48V penetration rates when using ICMs are quite consistent. Only the RPE case results in
a notable difference, and then only for cars. For the Mercedes fleet, the MHEV48V rates range
from 56 percent to 75 percent. This does not seem large given the different fleets and the
different cost inputs used. For PEVs, the fleet penetration rates range from 7 to 9 percent for
BEVs and 0 to 3 percent for PHEVs. Again, the difference in these results is small. The lack of
penetration of BEVs or PHEVs for trucks in the Draft TAR runs is explained by the fact that we
constrained the model's application of those technologies—for both ZEV program sales or for
compliance—on nearly all vehicles placed on the truck standard curves. Therefore, the
penetration rates would naturally be 0 percent unless there already existed some in the Mercedes-
Benz baseline fleet. For stop-start, the truck penetration rates are nearly identical, while the car
penetration rates change considerably depending on the run. This is most likely due to the
updates made to the Atkinson 2 and TRX22 effectiveness values along with the updates made to
exemplar vehicles. Note that the Draft TAR ICM run pushed the Mercedes car fleet further into
TDS24 and MHEV48V. In the Proposed Determination, the ICM run pushed the Mercedes car
fleet into Atkinson 2 and stop-start, and stopped short of adding as much TDS24 and
MHEV48V.
Missing from the Mercedes chart is differences in mass reduction, where the Draft TAR
showed the Mercedes car/truck/fleet as adding 11.5/6.8/9.7 percent mass reduction, respectively.
In the Proposed Determination, those percentages were 13.4/20/16.1 percent, respectively. This
difference in mass reduction penetrations would be expected to have impacts on the penetrations
of other technologies and on costs. The increase in mass reduction in the Proposed
Determination run would be an expected result of the placement of more vehicles, for all
manufacturers, onto the unibody-based mass reduction cost curve. There are significantly
different mass reduction alternatives and associated costs associated with the primary vehicle
architectures of unibody vs. body-on-frame. EPA believes it is important that each vehicle's
mass reduction be considered within the context of its primary body architecture. In the Proposed
Determination, Mercedes' entire fleet is placed on the unibody-based mass reduction cost curve
and only true pickups are placed on the body-on-frame mass reduction cost curve. All other
vehicles, even those placed on the truck standard curve, are placed on the unibody-based mass
reduction cost curve. Table 2-5 shows a breakdown of the percentage of each manufacturer's
fleet placed on each of the mass reduction cost curves in both the Draft TAR and the Proposed
Determination. As shown in Table 2-5, the cost of mass reduction for the entire Mercedes fleet is
based on the lower cost unibody-based curve, resulting in mass reduction being considerably
more cost effective and, hence, more mass reduction is projected for Mercedes. This is true for
105
-------
other manufacturers, and helps explain the increased levels of mass reduction for the FCA fleet
as questioned by FCA and Global.
Table 2-5 Percentage of the Fleet Placed on the Car-based versus Truck-based Mass Reduction Cost Curve
Draft TAR
Proposed Determination
Unibody Cost
Body-on-frame Cost
Unibody Cost
Body-on-frame Cost
Curve
Curve
Curve
Curve
BMW
76%
24%
100%
0%
FCA
38%
62%
83%
17%
FORD
34%
66%
74%
26%
GM
46%
54%
76%
24%
HONDA
68%
32%
100%
0%
HYUNDAI/KIA
89%
11%
100%
0%
JLR
21%
79%
100%
0%
MAZDA
91%
9%
100%
0%
MERCEDES
56%
44%
100%
0%
MITSUBISHI
96%
4%
100%
0%
NISSAN
71%
29%
93%
7%
SUBARU
95%
5%
100%
0%
TESLA
100%
0%
100%
0%
TOYOTA
67%
33%
87%
13%
VOLKSWAGEN
56%
44%
100%
0%
VOLVO
38%
62%
100%
0%
Fleet
59%
41%
89%
11%
This change also helps to explain some of the other cost differences between the Draft TAR
and the Proposed Determination as highlighted by Global Automakers. Importantly, the
allowance of plug-in hybrid and full electrification is consistent with the mass reduction cost
curve. As a result, more vehicles placed on the truck standard curve are now allowed these levels
of electrification in OMEGA where, in the Draft TAR, they were not. In the Draft TAR, these
levels of electrification were limited almost exclusively to cars and the smallest of cross-over
utility vehicles. In the Proposed Determination, all non-pickups are allowed to electrify. This
also impacts the ZEV program vehicle fleet since many ZEVs are now placed on the truck
standard curve where in the Draft TAR they were almost exclusively on the car standard curve.
These changes help explain some of the cost deltas questioned by Global.
106
-------
Table 2-6 shows the car/truck/fleet absolute costs in both the Draft TAR and the Proposed
Determination.
Table 2-6 Absolute Per Vehicle Costs in the Draft TAR and the Proposed Determination
Draft TAR (2013$)
Proposed Determination
Increase (dollar basis is
(2015$)
inconsistent)
Car
Truck
Fleet
Car
Truck
Fleet
Car
Truck
Fleet
BMW
$1,724
$1,942
$1,776
$2,212
$3,315
$2,467
$488
$1,373
$691
FCA
$1,789
$2,451
$2,254
$1,993
$2,566
$2,391
$204
$115
$137
FORD
$969
$1,777
$1,438
$1,203
$1,374
$1,303
$234
-$403
-$135
GM
$1,169
$2,248
$1,707
$1,293
$1,446
$1,368
$124
-$802
-$339
HONDA
$842
$967
$901
$880
$1,133
$1,003
$38
$166
$102
HYUNDAI/KIA
$1,447
$2,128
$1,529
$1,386
$1,584
$1,418
-$61
-$544
-$111
JLR
$5,090
$3,436
$3,782
$1,982
$3,377
$3,090
-$3,108
-$59
-$692
MAZDA
$772
$1,081
$866
$674
$1,300
$909
-$98
$219
$43
MERCEDES
$2,482
$2,732
$2,577
$2,217
$3,028
$2,551
-$265
$296
-$26
MITSUBISHI
$1,178
$1,333
$1,234
$1,590
$1,742
$1,649
$412
$409
$415
NISSAN
$1,148
$1,526
$1,298
$1,207
$1,669
$1,383
$59
$143
$85
SUBARU
$686
$691
$690
$720
$888
$852
$34
$197
$162
TESLA
$140
$0
$140
$143
$143
$143
$3
$143
$3
TOYOTA
$884
$1,547
$1,184
$838
$1,482
$1,123
-$46
-$65
-$61
VOLKSWAGEN
$2,751
$2,560
$2,679
$2,441
$3,869
$2,879
-$310
$1,309
$200
VOLVO
$2,351
$3,170
$2,777
$1,120
$2,084
$1,607
-$1,231
-$1,086
-$1,170
Fleet
$1,293
$1,864
$1,565
$1,327
$1,740
$1,521
$34
-$124
-$44
The most obvious change between the two analyses is to JLR car costs, decreasing from
$5000 to $3100. This is largely due to the reduced BEV penetration of the JLR car fleet in the
Proposed Determination where the technology penetration was 2 percent while in the Draft TAR
it was 30 percent. This large reduction in BEV penetration is in part due to the truck BEV
penetration having increased from 0 to 7 percent, since many trucks were "allowed" to electrify
in the Proposed Determination analysis while there existed more restrictions on electrification of
vehicles placed on the truck standard curve in the Draft TAR. This begs the question of why
there was not a large increase in JLR truck costs. This is due to the decreased mass reduction
costs, since JLR's entire truck fleet is placed on the unibody cost curve. Further, JLR's baseline
level of mass reduction on trucks in the Draft TAR was 11.9 percent, while in the Proposed
Determination it was 3.6 percent. That alone would serve to make mass reduction on JLR trucks
much costlier in the Draft TAR than in the Proposed Determination. If we look at JLR's
achieved CO2 levels, they were at 102 grams/mile in the Draft TAR and just 192 grams per mile
in the Proposed Determination. In the Draft TAR, with zero upstream emissions on BEVs,
OMEGA found BEVs to be an attractive means of compliance for JLR. In the Proposed
Determination, which unlike the Draft TAR includes an accounting for upstream emissions in
compliance calculations, those BEVs became less attractive, especially when truck mass
reduction was less costly due to use of the unibody cost curve.
Global also highlighted Ford's cost changes. The Ford truck cost decrease is due largely to
many of Ford's trucks being placed on the unibody cost curve. In the Draft TAR, nearly all of
107
-------
Ford's trucks were placed on the body-on-frame cost curve. For some of these trucks, such as the
small SUV Ford Escape, this placement was inappropriate. Ford's Escapes applied a full 5
percent more mass reduction in the Proposed Determination and for lower costs than in the Draft
TAR. As for Ford cars, their achieved CO2 has decreased from 166 g/mi to 161 g/mi, so more
control has been placed on Ford cars in the Proposed Determination, most likely due to the
inclusion of upstream emissions on PEVs.
For BMW, the increased car costs appear to be the result of BMW's projected penetration of
PEVs for meeting the standards in both the Draft TAR and the Proposed Determination, although
more of it in the Proposed Determination due to the inclusion of upstream emissions. In the
Proposed Determination, the bulk of that electrification is occurring in the truck fleet (which was
not allowed in the Draft TAR), which serves to increase truck costs. This requires less
electrification in the car fleet but BMW has pushed further into turbo/downsizing on their car
fleet in the Proposed Determination, including a 12 percent penetration of Miller-cycle where the
Draft TAR had none. These serve to increase somewhat the BMW car costs.
For VW, the car costs have decreased due to less electrification (from 14 percent down to 5
percent) while the truck costs have increased due to increased electrification (from 0 percent to
18 percent).
The Volvo changes are also quite large. This appears to be largely the result of Volvo's
baseline mass reduction level having been characterized as 6/7 percent and 3.8 percent for cars
and trucks, respectively, in the Draft TAR, and then 2.9 percent and 0.0 percent for cars and
trucks, respectively, in the Proposed Determination. This, along with the movement of vehicles
from the body-on-frame mass reduction cost curve to the unibody cost curve has resulted in
considerable changes to the Volvo estimates.
Global also highlighted some other differences by focusing on percentage changes between
the Draft TAR and the Proposed Determination, such as Honda trucks increasing 47 percent and
Mazda cars decreasing 35 percent. However, in absolute dollar terms, those changes were just
$166 and $98, respectively, which again seem relatively consistent in EPA's view and not
significantly changed as argued by Global. We do not mean to imply that $200 is a trivial
amount; however, $200 is less than 25 percent of our incremental cost estimate of $875/vehicle
and just 13 percent of our absolute cost estimate of $1521/vehicle. EPA believes the relative
consistency in the combined fleet costs per vehicle is important to consider given the role credit
transfers play in OMEGA which can drive fluctuations in the car/truck share of those combined
fleet costs. While there are differences, many manufacturer's combined fleet costs have remained
within $200 per vehicle which EPA believes represents a reasonable consistency given that we
are projecting forward 11 and 10 model years from our MY2014 and MY2015 based baseline
fleets.
It is also important to point out that we conducted a sensitivity that allowed no additional
mass reduction beyond that in the baseline fleet. So, while the use of the lower cost unibody
mass reduction curve on a higher percentage of the fleet in the Proposed Determination has a
significant impact on technology penetrations and costs, the impact of using different costs
should not be higher than the impacts of the sensitivity that allowed no additional mass reduction
beyond that in the baseline fleet.
108
-------
An important takeaway from these comparisons between the Draft TAR and the Proposed
Determination is the amount of available technology with similar cost effectiveness
characteristics. The relative similarities of these technologies in terms of cost effectiveness along
with EPA's allowance of transfers between car/truck fleets can mean that even very small
changes in technology cost and effectiveness can lead to changes in the technology penetrations
of other technologies with similar cost-effectiveness, but the fleet costs and the conclusions that
can be drawn from the results have remained relatively consistent throughout our analyses.
Mercedes and the automaker associations also were critical of the amount of variation in their
fleets as projected by OMEGA. We disagree with the suggestion that there is a proliferation of
variation and believe that the perception expressed by these commenters stems from a
misunderstanding or misinterpretation of results relative to the manner in which EPA interprets
the results. OMEGA is not meant to reflect a movement of each individual vehicle through the
design process, year-over-year, toward compliance with the standards. OMEGA is meant to be a
single model year's optimized look at the technology application and associated costs capable of
delivering compliance with the standards. As a result, while OMEGA output files show
individual vehicles containing sometimes multiple packages of technology, the fact is that
OMEGA works on platforms (these are OMEGA platforms and not necessarily what an
automaker would call a platform) containing many vehicles and applies technology to those
platforms. OMEGA is indifferent as to which individual vehicles within the given platform
actually apply the technology. For example, assume an OMEGA platform contained two
vehicles, each a 50 percent sales share within the platform, and OMEGA converted 50 percent of
that platform to a turbocharged/downsized configuration. One could interpret that to mean that
half of each vehicle on the platform is converted, or that all of one vehicle was converted while
the other remained unchanged. As said, OMEGA is indifferent to this. OMEGA would generate
output files that show the technology penetration as half of each vehicle being converted.
However, this is not necessarily the case - it is but one interpretation of the output. The OMEGA
code necessary to generate these technology tracking output files was developed after completion
of the 2012 FRM and was done only in an effort to make technology penetration rates more
easily gleaned from a given run; the code and its output file was not meant to suggest the actual
application of each technology to each individual vehicle but rather one possible application of
each technology to each individual vehicle.
That being said, EPA believes that a more reasonable approach when considering
proliferation of technology is to look at engine/transmission combinations within the OMEGA
baseline and compare that to the combinations in the OMEGA control case. For this look, EPA
has focused on the ICM, AEO reference fuel price run. We have considered the effective and
actual number of cylinders (which includes the presence or lack of turbocharging), the valvetrain
configuration (DOHC, SOHC, OHV) and the transmission type in terms of the TRX11 through
TRX22 coding used for modeling. We have also considered as unique "platforms" all hybrid,
plug-in hybrid and battery electric vehicles. In other words, these are the engine-transmission
pairings in the OMEGA baseline and control case fleets. We have not considered as unique
"platforms" any vehicles with stop-start or mild hybrid technology because we consider those
technologies to be relatively easy to add to vehicles as necessary for improving CO2
performance.
Some background first on "OMEGA platforms." An OMEGA platform is determined by the
OEM, the curb weight class (CWC), the vehicle type (1 through 29) and the car/truck
109
-------
determination. One example would be the BMW CWC215C which would represent a BMW
car in curb weight class 2 mapped to vehicle type 15. True OEM platforms are generally
determined by common mechanical elements, such as: floor pan, wheelbase, steering
mechanism, suspension type, engine type/placement; transmission type/placement. As a result, a
sedan and coupe and even a crossover utility vehicle might be built on the same platform
provided their wheelbases, suspension, engine and transmissions were similar. For example, the
Ford Focus, Escape, C-Max and Transit Connect are built on the same platform. Importantly,
while built on the same platform, there could be as many as 4 or 5 different engines and 2 or 3
transmissions available for this set of vehicles.
The point here is that OEM and OMEGA platforms are not one-to-one, and were never
intended to be. In the Ford example above, the Focus and Escape are clearly different curb
weight vehicles despite being built on the same platform. In OMEGA, the curb weight
differences would result in different OMEGA platforms. OMEGA needs that difference for
determining different effectiveness values for the different curb weight classes.
The automaker comments focus on the large number of technology combinations resulting
from the OMEGA runs. The Alliance argues that the baseline fleet consists of 2,600 vehicles and
moves to 7,200 vehicles in the control case run, a near tripling of technology combinations. The
methodology used in their determination of 7,200 vehicles, was not documented in the comment.
However, if we look at our baseline fleet and break that fleet into different
OEM/engine/transmission groupings, where the engine determination consists of actual number
of cylinders, effective number of cylinders (i.e., a turbo charged 4 cylinder engine would be
considered to have an effective 6 cylinders), valvetrain configuration (DOHC, SOHC, OHV),
and transmission (consists of type (manual, automated manual, automatic, continuously variable)
and number of gears), the baseline fleet would have 183 unique combinations. Again, these are
not meant to reflect OEM platforms, nor are they OMEGA platforms since the curb weight class
is not included.39
As this exercise clearly demonstrates, there exists considerable variety in the baseline fleet.
Table 2-7 shows each manufacturer's baseline and control case fleets broken into categories of
engines treated within OMEGA as 4, 6 and 8 cylinder engines, diesels, plug-in hybrids and
battery electric vehicles. A quick look at the FCA baseline fleet shows considerable variety with
four-cylinder dual-overhead cam engines equipped with four different transmissions and four
cylinder single overhead-cam engines equipped with three different transmissions. OMEGA
chooses to simplify this variety by converting FCA's four-cylinder fleet to Atkinson 2 while also
maintaining two transmissions.
There are also cases where OMEGA increases variability in the fleet; however, manufacturers
are expected to manage their product plans and compliance strategy, among other factors such as
customer preference, to keep their costs to a minimum. An example of this is GM's four-cylinder
39 Note that, when running OMEGA, there are 294 unique platforms stemming from the characterization of the pre-
ZEV fleet due to the inclusion of curb weight class in the platform determination. Of that 294, 253 serve as
possible ZEV-source platforms for modeling in OMEGA (i.e., 253 are non-pickup truck and non-EV/non-PHEV
in the baseline). We then add 253 OMEGA platforms to reflect ZEV program BEVs and 253 more OMEGA
platforms reflect the ZEV program PHEVs stemming from the 253 ZEV-source platforms, the end result being
800 unique OMEGA platforms. By doing this, we are not suggesting that ZEV program vehicles will be built on
unique OEM platforms. Their platforms are unique within OMEGA only for modeling purposes within OMEGA.
110
-------
vehicles. In the baseline, GM has four-cylinder vehicles with three different transmissions. After
running OMEGA, two of those engines remain and the third has improved its transmission from
TRX21 to TRX22. In addition, OMEGA has converted some engines to Atkinson-2 with two
different transmissions. So, GM's four-cylinder engines have gone from three
engine/transmission combinations to five. However, in reality, GM could move all four-cylinder
engines to Atkinson 2, at some cost but also at a lowering of their fleet CO2. The lowering of
their fleet CO2 on those four-cylinder engines could allow for less technology on some six- or
eight-cylinder engines, thereby reducing six- and eight-cylinder engine costs and, thereby
offsetting the increased costs on four-cylinder engines. The costs may or may not perfectly
balance, but that is part of the optimization goal of OMEGA - to seek the most cost-effective
approach to compliance. Given the variety in six- and eight-cylinder engine/transmission
pairings in GM's baseline fleet, OMEGA's move to five such pairings in the four-cylinder
control case does not appear unreasonable.
Most cases where OMEGA has increased variety compared to the baseline fleet are for those
manufacturers that offer several transmissions in the baseline fleet. For the most part, OMEGA
leaves 6 speed manual transmissions alone under the assumption that manufacturers offer those
transmissions because some buyers insist on them. Since OMEGA leaves them alone, the control
case fleet will generally have at least as many transmission offerings as were present in the
baseline fleet (with the exception of movement away from any transmissions having less than six
gears, denoted as TRXOO in Table 2-7).
Table 2-7 OEM, Engine Technology and Transmission Pairings in the OMEGA Baseline (pre-ZEV) and the
OMEGA Control Case Fleets
OEM
General
Engine
OMEGA Baseline
OMEGA Control in 2025
BMW
Engines
BMW
E04
A04
DOHC
M6
BMW
E04
A04
ATK2 DOHC M6
treated as
BMW
E04
A04
DOHC
TRX11
BMW
E04
A04
ATK2 DOHC TRX22
4 cylinders
Engines
BMW
E05
AO 3
TDS18
DOHC
M6
BMW
E06
A04
TDS18 DOHC M6
treated as
BMW
E05
AO 3
TDS18
DOHC
TRX11
BMW
E06
A04
TDS18 DOHC TRX22
6 cylinders
BMW
E06
A04
TDS18
DOHC
M6
BMW
E06
A04
TDS24 DOHC M6
BMW
E06
A04
TDS18
DOHC
TRX11
BMW
E06
A04
TDS24 DOHC TRX22
BMW.
_E06_
_A04_
JDS 18.
_DOHC_
TRX21
BMW.
BMW
_E06_
E06
_A06_
AO 6
_ATK2_DOHC_M6
ATK2 DOHC TRX22
Engines
BMW
E08
AO 6
TDS18
DOHC
M6
BMW
E08
AO 6
TDS18 DOHC TRX21
treated as
BMW
E08
AO 6
TDS18
DOHC
TRX11
BMW
E08
A08
DOHC TRX22
8 cylinders
BMW
E08
AO 6
TDS18
DOHC
TRX21
BMW
E08
A08
ATK2 DOHC TRX22
BMW
E08
AO 6
P2 TDS18 DOHC TRX21
BMW
E08
A08
ATK2 TURBM DOHC
M6
BMW
E10
A08
TDS18
DOHC
M6
BMW
E08
A08
ATK2 TURBM DOHC
TRX22
BMW
E10
A08
TDS18
DOHC
TRX21
BMW
E14
A12
TDS18
DOHC
TRX21
BMW
E12
A12
DOHC
TRX21
Diesels
BMW
E06
A04
DOHC
DSL TRX21
BMW
E06
A04
DSL TRX22
BMW
E08
AO 6
DOHC
DSL TRX21
BMW
E08
AO 6
DSL TRX22
EV& PHEV
BMW
E02
AO 2
SOHC
REEV40
BMW
REEV40
BMW
E05
AO 3
TDS18
DOHC
REEV40
BMW
E04
A04
TDS18 REEV40
BMW
EV75
BMW
EV75
BMW
EV100
BMW_EV200 (ZEV)
111
-------
FCA
Engines
FC A_ E 04_A04_ DO H C_T RXOO
FC A_ E 04_A04_AT K2_DOHC_M6
treated as
FC A_ E 04_A04_ DO H C_ M 6
FC A_ E 04_A04_AT K2_D 0 H C_T RX2 2
4 cylinders
FC A_ E 04_A04_ DO H C_T RX11
FC A_ E 04_A04_ DO H C_T RX21
FCA_E04_A04_SOHC_M6
FC A_ E 04_A04_S 0 H C_T RX 11
FC A_ E 04_A04_S 0 H C_T RX21
Engines
FCA_E06_A04_TDS18_DOHC_TRX11
FCA_E06_A04_TDS18_DOHC_M6
treated as
FCA_E06_A04_TDS18_SOHC_M6
FCA_E06_A04_TDS18_DOHC_TRX22
6 cylinders
FCA_E06_A04_TDS18_SOHC_TRX11
FCA_E06_A04_TDS24_DOHC_M6
FCA_E06_A06_DOHC_TRX00
FCA_E06_A04_TDS24_DOHC_TRX22
FCA_E06_A06_DOHC_M6
FCA_E06_A06_ATK2_DOHC_M6
FCA_E06_A06_DOHC_TRX11
FCA_E06_A06_ATK2_DOHC_TRX22
FCA_E06_A06_DOHC_TRX21
Engines
FCA_E08_A08_DOHC_TRX11
FCA_E08_A08_OHV_M6
treated as
FCA_E08_A08_xOHV_M6
FCA_E08_A08_OHV_TRX22
8 cylinders
FCA_E08_A08_xOHV_TRXll
FCA_E08_A08_DOHC_TRX22
FCA_E08_A08_xOHV_TRX21
FCA_E08_A08_ATK2_DOHC_TRX22
FCA_E08_A06_DOHC_TRX21
FCA_E08_A08_ATK2_TURBM_DOHC_TRX22
FC A_ E 10_A 10_x 0 H V_ M 6
FCA_E08_A06_TDS18_DOHC_TRX22
FCA_E10_A08_TDS18_xOHV_M6
FCA_E08_A06_TDS24_DOHC_TRX22
FCA_E 10_A08_TDS 18_xOHV_TRX21
FCA_E10_A08_TDS18_DOHC_TRX21
Diesels
FCA_E08_A06_DOHC_DSL_TRX21
FCA_E08_A06_DSL_TRX22
EV& PHEV
FCA_EV75
FCA_EV75
FCA_EV200 (ZEV)
FCA_REEV40 (ZEV)
Ford
Engines
Ford_E04_A04_DOHC_M6
Ford_E04_A04_DOHC_M6
treated as
Ford_E04_A04_DOHC_TRXll
Ford_E04_A04_DOHC_TRX22
4 cylinders
Ford_E04_A04_DOHC_TRX21
Ford_E04_A04_ATK2_DOHC_TRX22
Ford_E04_A04_P2_TRX21
Ford_E04_A04_P2_TRX21
Engines
Ford_E05_A03_TDS18_DOHC_M6
Ford_E06_A04_TDS18_DOHC_M6
treated as
Ford_E06_A04_TDS18_DOHC_M6
Ford_E06_A04_TDS18_DOHC_TRX22
6 cylinders
Ford_E06_A04_TDS18_DOHC_TRXll
Ford_E06_A06_DOHC_M6
Ford_E06_A06_DOHC_TRXll
Engines
Ford_E08_A06_TDS18_DOHC_TRXll
Ford_E08_A06_TDS18_DOHC_TRX22
treated as
Ford_E08_A08_DOHC_M6
Ford_E08_A08_DOHC_M6
8 cylinders
Ford_E08_A08_DOHC_TRXll
Ford_E08_A08_DOHC_TRX22
Ford_E10_A08_TDS18_DOHC_M6
Ford_E08_A08_ATK2_DOHC_TRX22
Ford_E10_A08_TDS18_DOHC_TRXll
EV& PHEV
Ford_E04_A04_DOHC_REEV40
Ford_REEV40
Ford_EV75
Ford_EV75
Ford_EV200 (ZEV)
GM
Engines
GM_E04_A04_DOHC_M6
GM_E04_A04_DOHC_M6
treated as
GM_E04_A04_DOHC_TRX11
GM_E04_A04_DOHC_TRX11
4 cylinders
GM_E04_A04_DOHC_TRX21
GM_E04_A04_DOHC_TRX22
GM_E04_A04_ATK2_DOHC_M6
GM_E04_A04_ATK2_DOHC_TRX22
112
-------
Engines
GM_E06_A04_TDS18_DOHC_M6
GM_E06_A04_TDS18_DOHC_M6
treated as
GM_E06_A04_TDS18_DOHC_TRX11
GM_E06_A04_TDS18_DOHC_TRX22
6 cylinders
GM_E06_A06_DOHC_M6
GM_E06_A06_OHV_TRX22
GM_E06_A06_DOHC_TRX11
GM_E06_A06_DOHC_TRX21
GM_E06_A06_xOHV_TRXll
Engines
GM_E08_A06_TDS18_DOHC_TRX21
GM_E08_A06_TDS18_DOHC_TRX22
treated as
GM_E08_A06_TDS18_DOHC_TRX11
GM_E08_A08_OHV_TRX22
8 cylinders
GM_E08_A08_xOHV_M6
GM_E08_A08_xOHV_TRXll
GM_E08_A08_xOHV_TRX21
GM_E10_A08_TDS18_xOHV_M6
GM_E10_A08_TDS18_xOHV_TRXll
G M_E10_A08_TDS 18_xOHV_TRX21
Diesels
GM_E06_A04_DOHC_DSL_TRX11
GM_E06_A04_DSL_TRX22
EV& PHEV
G M_E04_A04_RE EV40
GM_REEV40
GM_EV75
GM_EV75
GM_EV200 (ZEV)
Honda
Engines
Honda_E04_A04_DOHC_TRX00
Honda_E04_A04_DOHC_M6
treated as
H o n d a_E 04_A04_S 0 H C_T RXOO
Honda_E04_A04_DOHC_TRX22
4 cylinders
Honda_E04_A04_SOHC_M6
Honda_E04_A04_ATK2_DOHC_M6
Honda_E04_A04_SOHC_TRX21
Honda_E04_A04_ATK2_DOHC_TRX22
Honda_E04_A04_DOHC_M6
Honda_E04_A04_P2_M6
Honda_E04_A04_DOHC_TRX21
Honda_E04_A04_P2_TRX21
Honda_E04_A04_SOHC_TRX21
Honda_E04_A04_P2_M6
Honda_E04_A04_P2_TRX21
Engines
Honda_E06_A06_SOHC_TRX00
Honda_E06_A06_DOHC_M6
treated as
Honda_E06_A06_SOHC_M6
Honda_E06_A06_DOHC_TRX22
6 cylinders
Honda_E06_A06_SOHC_TRXll
Honda_E06_A06_SOHC_TRX21
Engines
None
None
treated as
8 cylinders
EV& PHEV
None
Honda_EV200 (ZEV)
Honda_REEV40 (ZEV)
Hyundai/
Engines
Hyundai/Kia_E04_A04_DOHC_M6
Hyundai/Kia_E04_A04_DOHC_M6
Kia
treated as
Hyundai/Kia_E04_A04_DOHC_TRXll
Hyundai/Kia_E04_A04_DOHC_TRX21
4 cylinders
Hyundai/Kia_E04_A04_P2_TRXll
Hyundai/Kia_E04_A04_DOHC_TRX22
Hyundai/Kia_E04_A04_ATK2_DOHC_M6
Hyundai/Kia_E04_A04_ATK2_DOHC_TRX22
Hyundai/Kia_E04_A04_P2_TRXll
Engines
Hyundai/Kia_E06_A04_TDS18_DOHC_M6
Hyundai/Kia_E06_A04_TDS18_DOHC_M6
treated as
Hyundai/Kia_E06_A04_TDS18_DOHC_TRXll
Hyundai/Kia_E06_A04_TDS18_DOHC_TRX22
6 cylinders
Hyundai/Kia_E06_A04_TDS18_DOHC_TRX21
Hyundai/Kia_E06_A06_DOHC_M6
Hyundai/Kia_E06_A06_DOHC_TRXll
Hyundai/Kia_E06_A06_DOHC_TRX21
Engines
Hyundai/Kia_E08_A08_DOHC_TRX21
Hyundai/Kia_E08_A08_DOHC_TRX22
treated as
8 cylinders
EV& PHEV
Hyundai/Kia_EV75
Hyundai/Kia_EV75
113
-------
Hyundai/Kia_EV200 (ZEV)
Hyundai/Kia_REEV40 (ZEV)
JLR
Engines
None
None
treated as
4 cylinders
Engines
JLR E06 A04 TDS18 DOHC TRX21
JLR E06 A04 TDS18 DOHC TRX22
treated as
JLR E06 A04 TDS24 DOHC TRX22
6 cylinders
Engines
JLR E08 A06 TDS18 DOHC TRX21
JLR E08 A06 TDS18 DOHC TRX22
treated as
JLR E08 A08 DOHC TRX11
JLR E08 A08 DOHC TRX22
8 cylinders
JLR E10 A08 TDS18 DOHC TRX11
JLR E08 A08 ATK2 DOHC TRX22
JLR E10 A08 TDS18 DOHC TRX21
JLR E08 A08 ATK2 TURBM DOHC TRX22
EV& PHEV
None
JLR EV75
JLR EV100
JLR_EV200 (ZEV)
JLR_REEV40 (ZEV)
Mazda
Engines
Mazda E04 A04 DOHC M6
Mazda E04 A04 DOHC M6
treated as
Mazda E04 A04 DOHC TRXOO
Mazda E04 A04 DOHC TRX22
4 cylinders
Mazda E04 A04 DOHC TRX11
Mazda E04 A04 ATK2 DOHC M6
Mazda E04 A04 ATK2 DOHC M6
Mazda E04 A04 ATK2 DOHC TRX21
Mazda E04 A04 ATK2 DOHC TRX11
Mazda E04 A04 ATK2 DOHC TRX22
Engines
Mazda E06 A06 DOHC TRX11
Mazda E06 A04 TDS18 DOHC TRX22
treated as
6 cylinders
Engines
None
None
treated as
8 cylinders
EV& PHEV
None
Mazda_EV200 (ZEV)
Mazda_REEV40 (ZEV)
Mercedes
Engines
Mercedes E03 A03 DOHC TRXOO
Mercedes E04 A04 ATK2 DOHC TRX22
treated as
4 cylinders
Engines
Mercedes E06 A04 TDS18 DOHC M6
Mercedes E06 A04 TDS18 DOHC M6
treated as
Mercedes E06 A04 TDS18 DOHC TRX21
Mercedes E06 A04 TDS18 DOHC TRX22
6 cylinders
Mercedes E06 A06 TDS18 DOHC TRX21
Mercedes E06 A04 TDS24 DOHC TRX22
Mercedes E06 A06 P2 DOHC TRX21
Mercedes E06 A06 ATK2 DOHC M6
Mercedes E06 A06 ATK2 DOHC TRX22
Mercedes E06 A06 P2 TRX21
Engines
Mercedes E08 A06 TDS18 DOHC TRX21
Mercedes E08 A06 ATK2 TURBM DOHC TRX22
treated as
Mercedes E08 A08 DOHC TRX21
Mercedes E08 A08 DOHC TRX22
8 cylinders
Mercedes E08 A08 SOHC TRX21
Mercedes E08 A08 SOHC TRX22
Mercedes E10 A08 TDS18 DOHC TRX21
Mercedes E08 A08 ATK2 DOHC TRX22
Mercedes E14 A12 TDS18 SOHC TRX21
Mercedes E08 A08 ATK2 TURBM DOHC TRX22
Mercedes E14 A12 TDS18 DOHC TRX21
Mercedes E08 A08 P2 TRX22
Diesels
Mercedes E06 A04 DOHC DSL TRX21
Mercedes E06 A04 DSL TRX22
Mercedes E08 A06 DOHC DSL TRX21
Mercedes E08 A06 DSL TRX22
EV& PHEV
Mercedes E08 A06 TDS18 REEV40
Mercedes TDS18 REEV40 TRX21
Mercedes EV75
Mercedes EV75
Mercedes EV100
Merced es_EV200
114
-------
Mitsubishi
Engines
Mitsubishi E03 A03 DOHC M6
Mitsubishi E04 A04 DOHC M6
treated as
Mitsubishi E03 A03 DOHC TRX21
Mitsubishi E04 A04 DOHC TRX22
4 cylinders
Mitsubishi E04 A04 SOHC M6
Mitsubishi E04 A04 ATK2 DOHC M6
Mitsubishi E04 A04 SOHC TRX21
Mitsubishi E04 A04 ATK2 DOHC TRX22
Engines
Mitsubishi E06 A04 TDS18 SOHC M6
Mitsubishi E06 A04 TDS18 DOHC M6
treated as
Mitsubishi E06 A06 SOHC TRX11
Mitsubishi E06 A04 TDS18 DOHC TRX22
6 cylinders
Mitsubishi E06 A04 TDS18 SOHC TRX11
Engines
None
None
treated as
8 cylinders
EV& PHEV
None
Mitsubishi_EV200 (ZEV)
Mitsubishi_REEV40 (ZEV)
Nissan
Engines
Nissan E04 A04 DOHC TRXOO
Nissan E04 A04 DOHC M6
treated as
Nissan E04 A04 DOHC M6
Nissan E04 A04 DOHC TRX22
4 cylinders
Nissan E04 A04 DOHC TRX21
Nissan E04 A04 ATK2 DOHC M6
Nissan E04 A04 ATK2 DOHC TRX22
Nissan E04 A04 TDS18 DOHC M6
Nissan E04 A04 TDS18 DOHC TRX22
Engines
Nissan E06 A04 TDS18 DOHC M6
Nissan E06 A04 TDS18 DOHC M6
treated as
Nissan E06 A04 TDS18 DOHC TRX21
Nissan E06 A04 TDS18 DOHC TRX22
6 cylinders
Nissan E06 A04 P2 TDS18 DOHC TRX21
Nissan E06 A04 TDS24 DOHC TRX22
Nissan E06 A06 DOHC TRXOO
Nissan E06 A04 TDS18 P2 TRX21
Nissan E06 A06 DOHC M6
Nissan E06 A06 P2 TRX21
Nissan E06 A06 DOHC TRX21
Nissan E06 A06 P2 DOHC TRX21
Engines
Nissan E08 A06 TDS18 DOHC TRX11
Nissan E08 A08 DOHC TRX22
treated as
Nissan E08 A08 DOHC TRXOO
Nissan E08 A08 ATK2 DOHC TRX22
8 cylinders
Nissan E08 A08 DOHC TRX21
Nissan E08 A06 TDS18 DOHC TRX22
EV& PHEV
Nissan E75
Nissan EV75
Nissan_EV200 (ZEV)
Nissan_REEV40 (ZEV)
Subaru
Engines
Subaru E04 A04 DOHC M6
Subaru E04 A04 DOHC M6
treated as
Subaru E04 A04 DOHC TRX11
Subaru E04 A04 DOHC TRX21
4 cylinders
Subaru E04 A04 DOHC TRX21
Subaru E04 A04 DOHC TRX22
Subaru E04 A04 P2 DOHC TRX21
Subaru E04 A04 ATK2 DOHC M6
Subaru E04 A04 ATK2 DOHC TRX22
Subaru E04 A04 P2 TRX21
Engines
Subaru E06 A04 TDS18 DOHC M6
Subaru E06 A04 TDS18 DOHC M6
treated as
Subaru E06 A04 TDS18 DOHC TRX21
Subaru E06 A04 TDS18 DOHC TRX22
6 cylinders
Subaru E06 A06 DOHC TRX21
Engines
None
None
treated as
8 cylinders
EV& PHEV
None
Subaru_EV200 (ZEV)
Subaru_REEV40 (ZEV)
Tesla
EV& PHEV
Tesla EV200
Tesla EV200
Toyota
Engines
Toyota_E04_A04_DOHC_TRX00
Toyota_E04_A04_DOHC_M6
treated as
Toyota_E04_A04_DOHC_M6
Toyota_E04_A04_DOHC_TRX21
4 cylinders
Toyota_E04_A04_DOHC_TRXll
Toyota_E04_A04_DOHC_TRX22
Toyota_E04_A04_DOHC_TRX21
Toyota_E04_A04_ATK2_DOHC_TRX22
Toyota_E04_A04_P2_DOHC_TRX21
Toyota_E04_A04_P2_TRX21
115
-------
Engines
Toyota_E06_A04_TDS18_DOHC_TRXll
T oyota_E06_A04_TDS 18_DO HC_M6
treated as
Toyota_E06_A06_DOHC_TRX00
Toyota_E06_A04_TDS18_DOHC_TRX22
6 cylinders
Toyota_E06_A06_DOHC_M6
T oyota_E06_A06_P2_TRX21
Toyota_E06_A06_DOHC_TRXll
Toyota_E06_A06_DOHC_TRX21
Toyota_E06_A06_P2_DOHC_TRX21
Engines
Toyota_E08_A08_DOHC_TRXll
Toyota_E08_A08_DOHC_TRX22
treated as
Toyota_E08_A08_DOHC_TRX21
Toyota_E08_A08_ATK2_DOHC_TRX22
8 cylinders
Toyota_E08_A08_P2_DOHC_TRX21
Toyota_E08_A08_P2_TRX21
EV& PHEV
Toyota_E04_A04_REEV40_DOHC_TRX21
Toyota_REEV40
Toyota_EV200 (ZEV)
vw
Engines
Volkswagen_E04_A04_SOHC_M6
VW_E04_A04_ATK2_DOHC_TRX22
treated as
Volkswagen_E04_A04_SOHC_TRXll
4 cylinders
Engines
Volkswagen_E06_A04_TDS18_DOHC_M6
VW_E06_A04_TDS18_DOHC_M6
treated as
Volkswagen_E06_A04_TDS18_DOHC_TRXll
VW_E06_A04_TDS18_DOHC_TRX22
6 cylinders
Volkswagen_E06_A04_TDS18_DOHC_TRX21
VW_E06_A04_TDS24_DOHC_M6
Volkswagen_E06_A04_P2_TDS18_DOHC_TRX21
VW_E06_A04_TDS24_DOHC_TRX22
Volkswagen_E06_A04_TDS18_SOHC_TRX21
VW_E06_A04_TDS18_P2_TRX21
Volkswagen_E06_A04_P2_TDS18_SOHC_TRX21
VW_E06_A06_ATK2_DOHC_M6
Volkswagen_E06_A06_DOHC_M6
VW_E06_A06_ATK2_DOHC_TRX22
Volkswagen_E06_A06_DOHC_TRXll
VW_E06_A06_ATK2_TURBM_DOHC_TRX22
Volkswagen_E06_A06_DOHC_TRX21
Engines
Volkswagen_E08_A06_TDS18_DOHC_M6
VW_E08_A08_ATK2_DOHC_TRX22
treated as
Volkswagen_E08_A06_TDS18_DOHC_TRX21
VW_E08_A08_ATK2_TU RBM_DOHC_M6
8 cylinders
Volkswagen_E08_A06_P2_TDS18_DOHC_TRX21
VW_E08_A08_ATK2_TURBM_DOHC_TRX22
Volkswagen_E08_A08_DOHC_M6
VW_E08_A08_P2_TRX22
Volkswagen_E08_A08_DOHC_TRX21
VW_E08_A06_TDS18_P2_TRX21
Volkswagen_E10_A08_TDS18_xOHV_TRX21
Volkswagen_E10_A08_TDS18_DOHC_TRX21
Volkswagen_E10_A10_DOHC_M6
Volkswagen_E10_A10_DOHC_TRX21
Volkswagen_E12_A12_DOHC_TRX21
Volkswagen_E14_A12_TDS18_DOHC_TRX21
Diesels
Volkswagen_E06_A04_DOHC_DSL_M6
VW_E06_A04_DSL_M6
Volkswagen_E06_A04_DOHC_DSL_TRXll
VW_E06_A04_DSL_TRX22
Volkswagen_E08_A06_DOHC_DSL_TRX21
VW_E08_A06_DSL_TRX22
EV& PHEV
Volkswagen_E08_A06_REEV40_TDS18_TRX21
VW_REEV40_TRX21
Volkswagen_E08_A08_REEV40_DOHC_TRX21
VW_E08_A06_TDS18_REEV40_TRX21
Volkswagen_E75
VW_EV75
VW_EV100
VW_EV200
Volvo
Engines
None
None
treated as
4 cylinders
Engines
Volvo_E06_A04_TDS18_DOHC_TRX21
Volvo_E06_A04_TDS18_DOHC_TRX22
treated as
Volvo_E06_A06_DOHC_TRXll
Volvo_E06_A04_TDS24_DOHC_TRX22
6 cylinders
Engines
Volvo_E07_A05_TDS18_DOHC_TRXll
Volvo_E08_A08_DOHC_TRX22
treated as
Volvo_E08_A06_TDS18_DOHC_TRXll
Volvo_E08_A08_ATK2_DOHC_TRX22
8 cylinders
116
-------
EV& PHEV
None
Volvo_EV200 (ZEV)
Volvo_REEV40 (ZEV)
Notes: E=Effective cylinders; A= Actual cylinders; TDS=turbo/downsized; 18/24=bar; DOHC=dual overhead cam;
SOHC=single overhead cam; OHV=overhead valve; TRX denotes the OMEGA transmission mapping, where
TRXOO denotes less than 6 gears (not actually used in OMEGA, used here only for presentation; M=manual
transmission; AKT2=Atkinson-2; TURBM=Miller cycle; P2=hybrid; REEV=plug-in hybrid electric vehicle;
EV=battery electric vehicle; ZEV=zero emission program vehicle and denotes introduction into the fleet for ZEV
program purposes only, where not shown it denotes OMEGA creation in addition to the ZEV program.
Toyota questions the application of electric power steering (EPS) to the LX570/Land Cruiser,
claiming that it is inappropriate for off-road use. EPA has previously observed (for example, in
the Draft TAR at p. 5-200) that EPS has been successfully implemented on all light duty vehicle
classes (including trucks) with a standard 12V electrical system. Due in part to this observation,
EPA has used the EPS code to denote either true EPS or electro-hydraulic power steering
(EHPS) and tracks both as EPS in OMEGA. Importantly, the LPM considers the vehicle type in
making the effectiveness determination for EPS (see the final TSD in support of the 2012 FRM
at section 3.4.3.1) and uses the same cost and effectiveness values for EPS as it would for EHPS.
While EPA disagrees with Toyota about EPS not being a viable technology for their trucks, the
use of EHPS in such vehicles would not change the OMEGA results.
Toyota also questions why the car/truck mix has changed in the Proposed Determination
relative to the Draft TAR. This change is the result of moving to the more recent AEO
projections of future fleet mixes, not due to changes made independently by EPA.
Toyota also highlights differences between the Draft TAR and the Proposed Determination
with respect to treatment of their RAV4 (see Table 3 of the Toyota comments). They express
confusion over why the RAV4's cost in the Draft TAR was $1531 while in the Proposed
Determination it was $1733, when only a transmission change appears to have been made (from
TRX21 to TRX22). In response, there were more aspects to the changes than understood by
Toyota, suggesting some misunderstanding of OMEGA. The costs shown in Toyota's table are
costs of a package when applied to their vehicle. The costs are not necessarily the final costs for
that vehicle. That said, there were minor differences in the way the RAV4 was characterized in
the Draft TAR baseline and in the Proposed Determination baseline. In the Draft TAR, the
baseline RAV4 was characterized as having 3.5 percent mass reduction and EPS. In the Proposed
Determination the baseline RAV4 had no EPS (erroneously in this case as discussed in 2.7.1 of
this document) and had no mass reduction. As a result, the baseline RAV4 had $243 (2014$) in
technology. In the Proposed Determination, the RAV4 had $290 in technology (2015$). At 3.5
percent mass reduction, the costs for the technology are negative, hence the higher cost in the
Proposed Determination, since a cost save associated with mass reduction is not included (this is
partly offset by the presence of EPS in the Draft TAR while not in the Proposed Determination).
For the package being added, the Draft TAR package had a cost of $1774 (2014$) while the
Proposed Determination had a cost of $2023 (2015$) with that cost differential being due to the
transmission difference along with the dollar-basis difference. The final increase in costs
associated with adding these packages is $1531 (2014$) in the Draft TAR ($1744 minus $243)
and $1733 (2015$) in the Proposed Determination ($2023 minus $290).
Toyota also expresses concerns regarding differences in results for their Sequoia and Land
Cruiser, where the Proposed Determination showed higher costs and lower final CO2 for those
117
-------
vehicles. Toyota does not express what their specific concerns are over the results. In the
Proposed Determination, those vehicles are part of the same OMEGA-platform which is placed
on the unibody mass reduction cost curve, making mass reduction more cost effective relative to
the Draft TAR. As a result, OMEGA pushes the platform further on mass reduction as well as
adding MHEV48V technology to some vehicles. These technologies became more cost effective
relative to other possibilities for the platform allowing for more control on the platform in favor
of reducing costs elsewhere. This outcome is reasonable especially in light of our interest in
showing a compliance pathway rather than a prediction for each individual vehicle.
The Global Automakers also questioned why the technology penetration of the Atkinson 2
technology went down in the Proposed Determination. In response, it seems reasonable for it to
have gone down given that a lower effectiveness value was applied to this technology (see
Chapter 2.5.1 of this document) while maintaining equivalent costs relative to the Draft TAR. In
combination with other changes, notably the movement of more vehicles to the unibody mass
reduction cost curve, the result is reasonable.
With respect to comments regarding EPA's expectation that the MY2021 reference fleet
would comply with MY2021 standards, both the Alliance and Toyota commented that
compliance with the MY2021 standards were not a "given." Both commenters point to a recent
study by their contractor, Novation Analytics, which projects that the MY2016 fleet will not
comply with the MY2016 standards. As the final MY2016 sales data is not yet available,
Novation Analytics was forced to rely on a number of alternative information sources including
sales projection data from vehicle manufacturers and vehicle registration data from IHS-Markit,
as well as their own projection of the application of credits. Without explicitly saying so, the
commenters seem to be extrapolating this estimated MY2016 compliance condition onto
MY2021, with the conclusion that the fleet will no longer be able to comply with future model
year standards and that MY2016 is the beginning of the end of fleet compliance. EPA disagrees
with this conclusion for several reasons. The estimated compliance situation provided by
Novation Analytics is not based on final MY2016 sales data and credit usage. Depending on
when the sales estimates were made, final data that has undergone the entire quality control
process required for establishing fleet compliance can be significantly different. In addition, a
single model year compliance situation does not reflect the long term compliance strategy for a
single manufacturer or for the fleet. All manufacturers are managing their compliance with
respect to their product plans and available credits. Individual manufacturer decisions to use
banked credits in lieu of making product changes to meet a single model year's standards may
reflect a legitimate short-term compliance strategy (for example, to use credits to better align
with redesign schedules) but not be an indication of future model year compliance capability. In
fact, as pointed out in Novation's MY2016 analysis, many vehicle manufacturers have over
complied in the initial model years of the GHG program so it is reasonable to expect some level
of under compliance in future model years. This is an example of one of the flexibilities inherent
in the credit averaging, banking and trading provisions in the 2012 FRM.
The Alliance also commented that the phase-out of Flexible Fuel Vehicle (FFV) credits and
the shift in sales from cars to trucks is negatively impacting the vehicle manufacturers' ability to
comply with current and future standards. The Alliance further concludes that the difference
between the projected CO2 performance from the 2010 FRM vs. the actual CO2 performance of
the 2012 to 2016 MY fleet is the result of over-optimistic technology deployment rates and
effectiveness. EPA does not agree with this conclusion or that the differences between the
118
-------
projected and actual performance are the result of technology penetrations and effectiveness.
The projections provided by EPA in the 2010 FRM were based on a car/truck sales mix that has
changed significantly since the FRM. Comparing the FRM projections to the actual fleet
performance is inappropriate. This is best exemplified by the fleet target shown on slide 31 of
the Novation Analytics study. On this slide, the fleet target established by Novation Analytics is
260 g/mile as compared to the projected target of 250 g/mile from the 2010 FRM. This change
in target is mainly due to the shift in sales from passenger cars to trucks. A key point here is that
the performance of today's fleet relative to a six-year old projection is irrelevant—the fleet
changes and the targets change with it due to the footprint-based nature of the standards. The
important point is that the fleet continues to improve along a trajectory consistent with the
original intent of the national program - that being to cut in half the greenhouse gas emissions
rate by 2025.
As for GM's comments regarding transmission technology penetration rates, this can be
explained by the updated technology effectiveness values as discussed in Chapter 2.5.4 of this
RTC document. The result of those changes was to move more vehicles to the TRX22
technology and away from the TRX21 technology. GM rightly points out that the TRX22
sensitivity shows higher costs than the central case illustrating the importance of this technology
although not a reliance on it for feasibility. We also discuss comments relating to perceived
volatility of results in Chapter 2.1 of this RTC document.
119
-------
Chapter 3: Economic / Consumer / Other Factors
3.1 Consumer Response
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Section B.1.2 of the Appendix to the Proposed Determination discussed the conundrum that
private markets appear to have been slow to provide fuel-saving technologies with short payback
periods, a phenomenon referred to as the energy paradox or the efficiency gap. The Appendix
section includes an overview of the comments EPA received on this topic in the Draft TAR and
EPA's responses to these comments.
Sections B. 1.1 and B. 1.3 of the Appendix to the Proposed Determination and Chapter 4.2 of
the TSD discuss consumer response to vehicles subject to the standards, with a focus on vehicle
sales. Sections B.l.l and B.1.3 of the Appendix and Chapter 4.2.1 of the TSD include an
overview of the comments EPA received on this topic in the Draft TAR and EPA's responses to
these comments.
Some commenters argued that a number of market and behavioral failures have contributed to
the existence of the energy efficiency gap. As EPA discussed in Appendix B.1.2 of the Proposed
Determination and Chapter 6.3 of the Draft TAR, there are a number of hypotheses to explain the
existence of the gap; there is relatively little research to test the hypotheses. The finding of a gap
is demonstrated in the technical analysis of cost and effectiveness of the standards.
Many of the comments focused on the role of fuel economy in consumers' vehicle purchases,
and what payback period for fuel-saving technologies consumers will consider. (What discount
rate people use to evaluate fuel savings is another way of considering the same issue, how people
trade off up-front costs and future savings over time). This value affects how people will respond
to standards that increase up-front costs while providing fuel savings over the vehicles' lifetimes.
Some commenters suggested that vehicle buyers take into account fuel savings over less than the
lifetimes of their vehicles, perhaps as little as 2-3 years of fuel savings, when they make their
purchase decisions. Consumers considering less than the lifetime of fuel savings in their
purchases is often considered evidence of the efficiency gap. EPA in Appendix B.1.2 of the
Proposed Determination and Chapter 6.3 of the Draft TAR reviewed the evidence on the role of
fuel savings in consumer purchase decisions. There, EPA agreed with the National Academy of
Sciences that there is not a definitive answer at present.40 Some studies show that consumers
undervalue fuel savings; others find no or little undervaluation; some find overvaluation.
The payback period that consumers use plays a key role in examining the effects of the
standards on vehicle sales. Some comments requested that EPA conduct a quantitative
assessment of the effects of the standards on vehicle sales. As EPA discussed in Section B. 1.3.3
of the Appendix to the Proposed Determination, evidence suggests that models of consumer
vehicle demand have not yet demonstrated their suitability for policy analysis, in part due to the
large uncertainty over the role of fuel economy in consumers' vehicle purchase decisions.
Nevertheless, as the Draft TAR presented in Chapter 6.2, the standards for MYs 2012-16 have
40 National Research Council (2015). Cost. Effectiveness and Deployment of Fuel Economy Technologies for Light-
Duty Vehicles. Washington, D.C.: The National Academies Press, pp. 9-16, 9-36. Docket EPA-HQ-OAR-2015-
0827-0273.
120
-------
not prevented vehicle sales from achieving record levels. The standards appear to play a far
smaller role in vehicle sales than broader macroeconomic forces.
For future years, some comments expressed concern that the costs associated with the
standards would be higher than EPA estimated in the Draft TAR, and that these higher costs
would reduce vehicle sales. EPA does not agree with these cost estimates. As discussed in
Chapter 4.2.1 of the TSD, some of the cost estimates from these commenters are not based on
costs of the technologies expected to be used to meet the standards, but rather on outdated
estimates of changing sales mix with fixed technologies, and therefore are not relevant to the
standards. Other estimates are based on divergences from price patterns in a few selected
overseas markets; as discussed in Appendix B. 1.6.2 to the Proposed Determination, this
approach does not provide a sound basis for estimating the effects of the standards on vehicle
prices.
Some comments suggested that costs would be higher because more electrification of the fleet
(both hybrid-electric vehicles and plug-in vehicles) is required than EPA has estimated. As
discussed in Appendix C. 1.1.3 to the Proposed Determination, EPA finds that the standards can
be met almost fully with advanced gasoline technologies. For a small segment of the public,
PEVs already are suitable for their purposes. As the technology of PEVs evolves, especially as
range and fueling infrastructure expand, it is likely that a larger segment could find PEVs
suitable.
Summary of Comments on the Proposed Determination
The comments on the Proposed Determination repeat many of these points.
Impact on sales
The Blue-Green Alliance, Center for Biological Diversity (CBD), Consumer Federation of
America (CFA), Consumers Union (CU) Environmental Defense Fund (EDF), International
Council for Clean Technology (ICCT), Michigan League of Conservation Voters, Natural
Resources Defense Council (NRDC), Union of Concerned Scientists (UCS), Utah Physicians for
a Healthy Environment, and the joint comments of 74 environmental groups argue that the
standards provide significant net benefits to consumers. Some point to polls indicating the
popularity of the standards across diverse populations. Various of them also point out that sales
are high, the auto industry is profiting while the industry is over-complying with the standards,
options for fuel-efficient vehicles have increased, and exports have increased.
Automakers express concerns about the effects of the standards in the MY2022-25 period,
generally because they expect costs to be higher than EPA estimates, due to greater
electrification than EPA considers necessary to meet the standards. Fiat Chrysler argues that its
sales elasticity shows potential reductions in sales in the MY2022-2025 period. The Alliance
argues that auto sales are inelastic, citing a report from the Center for Automotive Research
(CAR) (MacAlinden et al. 2016). The Alliance, Global Automakers, National Automobile
Dealers Association (NADA), and Fiat Chrysler argue that EPA should have done a quantitative
vehicle sales analysis; CBD disagrees with the need for quantification. Fiat Chrysler suggests
that EPA look at sales of "high efficiency" vehicles relative to other segments of the market;
Global Automakers considers sales of electrified vehicles to be proxies for consumer acceptance
of efficiency technologies. Ford and Toyota request that EPA include macroeconomic
121
-------
forecasting to develop its baseline and sales projections. CBD, NRDC, and UCS raise concerns
about models of consumer preferences for vehicles. CBD and NRDC mention that the models
are typically based on existing vehicles, not future vehicles with innovations; models based on
existing vehicles will not produce good results. UCS points out that the lack of consensus on
consumer willingness to pay for vehicle attributes arising from these models suggests that the use
of these models for quantitative analysis is not appropriate. CBD argues that "consumer
preference is not one of the factors set forth in the pertinent regulation that EPA must consider."
The Alliance, NAD A, Fiat Chrysler, and Ford continue to cite MacAlinden et al. (2016). The
Alliance says that EPA has not refuted it; NADA argues that, although it is not perfect, it is
superior to EPA's analysis because it does provide quantitative estimates. The International
Center for Clean Transportation (ICCT), and UCS support EPA's critique of the CAR Report.
CAR defends its study (MacAlinden et al. 2016), arguing that:
• It did not base its costs on a 1991 study, but rather its cost assumptions of $2000 -
$6000/vehicle "were used in place of a single, derived fuel efficiency scenario." It
argues that its low value, $2,000, includes a retail price equivalent multiplier of 1.86;
without that multiplier, the cost, $1075, is similar to EPA's estimated technology cost.
• It did not use discounting studies in its assessment of consumer payback periods
because of the assumptions needed to convert them to payback periods.
• Its use of two very similar models that produce different results is due to different
purposes for the models (one for revenues, one for consumer expenditures). They
appear to contradict each other because one is in nominal dollars, while the other is in
real dollars.
• If increased prices cause expenditures to decrease, then sales decrease, and
employment decreases. The decrease in sales is a signal that consumers do not want
the added content associated with the price increase, and thus employment decreases
due to sales decreases account for "substitution effect" employment. In addition, new
technologies substitute for existing technologies, rather than adding content.
The Alliance and Fiat Chrysler consider speculative EPA's statement that the impacts of the
standards on vehicle sales are likely to be secondary to broader economy-wide impacts. This
claim is based on the lack of quantification of sales and employment impacts.
Mercedes-Benz expresses concerns about the impacts to leasing and residual values due to the
program. In particular, with most lease terms for 36 months, it is concerned that "the payback
period of new technologies required will not be short enough to be attractive to these
consumers."
Role of Fuel Economy in Consumer Vehicle Purchase Decisions and the Energy Paradox
Fiat Chrysler, the Institute for Energy Research (IER), and an assessment by the Defour
Group provided by the Alliance (as an attachment to their comments) express skepticism of the
existence of the energy paradox, noting that consumers may not undervalue fuel economy.
Instead, according to IER and Toyota, consumer heterogeneity might explain what appears to be
122
-------
an efficiency gap, citing a paper by Bento et al.41 Defour states that people might rationally wish
to spend on attributes other than fuel-saving technologies, and that the standards require
foregoing other vehicle attributes; indeed, it says EPA does not provide good evidence of
consumer irrationality, and fuel economy labels should be sufficient to overcome that
irrationality. It argues that OMB Circular A-4 requires demonstration of consumer irrationality.
On the other hand, CFA points to a large number of market imperfections. NRDC points out
that manufacturer risk aversion may contribute to undersupply of fuel-saving technologies. CU
asserts that automakers seem resistant to adding technologies that save consumers money; it
argues that consumers prefer to spend more on vehicles than on fuel, and that "it is a disservice"
not to provide consumers with cost-effective fuel-saving technology. UCS observes that
manufacturers will under-supply fuel efficiency technologies if they assume that consumers
require a 2-3 year payback period for costs when consumers may be willing to accept longer
payback periods.
Global Automakers continues to argue that consumers consider only 2-3 years of fuel savings
in their vehicle purchase decisions; the Center for Automotive Research (CAR) points out that
EPA mentions this finding. In contrast, CBD, citing CU, supports a 5-year payback, although
Toyota questions this finding. NADA seeks for EPA to conduct a meta-analysis of the
willingness-to-pay for fuel economy, using Strategic Vision's results on how consumers rank
fuel economy in their purchase decisions, and considers EPA's conclusions to be "seriously
flawed" without using those data. Global Automakers and Toyota argue that EPA should use
higher rates than 3 and 7 percent to discount fuel savings; Toyota recommends credit-card
interest rates. UCS, on the other hand, says that estimates of the willingness to pay for fuel
economy are highly varied, and admonishes against cherry-picking results. It notes that Toyota's
credit card interest rate corresponds to a payback period of 5-6 years, greater than the assertion
of 2-3 years.
Global Automakers argues that EPA did not rely on real-world data, but rather relied on
academic studies that do "not provide definitive answers" in assessing consumer demand for fuel
efficiency and electric-drive vehicles. NADA would like EPA to assist in influencing consumers
to buy more efficient vehicles. Fiat Chrysler recommends that EPA compare sales of "high
efficiency vehicles" relative to those in other segments.
Global Automakers disputes an EPA interpretation of a NADA finding discussed in the
Proposed Determination. NADA found that 68 percent of consumers would pay $30 or less per
month for a 17-mpg increase in fuel economy. EPA pointed out that, with a 5-year payback
period at a 5 percent interest rate, $30/month is equivalent to $1,558 in up-front costs, which is
less than the cost of the MY2022-2025 standards. Global Automakers points out that the fuel
economy gain over the MY2022-2025 program is 5.1 mpg, not 17 mpg, as in NADA's scenario.
Role of Lower Gasoline Prices
The Alliance, Global Automakers, and GM say that EPA does not adequately consider how
low fuel prices reduce the importance of fuel economy in consumer vehicle purchases, and
41 Bento, Antonio, Shanjun Li, and Kevin Roth (2012). "Is there an energy paradox in fuel economy? A note on the
role of consumer heterogeneity and sorting bias." Economics Letters 115: 44-48. An early version of this paper is
docketed at EPA-HQ-OAR-2010-0799-11940.
123
-------
reduce consumer acceptance of efficient vehicles. Ford and Toyota point to diminishing returns
to increasing fuel economy. Ford argues that it needs to add other attributes to get people to buy
more efficient vehicles. In addition, Ford argues that increased costs will shift the market to
trucks, because the fuel economy gains are greater in that segment.
NADA points out that efficiency gains cannot be achieved if vehicles aren't sold, and argues
that EPA should aim to maximize fleet turnover.
Payback Period Analysis
Fiat Chrysler argues that EPA's payback calculation used the prime interest rate; this interest
rate, available to borrowers with strong credit ratings, is lower than that for subprime borrowers.
Toyota says that the payback calculation "conflates the intended concern of vehicle affordability
raised in the comments with a myopic view of payback."
Response to Comments on the Proposed Determination
Impact on Sales
As EPA discussed in the Sections B. 1.1 and B. 1.3 of the Appendix to the Proposed
Determination and Chapter 4.2 of the TSD, and in Section I. A. of the Final Determination, the
standards to date have not stopped the auto market from achieving record sales levels. We note
here the argument from the Alliance and Fiat Chrysler, that EPA is speculating when it claims
that the standards play a role secondary to broader macroeconomic conditions in determining
sales. In the context of the last 5 years, this claim from the Alliance and Fiat Chrysler, that the
standards play a significant role in sales and in broader macroeconomic conditions, suggests that
the standards may have contributed significantly to the sales boom and economic growth. EPA
does not have evidence for such a finding. We recognize that the Alliance and Fiat Chrysler
instead seek to imply that, in the future, the standards will play a more significant role in sales.
This argument is based on the premise that costs to meet the standards will be much higher than
EPA has estimated. We continue to support our cost estimates, and therefore do not expect the
effect on sales to be very large. The effects of the standards on vehicle prices are discussed
further in Section 3.3, below.
The arguments for quantifying vehicle sales impacts seem to claim that even flawed
quantification is better than no quantification. EPA does not agree that quantification is
inherently superior to qualitative results; the discussion in Section 3.4, below, relating to
quantification of employment impacts, applies here as well. As discussed in the Section B.1.2 of
the Appendix to the Proposed Determination (and previously in the 2012 FRM, 77 Federal
Register 62946-62950, and RIA Chapter 8.1), a key parameter in determining the effects of the
standards on vehicle sales is the role of fuel economy in consumer vehicle purchases. As
discussed there and below, and as the comments suggest, the research that has been done on the
appropriate value has not reached anything approaching consensus. Without that key parameter,
estimates of the effects of the standards on vehicle sales range from reductions, as CAR and
others cite, to increases. EPA considers examination of consumer preference to provide useful
information, even if it is not quantified.
The elasticity of vehicle demand (sales) with respect to price is another key parameter; it
measures how sales respond to a change in price. In response to Fiat Chrysler, we note that the
elasticity does not by itself determine whether sales will increase or decrease; the role of fuel
124
-------
economy in vehicle purchases—the share of future fuel savings that consumers take into account
when buying vehicles—is needed for that directional indicator. If that share of fuel savings
exceeds technology costs, then sales would be expected to increase. The elasticity determines the
magnitude of the sales impact. We observe that the inelastic response—a smaller percent change
in sales than percent change in price—argued by the Alliance suggests smaller sales impacts than
an elastic response.
It is not obvious how to implement Fiat Chrysler's suggestion of comparing sales of "high
efficiency" vehicles to sales of vehicles in other segments. The standards are intended to increase
the efficiency of vehicles across the fleet, not just in limited segments. If Fiat Chrysler means
that EPA should compare electrified vehicles to other vehicles (and as Global Automakers
specifically indicates), EPA disagrees that this approach will produce valid insights. First, EPA
projects that the standards can be met with very low levels of electrification. Second, these
vehicles are often not directly comparable in other attributes to conventional vehicles.
Regarding the request for the use of macroeconomic forecasting for the baseline and sales
projects, EPA notes that the reference fleet is based on macroeconomic projections from the EIA
and IHS; the baseline is the actual data from MY2015 (see TSD Chapter 1). We agree with CBD
that models of consumer demand for vehicles are typically based on existing or historic fleets
and may not reflect consumer preferences for future vehicles.
Mercedes-Benz's concern about payback period for leased vehicles is based on an expectation
that the first owner/leaser must pay for the full value of the technology in those three years. After
the lease period ends, the vehicles are expected to be resold. The studies on the role of fuel
economy in vehicle purchases, discussed in Appendix Section B.1.2 of the Proposed
Determination, find that fuel economy continues to play a role in purchases in the used vehicle
market. Thus, the full payback for new technologies need not be the responsibility of the first
vehicle owner/leaser.
EPA continues to find the CAR Report not applicable to the standards, for the reasons stated
in TSD Chapter 4.2.1 and Section B.1.3 of the Appendix to the Proposed Determination. In
response to the points raised by CAR in its comments here, we respond:
• We continue to disagree with the cost estimates from CAR; regardless of their source,
they are not based on an analysis of the costs of technologies sufficient to meet the
standards. We also disagree that the cost estimate without the retail price equivalent
multiplier, $1075, is similar to EPA's primary estimate of $875 (see Table ES-1 of the
Proposed Determination), because EPA's estimate already includes use of indirect
cost multipliers, which account for indirect costs related to production (which can be
reflected in the retail price); thus, CAR's low estimate of $2000 is more than twice
EPA's primary estimate, and reflects a significant element of double counting costs.
• In deciding which studies to consider in assessing the role of fuel economy in
consumer purchase decisions, CAR decided not to use studies that estimate
consumers' implicit discount rates. Discount rates are another way of estimating the
role of fuel economy in consumer decisions; rather than considering fuel savings for
only a few years (the payback period approach), the discount rate approach considers
the lifetime of fuel savings, and estimates the weight (the discount factor) that
consumers put on those future fuel savings. It is true that assumptions are needed to
125
-------
convert one approach into the other. Nevertheless, these studies (CAR itself cites 7),
typically based on consumers' actual behavior, provide useful insights into consumer
decision-making.42 Studies of payback typically use stated-preference data, not
revealed behavior.
• CAR's explanation of getting different results from similar models is that one is in
nominal dollars, and the other is in real (constant) dollars, where the difference is
inflation. However, inflation is just a multiplicative constant in these regressions; it
should not by itself affect the relationship between expenditures (or revenues) and
price. EPA still considers these models not to be of sufficient quality to estimate
elasticities or to forecast revenues or expenditures.
• CAR's argument for excluding the substitution effect does not reflect the purpose of
this effect. The substitution effect is that part of the increased cost of the standards
that is due to labor inputs to production. (Indeed, the vehicle manufacturer comments
emphasized the importance of the substitution effect. See e.g. the Alliance comments
at p. 23). CAR's explanation instead is about the "output effect," which is the effect
of changes in sales on employment, that EPA does not quantify. CAR further argues
that the new technologies primarily substitute for other technologies. If that were
strictly true, the standards would not impose additional costs other than materials.
EPA finds that the fuel-saving technologies have added costs, and part of those
increased costs is labor. This increase in labor is the substitution-effect employment;
see Section B.2 of the Appendix to the Proposed Determination, where we estimated
substitution-effect employment impacts to be between 1000 and 12,000 job-years in
MY2025.
Role of Fuel Economy in Consumer Vehicle Purchase Decisions and the Energy Paradox
EPA continues to agree with the National Academy of Sciences that the role of fuel economy
in consumer vehicle purchase decisions is unresolved. As EPA has previously pointed out (see
Section B.1.2 of the Appendix to the Proposed Determination), the 2-3 year payback cited by
Global Automakers is the low end of the possible scale; at the other end, as argued by Defour
and IER, consumers may not undervalue fuel savings, and instead are acting rationally. A
payback period of 2-3 years suggests substantial undervaluation and is thus potential evidence of
the existence of the energy paradox. The use of alternative discount rates for fuel savings, as
discussed above, is a proxy for different payback periods; higher discount rates than opportunity
costs tend to be considered evidence of consumer myopia, and thus of the energy paradox. EPA
continues to disagree with Toyota that credit-card rates are the opportunity cost for borrowing
when auto loan rates are the relevant opportunity cost metric; vehicles are rarely purchased via
credit cards. In comments on the Draft TAR, for instance, Fiat Chrysler cited evidence that
almost 86 percent of vehicles are financed, which we assume does not include credit cards. We
also note that several commenters endorse the use of survey data from Strategic Vision on the
importance of fuel economy in consumer vehicle purchases; other commenters point to survey
data, including that from Consumers Union (CU), showing widespread consumer support for
42 CAR's list, in addition, is outdated and incomplete. For instance, it cites working paper versions of Allcott and
Wozny from 2010, and Sallee et al. from 2011; they are both now published and discussed in the Draft TAR and
the Proposed Determination (Docket EPA-HQ-OAR-2015-0827-0107 and EPA-HQ-OAR-2015-0827-0120); and
it omits Busse et al. 2013 (Docket EPA-HQ-OAR-2015-0827-0110).
126
-------
strong fuel economy standards; and some argue against the use of showing general public
attitudes toward fuel economy, as not reflecting actual consumer behavior. We consider all these
sources to inform, but not be decisive for, the role of fuel economy in consumer vehicle
purchases.
As noted, Defour and IER argue that consumers are rationally choosing to spend on vehicle
attributes other than fuel economy. Fuel economy is unusual in being the only attribute of a
vehicle that has the potential to pay for itself in cash savings. In addition, as discussed in TSD
Chapter 4.1 and Section B. 1.4 of the Appendix to the Proposed Determination (and in Section
3.2 below), fuel economy and other vehicle attributes are not mutually exclusive, so there is no
necessary tradeoff between fuel economy and other vehicle attributes. Purchasing both comes at
a cost; as noted, though, the fuel-saving technologies can pay for themselves. Section B. 1.6.3 of
the Appendix to the Proposed Determination and TSD Chapter 4.3.3 further discuss EPA's
assessment of the effects of the standards on access to the credit market. As discussed there, it is
possible that there may be an effect, but the limits on access to credit do not appear absolute.
Defour argues that EPA does not provide good evidence of consumer irrationality, and claims
that such evidence is required by OMB Circular A-4. EPA notes, first, that it is not issuing the
standards to "correct" for consumer irrationality; rather, it is issuing the standards to reduce
vehicle GHG emissions. Second, EPA's assessment of the energy paradox is based in its
technology assessment: we find that fuel savings outweigh technology costs, and, as discussed in
Chapter 4.2 of the TSD and Appendix Section B.1.5 of the Proposed Determination, we have not
found evidence of "hidden costs" of the technologies. Thus, the finding of the gap does not rely
on findings about consumer behavior. Third, EPA discussed extensively (see Proposed
Determination Appendix Section B.1.2) the possible reasons for the efficiency gap, including the
possibility of consumer heterogeneity (see Proposed Determination Appendix p. A-29: "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"). Because EPA was summarizing previous discussions, it
did not cite all the pertinent literature, such as Bento et al., but it has previously considered that
and many other papers. Finally, as discussed in that section, EPA also notes the possibility of
producer-side factors as sources of the efficiency gap, such as strategic decisions over which
attributes to offer in which vehicle segments. NRDC, CFA, and CU point to potential producer-
side explanations. The commenters are correct that EPA has not made definitive findings on the
source of the efficiency gap; nevertheless, the technology assessment is the basis for the
observation that there continue to be technologies for which the expected present value of fuel
savings exceeds estimated technology costs. Finally, OMB Circular A-4 does not require a
finding of consumer irrationality; instead, it provides guidance and asks agencies to "examine
and discuss why market forces would not accomplish these gains in the absence of regulation."43
As noted, EPA has discussed extensively why these gains might not happen in the absence of the
standards.
Global Automaker's concern that EPA relied on academic studies rather than "real-world
data" is puzzling, because academic studies are commonly based on real-world data. It seems to
be requesting that EPA instead look at slow sales of electrified vehicles as evidence for
43 Office of Management and Budget (2003). "Circular A-4." https://www.whitehouse.gov/omb/circulars_a004_a-4/,
Docket EP A-HQ-0 AR-2015-0827-0803.
127
-------
consumers not being interested in fuel-saving technologies. EPA's analysis suggests that only
low levels of electrification are necessary to meet the standards. EPA further discussed consumer
acceptance issues related to electrified vehicles in Appendix Section B. 1.5.2 of the Proposed
Determination and Section 3.2 below.
Regarding the NAD A finding that most consumers would pay no more than $30/month for a
17 mpg increase in fuel economy, Global Automakers is correct to point out that the 17 mpg
increase is not the projection for the MY2022-25 standards, but is instead approximately the
increase expected from MY2016 to MY2025. In Table 12.44 of the Draft Technical Assessment
Report, EPA estimates that the cost of going from MY2016 standards to MY2025 standards is
$1,287. Five years of $30/month at a 5 percent discount rate is $1,558, which exceeds that value.
EPA's primary point here is that NADA's estimate of $30/month, even if only for part of the
population, suggests a significant willingness to pay for additional fuel-saving technology.
Role of Lower Gasoline Prices
EPA does not agree that we have not adequately considered the role of lower fuel prices.
EPA's analysis was done using AEO estimates that reflected a lower fuel price scenario, and
included sensitivity analyses for prices below the expected (reference) values. In addition, the
fuel savings calculations are done by estimating volumes of fuel saved, not miles per gallon. We
agree that there are diminishing returns to increasing mpg, but fuel savings increase with gallons
saved. Though Toyota considers diminishing returns to be especially an issue for low-priced
vehicles, calculating fuel savings rather than mpg increases avoids this concern for any segment.
We agree with NADA that efficiency gains will increase as more new vehicles penetrate the
market. As discussed above, we have not made a prediction on the effects of the standards on
vehicle sales. As discussed above, though, we do not consider the effects of the standards on
sales to be as negative as the CAR Report implies.
Payback Period Analysis
EPA's calculation of the returns for those borrowing on credit reflect the average rate, 4.25
percent, at the time of the calculation (see Proposed Determination Appendix Section C.2.4). We
recognize that some will get lower rates, and some will get better rates. The payback period
analysis nevertheless indicates that fuel savings will recover technology costs in approximately
5-6 years (see Proposed Determination Appendix Section C.2.4). We do not understand the
intent behind Toyota's statement of the conflation of affordability and "a myopic view of
payback." Toyota does not provide additional explanation of this comment.
3.2 Consumer Impacts of New Technologies
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Section III. A. of the Proposed Determination, Sections B.1.4 and B.1.5 of the Appendix to the
Proposed Determination, and Chapters 4.1 and 4.2.2 of the TSD discussed EPA's assessment of
the effects of greenhouse gas (GHG)-reducing technologies on other vehicle attributes. The
Appendix sections include an overview of the comments EPA received on this topic in the Draft
TAR and EPA's responses to these comments.
Some comments suggested that other vehicle attributes, especially vehicle power or size,
might have increased in the absence of the standards, and requested that EPA develop a reference
128
-------
fleet based on larger and more powerful vehicles. Other comments suggested that EPA base its
reference fleet on MY2010 vehicles, because any improvements in other vehicle attributes are
due to automakers' choices not to use technological progress for reduced GHG emissions. Yet
other comments pointed out that some of the GHG-reducing technologies provide ancillary
benefits to other vehicle attributes, such as lightweighting providing better handling and
performance. As discussed in the Proposed Determination Appendix Section B.1.4, EPA has
updated the reference fleet to MY2015, to use the most recent final data, and maintains other
vehicle attributes at their MY2015 levels. Because, as many comments pointed out, the standards
appear to be contributing to major innovations, the standards may provide significant ancillary
benefits as well as potential opportunity costs (although, as noted in Chapter 2.2.5 of this
Response to Comments, the EPA analysis holds performance constant; i.e., the analysis holds
acceleration constant (that is, includes costs to preserve acceleration) as a proxy for standards
which preserve all performance attributes). In addition, as some comments indicated, it is not
clear that consumer demand for additional power is large; and some innovations, such as
improved infotainment systems, are not directly related to GHG emissions. In light of these
issues, EPA maintained the static baseline in its modeling.
Some comments asked questions about EPA's analysis of professional auto reviews as a
source of insight into consumer response to fuel-saving technologies. As discussed in Section
B. 1.5.1.2 of the Appendix to the Proposed Determination, EPA considers these analyses to
provide useful insights into potential "hidden costs" of the new technologies. As documented in
that section and in TSD Chapter 4.2.2, the studies found more positive associations with the
existence of the technologies than negative associations. The evidence suggests that it is possible
to implement these technologies well, and that automakers may improve their implementation
over time.
Other comments expressed concern that vehicle buyers will not accept the higher costs
associated with high levels of vehicle electrification. As discussed in the Proposed Determination
Section III.A and the Appendix B. 1.1, with the very low proportion of PEVs projected to be
needed for compliance, EPA expects that compliance will mostly depend on advanced gasoline
technology vehicles. Moreover, as discussed in Chapters 2.2.4.4.5 and 2.2.4.4.6 of the TSD, the
market for electrified vehicles is evolving rapidly. As discussed in Section B.l.5.2 of the
Appendix to the Proposed Determination, widespread consumer acceptance of PEVs may
depend, not only on technological advances, but also on the feedback loop associated with other
consumers purchasing PEVs.
Summary of Comments on the Proposed Determination
Potential Opportunity Costs of the Standards
The Defour Group (in an attachment to comments from the Alliance), Ford, Global
Automakers, the Institute for Energy Research (IER), and Toyota express concerns that EPA
ignores opportunity costs, because they claim the standards deny people the opportunity to buy
vehicles with other attributes that they prefer to fuel savings, or because other attributes might be
adversely affected. Global Automakers argues that the analysis of auto reviews conducted by
EPA does not reflect consumer responses, nor does it reflect all technologies that will be used to
meet the standards. The Natural Resources Defense Council, in contrast, considers EPA's
analyses to be a fair assessment of the new technologies. IER points specifically to two studies
(Whitefoot et al. (2011) and Klier and Linn (2016)) suggesting that consumers will be adversely
129
-------
affected.44 The Union of Concerned Scientists (UCS) points out that Whitefoot et al. note
consumer surplus gains, and that the Klier and Linn paper finds that fuel savings exceed the
opportunity costs of performance. An anonymous commenter expresses concern that the
standards are forcing people into larger vehicles and eliminating manual transmissions.
The Blue-Green Alliance (BGA), Business for Innovative Climate and Energy Policy
(BICEP), Center for Biological Diversity (CBD), Consumer Federation of America, the
Michigan League of Conservation Voters, and a joint letter from 74 environmental groups, in
contrast, point to the tremendous innovations stimulated by the standards that drive down costs.
BICEP, the Environmental Defense Fund (EDF), International Council for Clean Technology
(ICCT), UCS, and Woodward argue that this innovation increases the competitiveness of the
U.S. auto industry in the world economy. BGA, BICEP, ICCT, and UAW point out that the
standards provide certainty necessary for the large investments being used to meet the standards;
indeed, ICCT points to the uncertainty, and hence reduced investment, that would arise if the
Determination is not finalized. In addition, BGA observes that research on electric vehicles for
components such as start-stop and regenerative braking complements technology for
conventional vehicles.
Consumer Acceptance of Electrified Vehicles
Ford, Global Automakers, Mercedes-Benz, Nissan, Subaru, and Toyota express concerns
about consumer acceptance of electric-drive vehicles, and the potential need for cross-
subsidization or government incentives to achieve the standards, based on its disagreement with
EPA over the level of electrification needed to achieve the standards. Global Automakers
observes, e.g., that 75 percent of HEV or PEV owners who replaced those vehicles in 2016 chose
gasoline vehicles. The High Octane-Low Carbon Fuel Alliance argues that people prefer liquid
fuel-powered cars. On the other hand, the California Air Resources Board (CARB), Faraday
Future, the Center for Biological Diversity (CBD), and UCS argue that consumer acceptance for
electrified vehicles is growing. UCS points to the advances in electric vehicle technology, such
as longer range and lower costs. CARB points out that factors beyond fuel prices, including
awareness of electrified vehicles, interest in driving, and using the latest technology, influence
purchase of electrified vehicles; it points to survey data showing owner satisfaction with PEVs
and intent to purchase another in the future. Toyota argues that more efficient conventional
vehicles do not in fact have fuel economy comparable to hybrids, noting that the most efficient
mid-sized gasoline car gets 36 mpg (combined city and highway), while the hybrid Prius gets 56
mpg combined.
44 Whitefoot, Kate, Meredith Fowlie, and Steven Skerlos (2011). "Product Design Responses to Industrial Policy:
Evaluating Fuel Economy Standards Using an Engineering Model of Endogenous Product Design." Energy
Institute at Haas Working Paper No. 214, Docket EPA-HQ-OAR-2010-0799-11895; Klier, Thomas, and Joshua
Linn (2016). "(2016). "The Effect of Vehicle Fuel Economy Standards on Technology Adoption." Journal of
Public Economics 133: 41-63, Docket EPA-HQ-OAR-2015-0827-0142.
130
-------
Response to Comments on the Proposed Determination
Potential Opportunity Costs of the Standards
As discussed in the Summary of Comments on the Draft TAR Addressed in the Proposed
Determination above, EPA has extensive discussion and analysis of the potential opportunity
costs of the standards in Section III. A. of the Proposed Determination, Sections B.1.4 and B.1.5
of the Appendix to the Proposed Determination, and Chapters 4.1 and 4.2.2 of the TSD. In those,
EPA has not found systematic evidence of tradeoffs between fuel economy and other vehicle
attributes, including performance. EPA agrees with various commenters that the standards have
stimulated significant innovation; we argue that the innovation stimulated by the standards
allows the possibility—in some cases the reality (e.g., the Ford F-150)—of getting improvements
in both fuel economy and other vehicle attributes. EPA agrees with Global Automakers that its
analysis of auto reviews does not measure consumer response. Nevertheless, we consider
professional auto reviewers to be able and expected to identify any significant problems if they
are observed.
IER cites an outdated version of Whitefoot et al.; the most recent version, from 2013,
produces much lower estimates of consumer impacts.45 UCS's finding of consumer surplus is
likely to arise because Whitefoot et al. (2011) considered several scenarios; consumer surplus
increased in some, and decreased in others. The 2013 version presents fewer scenarios and shows
some losses in each. We note that Whitefoot et al. use old (2008) cost data, and expected that
automakers could only comply with the standards by decreasing acceleration; as discussed in
Chapter 3.1.5 of the Draft TAR, horsepower has instead increased. For that reason, we do not
rely on this paper's results. IER cited one paragraph of EPA's assessment of Klier and Linn
(2016), claiming it is "hardly enough to dispose of the serious objections" of this and similar
studies. EPA's assessment of that and similar studies, in Chapter 4.1.3.1 of the Draft TAR,
Appendix Section B.1.4 of the Proposed Determination, and Chapter 4.1.2 of the TSD, are much
more detailed than the one (summary) paragraph referred to in IER's comments. In essence,
while EPA does not dismiss the potential concerns of tradeoffs between fuel economy and other
vehicle attributes, neither does it find that these concerns have to date actually occurred, except
via the predicted pathway of costs of achieving the standards. As discussed in those sections of
the Draft TAR and Proposed Determination, EPA expects that innovation related to other vehicle
attributes will continue in the presence of these standards, and there may as a result be
improvements in some other vehicle attributes during this time independent of the standards.
Regarding the anonymous commenter's concerns, EPA has not observed a lack of small
vehicles. Although the average vehicle has increased slightly in size, small vehicles are still
produced and sold; fueleconomy.gov, for instance, lists 104 subcompact cars in MY2016.46 The
declining market share of manual transmissions follows a historic trend; it unlikely to be due to
increasingly stringent emissions standards, and instead is more likely the result of changing
consumer preferences and manufacturer decisions about which models to offer with a manual
45 Whitefoot, Kate, Meredith Fowlie, and Steven Skerlos (2013). "Compliance by Design: Industry Response to
Energy Efficiency Standards." https://nature.berkeley.edu/~fowlie/whitefoot_fowlie_skerlos_submit.pdf,
accessed 12/30/2016.
46 Fuel Economy Guide 2016 Datafile, http://fueleconomy.gov/feg/download.shtml, accessed 01/09/2017.
131
-------
transmission given the relatively low market demand. The same database lists 201 vehicles
offered with manual transmissions.
We agree with various commenters that the standards not only stimulate innovation, but also
that innovation contributes to the competitiveness of the U.S. auto industry in a world where
many countries seek to reduce vehicle GHG emissions. ICCT, for instance, provides a chart
showing existing and future standards becoming more stringent for a number of countries. We
also agree that finalizing the Determination will provide certainty that will promote more
investment in fuel-saving technologies.
Consumer Acceptance of Electrified Vehicles
EPA continues to rely on its analysis of the technology cost and effectiveness of vehicles,
which indicates that the standards can be met with very low levels of electrification. Automakers
are able to choose their own compliance paths, including increased use of electrification, and
may use the flexibilities under the standards to bank or trade some of the credits associated with
that choice. Those flexibilities are expected to reduce the costs of electrified vehicles.
Toyota observes that a 36-mpg midsize conventional gasoline car does not approach the
efficiency of the hybrid Prius, at 56 mpg. The Prius is not the only midsize hybrid, though. The
2016 Toyota Camry Hybrid, depending on trim, gets 40-41 mpg, while the Kia Optima Hybrid
gets 37-38 mpg.47 The advantages of these HEVs over high-efficiency conventional midsize cars
is obviously much smaller than the advantages of the Prius relative to these vehicles, and may
help to explain why HEV sales are not very high.
Section B.l.5.2 of the Appendix to the Proposed Determination assessed factors relevant for
consumer acceptance of PEVs. We continue to expect consumer acceptance to be sufficient for
the levels of electrification that EPA projects. We agree with various commenters that increased
marketing and incentives for electrified vehicles will aid in their adoption. We also agree with
various automakers that there are limits to their abilities to cross-subsidize vehicle prices while
still remaining profitable; given the longstanding differences in the profitability of different
vehicles, though, those limits are unlikely to be zero. While Global Automakers observes that
many HEV and PEV owners replaced their vehicles with gasoline vehicles, Ford, in a recent
press release, observes that 92 percent of Ford EV owners expect to purchase another EV as their
next vehicle, and 87 percent of PHEV owners want another PHEV for their next vehicle.48 Thus,
for the low levels of electrification projected for achievement of the standards, EPA does not
foresee significant obstacles to consumer acceptance of electrified vehicles.
3.3 Affordability
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Section III. A. of the Proposed Determination, Section B. 1.6 of the Appendix to the Proposed
Determination, and TSD Chapter 4.3 discussed EPA's assessment of the effects of the standards
47 Fuel Economy Guide 2016 Datafile, http://fueleconomy.gov/feg/download.shtml, accessed 01/09/2017.
48 Ford News (January 2017). "Ford Adding Electrified F-150, Mustang, Transit by 2020 inMajorEVPush;
Expanded U.S. Plant to Add 700 Jobs to Make EVs, Autonomous Cars."
https://media.ford.com/content/fordmedia-mobile/fna/us/en/news/2017/01/03/ford-adding-electrified-f-150-
mustang-transit-by-2020.html, accessed 01/04/2017.
132
-------
on affordability. These sections include an overview of the comments EPA received on this topic
in the Draft TAR and EPA's responses to these comments.
As discussed in Appendix Section B. 1.6.1, some comments discussed whether the standards
have a more significant impact on upper-income or lower-income households (the progressivity
of the standards). There does not appear to be agreement in the cited studies. Other comments in
this section argued that the standards will make lower-income households better off, because
they spend more on fuel than they do on vehicles. Comments discussed in Appendix Section
B. 1.6.2 argued that prices of vehicles have been increasing; as EPA discussed there, when
adjusted for inflation, prices have not been increasing; and the trends do not account for changes
in sales mix, which in itself can increase average prices. See Chapter 3.1 above. Some
comments in the Proposed Determination Appendix B. 1.6.2 and B.l.6.3 expressed concern that
access to credit and other indicators of vehicle affordability will be adversely affected as
macroeconomic conditions, such as loan rates, change. EPA agreed that vehicle affordability will
be affected by macroeconomic conditions, but the standards are likely to have at most a small
role in any of those changes.
Summary of Comments on the Proposed Determination
Progressivity/regressivity of the standards
Since the Proposed Determination was issued, two working papers, cited by Defour Group (as
an attachment to comments from the Alliance of Automobile Manufacturers), have asked
whether the vehicle GHG standards provide proportionately more benefits to lower-income
households (that is, they are progressive) or to higher-income households (regressive) than a
gasoline tax. Both Levinson49 and Davis and Knittel50 examine this question by considering fuel
economy/greenhouse gas standards as a tax on inefficient vehicles (those below their target
efficiencies) and a subsidy on efficient ones (those above their target efficiencies). Defour Group
further argues, citing Levinson, that the footprint-based standard is more regressive than a flat
standard. Because low-income households disproportionately buy used vehicles, and because
larger vehicles tend to survive longer than smaller used vehicles, Defour Group argues that the
standards will lead to lower-income households having more powerful and less efficient
vehicles.
The American Council for an Energy-Efficient Economy (ACEEE), Consumer Federation of
America (CFA), Consumers Union, and the Center for Biological Diversity (CBD), in contrast,
say that low-income households spend more on fuel than on vehicles, and that that low-income
households will benefit more than average, both financially and in terms of health benefits.
ACEEE points out that Davis and Knittel did not consider fuel savings in their calculation; if
they had, the standards would be progressive. It also observes that Davis and Knittel did not
account for innovation, which would be expected to decrease the costs of meeting the standards
over time. CU points out that lower-income households benefit from the depreciated prices of
used vehicles. The Environmental Defense Fund argues that used-vehicle buyers will get more
49 Levinson, Arik (December 2016). "Energy Efficiency Standards Are More Regressive than Energy Taxes: Theory
and Evidence." NBER Working Paper 22956, http://www.nber.org/papers/w22956.
50 Davis, Lucas W., and Christopher R. Knittel (December 2016). "Are Fuel Economy Standards Regressive?"
NBER Working Paper 22925, http://www.nber.org/papers/w22925.
133
-------
diverse and more efficient choices, because the standards lead to improvements in fuel economy
across the fleet.
Vehicle price changes
Since the Proposed Determination was issued, some new analyses raise questions about trends
in vehicle prices.
Baum and Luria51 as well as the Union of Concerned Scientists (UCS) and Consumers Union
argue that increases in average vehicle prices are due to: higher market share for trucks, which
are more expensive than cars; a stronger economy that supports higher prices; more content,
beyond emissions and safety requirements, being added to vehicles; and the high income of those
who buy new vehicles. D. Simmons, Furth of the Heritage Foundation,52 and the Institute for
Energy Research (IER, citing Heritage Foundation's assessments) argue that EPA's assessment
of price trends is incorrect. First, Furth argues that EPA's interpretation of the price trend is
incorrect. EPA's understanding was that this price index did not hold sales mix constant. If so,
then even if vehicle prices are constant, average vehicle prices would increase if people are
buying more expensive vehicles (as Baum and Luria argue). Furth instead states his belief that
the price index holds sales mix constant. IER proposes two interpretations of the price trends
since the recession: one, that the increase in vehicle prices was due to higher standards rather
than improvements in vehicle quality (its preferred interpretation); alternatively, it asserts that
EPA's interpretation is that people have "wanted to splurge on getting fancier models." Ford
states that "content-equivalent real vehicle prices have been flat in recent years during a period
of industry growth and declined during the pre-recession years" reflecting consumer expectations
for content-equivalent vehicles. With expectations of reduced real disposable income growth in
coming years, Ford is concerned that price increases will be difficult to support in the future.
UCS disagrees with Furth, arguing that the Heritage Foundation study53 does not account for
mix shift and for improvements in other vehicle attributes. The American Council for an Energy-
Efficient Economy (ACEEE) presents price data indicating that prices of cars have dropped in
real terms, while prices of light trucks have increased slightly; it finds that, if the fleet mix had
stayed at 2009 levels, average expenditures would be $1331 lower than they are.
Fiat Chrysler states that EPA has not properly considered effects on the used car market,
including the potential for the "jalopy effect," where higher prices for new vehicles lead people
to hold onto used vehicles longer.
51 Baum, Alan, and Dan Luria (December 2016). "Analyst Brief: Affordability of Vehicles Under the Current
National Program in 2022-2025 for Detroit Three Automakers." https://www.ceres.org/files/analyst-brief-
affordability-of-vehicles-under-the-current-national-program-in-2022-2025-for-detroit-three-
automakers/at download/file, accessed 12/28/2016.
52 Furth, Salim (December 2016). "Issue Brief: Regulation Continues to Increase Car Prices."
http://www.heritage.org/research/reports/2016/12/regulation-continues-to-increase-car-prices, accessed
12/28/2016.
53 Furth, Salim, and David W. Kreutzer (2016). "Fuel Economy Standards are a Costly Mistake." The Heritage
Foundation Backgrounder, http://www.heritage.org/research/reports/2016/03/fuel-economy-standards-are-a-
costly-mistake, downloaded 5/20/2016.
134
-------
Effects on Access to Transportation
Defour Group (in an attachment to comments from the Alliance of Automobile
Manufacturers) argues that lower-income households especially need access to transportation,
arguing that it is irrelevant whether there is a commonly accepted definition of access. It cites
studies from 2003 emphasizing the role of personal transportation—vehicle ownership—to
lower-income households.
Access to Credit
Global Automakers and Toyota point out the increasing average length of loans on new
vehicles as a way for consumers to reduce their monthly loan payments, and considers EPA's
analysis of related issues inadequate. NADA argues that lenders do not take potential fuel
savings or reduced operating costs into account for loans. IER points out that EPA did not cite a
2012 study54 on credit-constrained consumers; Ford criticizes EPA for not estimating how many
people would not be able to get financing due to the standards.
Response to Comments on the Proposed Determination
Progressivity/regressivity of the standards
EPA agrees with ACEEE that both Levinson (2016) and Davis and Knittel (2016) look only at
the implicit tax/subsidy on vehicles due to the standards. They do not allow for the possibility
that the standards lead to fuel savings that exceed technology costs. As Levinson's paper says,
"if the regulation makes consumers better off even ignoring the environmental benefits, then the
distributional comparison with a gasoline or energy tax becomes moot" (p. 6). As shown in
Table ES-4 of the Final Determination, EPA expects the fuel savings to exceed the technology
and maintenance costs by $56 billion (3 percent discount rate; $26 billion at 7 percent), before
consideration of the environmental benefits. EPA notes that the Levinson and Davis and Knittel
papers focus on comparing the progressivity of fuel taxes to standards. Because EPA does not
have statutory authority to tax gasoline, the relative merits of a gasoline tax are not relevant to
EPA's determination on the appropriateness of the MY2022-2025 standards. In addition,
Levinson's assessment of the footprint-based standard compared to the flat standard, cited by
DeFour, is based on a stylized model where all vehicles are alike. As EPA previously pointed out
in Appendix Section B. 1.6.1 of the Proposed Determination, vehicle size may play a role in
consumer welfare from the standard; the reduced incentives to downsize vehicles under the
footprint standard may be an important feature that Levinson (cited by DeFour) has not
incorporated into his model. In comments on the Draft TAR, Greene and Welch presented a
paper that argued, not only that the standards are progressive, but that all income groups benefit
from the standards. EPA expects that the gains to consumers from the net fuel savings will be
widespread, even if it is not yet definitive which income groups benefit more. For these reasons,
EPA continues to find that the evidence on the progressivity or regressivity of the standards is
inconclusive. EPA expects that the gains to consumers from the net fuel savings will be
widespread, even if it is not yet definitive which income groups benefit more.
54 Wagner, D., P. Nusinovich, and E. Plaza-Jennings, National Automobile Dealers Association (February 13, 2012).
"The Effect of Proposed MY 2017-2025 Corporate Average Fuel Economy (CAFE) Standards on the New
Vehicle Market Population." Docket EPA-HQ-OAR-2010-0799-9575.
135
-------
Vehicle price changes
We agree with Baum and Luria and others that the factors they cite might all contribute to
higher vehicle prices. As discussed in the Proposed Determination Appendix Section B.l.6.2 and
TSD Chapter 4.3.3.3, EPA found that prices in recent years, adjusted for quality and inflation,
have been flat, not increasing. Because this price series adjusts for changes in content (i.e.,
quality), the flat trend does not contradict Baum and Luria's comment on increasing prices,
which is based in part on added content. The International Energy Agency uses cross-country
comparisons to conclude similarly that "average vehicle price is not strongly driven by fuel
economy parameters, but rather by other attributes" (p. 39).55 It finds that the major factors
affecting vehicle price are vehicle segmentation and the market share of premium brands, not
fuel economy.
The memo to the docket "Review of Heritage Foundation analyses, 'Fuel Economy Standards
are a Costly Mistake' by Furth and Kreutzer, and 'Issue Brief: Regulation Continues to Increase
Car Prices' by Furth" provides a more detailed assessment of these two analyses from the
Heritage Foundation.56 We continue to disagree with the Furth and Kreutzer argument that the
standards have increased prices relative to counterfactual price trends, because we do not find
that the price trends that Furth and Kreutzer cite meet the statistical criteria for representing
what could have been expected to happen in the absence of the standards: closely matching price
trends before the shock, and being expected to respond to common price shocks in the same
way.57 We do not find that the price trends cited for comparison—furnishings and durable
household equipment, or vehicle prices in the U.K. or Australia—meet these criteria.
As we note, Furth does not provide citations to document his claim that the price indices do
account for sales mix. EPA has confirmed with both the Bureau of Labor Statistics (BLS) and
the Bureau of Economic Analysis (BEA) that changes in sales mix do in fact affect their price
indices. For instance, documentation of BLS's sampling methodology58 in the section of
Selection Procedures within Outlets (p. 15) states, "The probabilities of selection are
proportional to the sales of the items included in each group." In other words, if consumers are
buying more expensive vehicles, the probability for being selected as a sample for Consumer
Price Index would be higher for expensive vehicles. We note that Ford observes flat content-
equivalent prices in recent years. In that case, addition of content—added either due to the
standards or for other reasons—may contribute to the unobserved price increases cited.
EPA, similar to ACEEE, conducted an exercise to look at how changing market shares would
affect average vehicle prices. This exercise, documented in the memo cited above, used sales and
55 International Energy Agency (2017). "International Comparison of Light-Duty Vehicle Fuel Economy 2005-2015:
Ten years of fuel economy benchmarking." http://www.globalfueleconomy.org/media/418761/wpl5-ldv-
comparison.pdf, accessed 01/11/2017.
56 Assessment and Standards Division, Office of Transportation and Air Quality, U.S. EPA (December 2016).
"Review of Heritage Foundation analyses, 'Fuel Economy Standards are a Costly Mistake' by Furth and
Kreutzer, and 'Issue Brief: Regulation Continues to Increase Car Prices' by Furth." Docket EPA-HQ-OAR-2015-
0827.
57 Abadie, Alberto (2005). "Semiparametric Difference-in-Differences Estimators." Review of Economic Studies
72(1): 1-19; Dimick, JustinB. and Andrew M. Ryan (2014). "Methods for Evaluating Changes inHealth Care
Policy: The Difference-in-Differences Approach." JAMA 312(22): 2401-2402.
58 U.S. Department of Labor, Bureau of Labor Statistics. "Chapter 17. The Consumer Price Index (Updated
06/2015)." https://www.bls.gov/opub/hom/pdf/homchl7.pdf, accessed 1/5/2017.
136
-------
price data from model years (MYs) 2010 and 2015 to examine the effects of sales mix (at the
class level) on vehicle prices between these years. Similar to ACEEE's findings, EPA found that,
if the prices for MY2015 are used with the sales mix in MY2010, then average vehicle price
would be about $1000 lower than the actual average price, just because of the different sales mix.
Similarly, the sales mix in MY2015 at MY2010 prices would have an average price over $1000
more expensive than the actual average price. (All values are in 2015$). In other words, with
vehicle prices held constant (at either 2010 or 2015 values), the sales mix is a more expensive
one in MY2015 than in MY2010. Thus, changes in sales mix can contribute to average price
increases. We note that this exercise does not account for changes in vehicle quality during that
time; as a result, we cannot determine whether the increase in price for a constant sales mix is
due to changes in vehicle content added due to the standards or for other reasons. We thus
disagree with IER that the full increase in vehicle prices in recent years is due to the standards;
instead, based on these analyses, we find that the increase in vehicle prices is due in part to
people buying more expensive vehicles.
As EPA discussed in Appendix Section B. 1.6.2 of the Proposed Determination, EPA expects
(and has observed) that used vehicle prices respond fairly closely to changes in the new vehicle
market. The "jalopy effect" might happen if the standards lead to decreased new vehicle sales,
which would lead to people buying used vehicles, or holding onto used vehicles longer. Because
EPA has not projected the effects of the standards on new vehicle sales, it does not conclude
whether the standards might lead to higher or lower used vehicle prices.
Effects on Access to Transportation
EPA disagrees with the Defour Group that it is possible to evaluate trends in access to
transportation in the absence of a way to evaluate or define access to transportation. The market
for mobility services is changing rapidly, with increased urbanization, new opportunities for ride-
sharing, changing interest in demographic groups in vehicle ownership, and the possibility in the
near future that owning a vehicle may not be a prerequisite to quality transportation services.
Thus, access to transportation for many does not now require a personal vehicle, and changes in
the definitions of access to mobility are likely to continue.
Access to Credit
Issues related to access to credit depend heavily on changes in vehicle prices. As discussed
above, trends for vehicle prices in recent years appear to be due in part to changes in the mix of
vehicles sold. EPA agrees that access to credit is likely to be more difficult for more expensive
vehicles, but with some caveats. First, as EPA has previously discussed, we do not expect price
increases as large as those provided in the CAR Report cited by some automakers. Second, a
number of lending institutions do provide some reduction in interest rates for more efficient
vehicles. Third, EPA has examined the role of the debt-to-income ratio in loans for new vehicles,
and there does not appear to be a strict boundary beyond which a potential borrower cannot get a
vehicle loan. Fourth, lower-priced vehicles continue to exist, and appear to be gaining new
features while staying low-priced. EPA agrees that longer-term loans reduce monthly payments
by extending loan periods; these longer-term loans may put some purchasers at risk of owing
more than their vehicles are worth at some future time. Baum and Luria,59 cited by UCS, argue
59 Baum, A., and D. Luria. (December 21, 2016). "Fuel economy rules a bogeyman for long-term trends in auto
137
-------
that these trends are due in large part to factors other than the standards; they argue that
automakers are focusing their sales efforts on higher-profit, higher-trim models, for instance.
Many of the concerns raised about the future of vehicle sales due to rising interest rates and
stagnant incomes are happening regardless of the standards. As we have previously argued, the
role of the standards in vehicle sales is likely to be secondary to these larger impacts. EPA
observes, regardless, that the fuel savings provide more money for people to pay their loans.
EPA did not cite the Wagner et al. (2012) study because, in the 2012 FRM, it found it
significantly flawed (see 77 Federal Register 62950-62951). We also do not estimate the number
of households who would not be able to get loans due to the standards, as recommended by Ford,
because doing so would require identifying which households were likely to buy new vehicles;
we are not aware of data that reflect both likelihood to purchase a vehicle, what the expected
vehicle price would be for that household, and the detailed financial data needed to identify the
household's likely ability to get a loan. (One of the flaws in Wagner et al. was that it did not
separate households likely to purchase new vehicles from those unlikely to purchase new
vehicles).
3.4 Employment
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Section IV.B. of the Proposed Determination and Section B.2 of the Appendix to the
Proposed Determination discussed EPA's assessment of the effects of the standards on
employment. These sections, plus Section B.1.3 of the Appendix and TSD Chapter 4.2.1 include
an overview of the comments EPA received on this topic in the Draft TAR and EPA's responses
to these comments.
Some comments criticized EPA for not quantifying the effects of changes in vehicle sales on
employment in the auto sector. As EPA discussed in Section B.2.4 of the Appendix to the
Proposed Determination, EPA did not quantify this effect because of its lack of confidence in
estimates of the sales impacts of the standards. Sections B. 1.3.3 of the Appendix and TSD
Chapter 4.2.1 include comments on quantified estimates of the effects of the standards on sales
and employment. Some comments asked EPA to incorporate "multiplier" effects into its
employment estimates. Multipliers estimate the effects in the broader macroeconomy of impacts
in the regulated sector. Appendix Section B.2.3 points out that using multipliers for national-
scale analysis is only appropriate when the economy suffers from significant involuntary
unemployment. Other comments pointed out the potential for both employment gains and
employment losses in industries related to the auto industry. As Appendix Sections B.2.4 and
B.2.5 respond, net effects on overall employment depend heavily on the state of the
macroeconomy, as jobs may shift among sectors. The Proposed Determination concluded that for
the regulated sector, the partial employment impact due to the substitution effect of increased
costs of vehicles is expected to be positive. EPA did not estimate the total effects of the standards
in the regulated industry because the total effect of the standard on motor employment depends,
at least in significant part, on changes in vehicle sales; EPA found these estimates too unreliable
to quantify. (EPA also indicated that such quantification is not required by either the Midterm
Evaluation rules, or section 202(a)(1) of the Act, concluding that it is better to make reasonable
industry." The Hill, http://thehill.com/blogs/congress-blog/energy-environment/311281-fuel-economy-rules-a-
bogeyman-for-long-term-trends-in, accessed 01/09/2017.
138
-------
albeit qualitative predictions than insufficiently supported quantitative estimates. See Proposed
Determination at p. A-87).
Summary of Comments on the Proposed Determination
The Blue-Green Alliance (BGA), Center for Biological Diversity, Environmental Defense
Fund (EDF), Woodward, and a joint letter from 74 environmental groups point to the growth in
employment in the auto industry since the recession, the first sustained period of job growth in
the industry since 1999 (BGA). They argue that the standards have increased employment, in
part because of the innovation that the standards have triggered. BGA points to "more than 1,200
factories and engineering facilities in 48 states" (p. 2) producing fuel-saving technologies or the
materials for them. EDF points to employment gains associated with the innovation in the Ford
F-150, listing 7 domestic plants with significant employment and investment associated with that
vehicle. BGA argues that job growth depends on innovation, investment, and manufacturing in
the U.S., to which the standards have contributed. UAW also notes that the GHG standards have
spurred investments in new products that employ tens of thousands of its members.
The Alliance of Automobile Manufacturers and Fiat Chrysler assert that EPA is deficient
because it has not fully quantified employment impacts. The Alliance claims that EPA has not
assessed economic impacts to the auto industry. The Union of Concerned Scientists cites a study
from Ceres arguing that the industry will be profitable under the standards, even with low fuel
prices. The Alliance points out that EPA cited a "projection and analysis of the likelihood of
mass unemployment and bankruptcy for a large U.S. automaker based on connection with the
promulgation of much less stringent standards" (p. 20), 74 Fed. Reg. 49485, a reference to the
reference case analysis in the NPRM for the MY2012-2016 standards. Fiat Chrysler argues that
the partial quantification that EPA uses is misleading, and suggests that EPA wait because it "is
currently developing a comprehensive model for studying the economy-wide effects of the
National Program" (p. 24).
The Center for Automotive Research (CAR) explains that it does not quantify substitution
effect employment in its analysis because it finds that expenditures and sales on vehicles
decrease, indicating that consumers do not want the added content associated with the increased
cost. In addition, CAR argues that the new technologies substitute for existing technologies, and
thus do not add labor. EDF and UCS, on the other hand, state that suppliers gain from the
standards. BGA, the International Council for Clean Transportation (ICCT), the Natural
Resources Defense Council, and UCS point out that innovative technologies spur employment in
other industries; indeed, weakening standards would put jobs and competitiveness at risk.
The Alliance, in comments on EPA's substitution effect analysis, requests that EPA adjust its
substitution effects multiplier due to increases in productivity over time, and should account for
"factor shift effects" due to decreases in employment for some technologies. In discussion of the
substitution effect, the Alliance commented that electrified vehicles pose particular threats to
auto sector employment, citing the potential for significantly reduced labor content in large,
consolidated parts such as batteries and electric motors, and speculating (based on an informal
comment describing the aspirations of one manufacturer) that BEV production lines may
someday eliminate workers entirely (p. 23-24, the Alliance comments). Also discussing the
substitution effect, FCA suggested that many of the jobs created by production of lithium-ion
batteries (implicitly in a scenario of higher market penetration than anticipated by the EPA
139
-------
projections) will be created overseas where the currently leading manufacturers are based, not in
the U.S.
EDF points out that exports have increased; Honda for instance, is a net exporter from the
U.S.
BGA, EDF, and UCS also observe that improved fuel economy reduces the exposure of the
auto industry to fuel price shocks, and therefore reduces employment fluctuations that have
resulted from that historic exposure.
Consumer Federation of America (CFA) points to employment gains from consumer savings
on fuel: as consumers switch from spending on fuel to other goods, they will increase economic
output and personal welfare.
UCS mentions a study from Dziczek et al. (2016)60 on how differences in fleet mix for light
trucks across the auto industry leads to non-uniform employment impacts, focusing on whether
the most profitable trucks will be able to meet the standards. UCS notes issues with the technical
analysis in the study (such as including medium- and heavy-duty pickups that are not subject to
these standards), and comments that its use of multipliers for employment impacts overstates the
likely impacts.
Response to Comments on the Proposed Determination
Employment has increased in the auto industry as it has recovered from the recession, as
various commenters observe. We also agree that the standards have stimulated research and
innovation, and that there are employment gains both directly in the innovation effort and in
producing fuel-saving technologies. We also agree that the auto industry has been able to achieve
the standards while being highly profitable. EPA describes the employment impacts associated
with producing the new technologies the "substitution effect," discussed more below.
EPA disagrees that it has not assessed economic impacts to the auto industry. The basis for
this claim appears to be EPA's not quantifying sales and employment impacts of the standards.
In the Proposed Determination, consistent with section 86.1818-12(h)(v) and (viii), EPA
evaluated potential employment impacts. See PD App. B.2. We considered potential impacts on
the regulated sector, as well as on other directly related sectors (e.g. motor vehicle parts
manufacturing, auto dealers, and fuel suppliers). We considered costs for the regulated sector
both for the industry as a whole and at the individual firm level. In doing so, we discussed and
evaluated both the output effect (relation of production changes and labor demand) and
substitution effect (if output is constant, how regulation affects labor-intensity of production),
quantifying the second of these. Id. at B.2.4.2.
Commenters reiterated their arguments that this analysis was necessarily deficient because it
failed to quantify output effects and to fully quantify substitution effects. As EPA indicated,
however, neither the rules establishing the MTE, nor section 202(a) of the Act, contain any
requirement that estimates of employment impacts be quantified. See, e.g., id. at A-87. The
commenters did not cite any legal principle that would compel such quantification here, nor is
60 Dziczek, K., B. Smith, Y. Chen, M. Schultz, and D. Andrea (2016). "The economic implications of potential
NHTSA and EPA regulatory revisions on U.S. light truck sales and manufacturing." Center for Automotive
Research. www.cargroup.org/?module=Publications&event=View&pubID=145, accessed 01/05/2017.
140
-------
EPA aware of any. Indeed, courts have supported qualitative determinations for decisions that
are even more consequential. S qq American Trucking Associations v. EPA, 283 F. 3d 355, 362
(D.C. Cir. 2002) (upholding qualitative determinations in establishing primary National Ambient
Air Quality Standards); American Farm Bur. Fed. v. EPA, 559 F. 3d 512, 535 (D.C. Cir. 2009)
("It is true that the EPA relies on a qualitative analysis to describe the protection the coarse PM
NAAQS will provide. But the fact that the EPA's analysis is qualitative rather than quantitative
does not undermine its validity as an acceptable rationale for the EPA's decision"). EPA
reasonably explained, and repeats here, that it chose not to further quantify employment impact
estimates due to legitimate uncertainties in the validity of such estimates - based on a close
survey of the peer-reviewed literature, as well as careful consideration and discussion of the
various quantitative estimates in the public comments. Proposed Determination Appendix B.2.4
and B.2.5. EPA concluded that "EPA views it preferable to consider an issue with reliable
qualitative information than unhelpfully wide-ranging estimates." Id. at A-87. Such an approach
again has judicial support. See, e.g., State of Mississippi v. EPA, 744 F. 3d 1334, 1352 (D.C.
Cir. 2013) (referring to "the inviolable law" of "garbage in, garbage out" in upholding EPA
decision to place minimal weight on quantified risk assessment in revising the National Ambient
Air Quality Standard for ozone).
The Alliance implies since EPA evaluated the likelihood of mass unemployment and
bankruptcy for a large U.S. automaker" under "much less stringent standards" the current
standards could therefore be expected to lead to mass unemployment and bankruptcies for
automakers. This is an incorrect implication. The quote refers to the financial status of Chrysler
in 2009, which was highly uncertain. In developing its reference case for the MY2012-2016
standards, which EPA proposed in 2009, it acknowledged the then-uncertain status of that
company. Clearly Chrysler's financial status was not due to compliance with those standards,
which had not even been proposed at the time of Chrysler's difficulties, and there is no reason to
believe that the current standards will lead to mass unemployment or bankruptcy for a large U.S.
automaker.
Unlike Fiat Chrysler, EPA does not consider its partial quantification misleading. The
discussion of the effects is clear that we do not fully quantify employment effects; Fiat Chrysler
as well as others understood EPA's employment analysis sufficiently to criticize us for not fully
quantifying effects. Fiat Chrysler is incorrect that EPA's Science Advisory Board panel is
"developing" a model for economy-wide effects; rather, the panel is examining the suitability of
such models for regulatory analysis.61 While the findings of the panel will be helpful in
informing future regulatory work, they will not provide a model suitable for use in this
determination, regardless of timing.
EPA addresses CAR's comments on substitution effect employment in Chapter 3.1 of this
RTC document, in the context of other comments on CAR's report. As we note there, CAR's
explanation appears to be about the output effect, not the substitution effect. In addition, EPA
finds that part of the costs of the new technologies is labor, regardless of whether the
technologies substitute for existing components.
61 McGartland, A1 (2015). "Transmittal of Charge to the Science Advisory Board Advisory Panel on
Economy-Wide Modeling of the Benefits and Costs of Environmental Regulation."
https://yosemite.epa. gov/sab/sabproduct.nsf/0/07E67CF77B54734285257BB0004F87ED/$File/Charge%20Quest
ions%202-26-15.pdf, accessed 01/09/2017.
141
-------
EPA long ago adopted the practice requested by the Alliance that it adjust its multipliers due
to increases in productivity over time; it uses historic trends in productivity to adjust its future
multipliers (see, e.g., pp. A-90 to A-91 of the Appendix to the Proposed Determination). We note
that automation has been a significant force behind changes in employment in the auto sector for
decades, independent of the standards. Though "factor shift" employment might be lower for
some technologies, such as battery and motor production, it may increase with other
technologies; for instance, in the RIA for the MY2017-2025 standards (Table 8.2-4, p. 8-31),
EPA provided estimates of the increased labor associated with some of the technologies expected
to be used to meet the standards, based on the FEV teardown studies mentioned by the Alliance.
Regarding the employment effects of electrification, while early dominance in lithium-ion
batteries was overseas, domestic production has been introduced almost in tandem with
introduction of volume-production PEVs such as the Leaf and Volt: Nissan and LG Chem both
built large battery manufacturing facilities in the U.S. several years ago. Tesla has just started
production at the Gigafactory, which is expected by 2018 to produce over 35 GWh/year, "nearly
as much as the rest of the entire world's battery production combined,"62 and create 6,500 jobs
by 2020. We note that the weight of battery packs, and shipping considerations (flammability)
likely have played a part in the decision to manufacture these bulky and volatile items
domestically; for those manufacturers not currently producing battery packs in the U.S., these
factors are likely to continue to exert pressure as a way to keep costs low in the future. The
Bloomberg article cited above says 95% of the Tesla Model 3 components will be made in the
U.S., suggesting that electrification need not inherently send jobs overseas.
EPA agrees with BGA, EDF, and UCS that reducing auto industry exposure to shocks in fuel
prices should reduce fluctuations in sales and employment in the auto industry. EPA also agrees
with CFA that consumers will benefit from being able to transfer expenditures on fuel to other
goods. Though, as discussed in Appendix Section B.2.5.2 and B.2.5.3, there may be reduced
employment in sectors related to fuel supply, there are likely to be increases in employment due
to substituting expenditures on fuel to expenditures in other sectors. EPA agrees with UCS that
the use of multiplier impacts, as in the Dziczek et al. study, is not appropriate in the context of a
national program; see the discussion in Section B.2.3 of the Appendix to the Proposed
Determination.
62 Tesla Team (January 4, 2017). "Battery Cell Production Begins at the Gigafactory."
https://www.tesla.com/blog/battery-cell-production-begins-gigafactory , accessed 01/05/2017; Randall, Tom
(January 4, 2017). "Tesla Flips the Switch on the Gigafactory." Bloomberg News,
https://www.bloomberg.com/news/articles/2017-01-04/tesla-flips-the-switch-on-the-gigafactory, accessed
01/05/2017.
142
-------
3.5 Economic and Other Key Inputs
Summary of Comments on the Draft TAR addressed in the Proposed Determination
In the Draft TAR, the agencies documented in detail how we developed several input values
used in various economic and other analyses. Except as discussed below, EPA did not receive
comments on the following economic and other input values presented in the Draft TAR:
• On-road Fuel Economy Gap
• Fuel Prices
• Rebound Effect
• Energy Security
• Reduced Fueling Time Benefits
• Impacts of Additional Driving
• Discounting Future Benefits/Costs
• Additional Costs of Vehicle Ownership
Regarding the social cost of carbon (SCC), several comments received on the Draft TAR
stated that the SC-CO2, SC-CH4, and SC-N2O underestimates climate-related benefits and
discussed some of the technical details of the modeling conducted to develop these estimates. As
noted in the TSD Chapter 3.7, EPA recognizes the importance of the estimates to be as complete
as possible and will continue to follow and evaluate the latest science on impact categories that
are omitted or not fully addressed in the integrated assessment models.
One commenter also recommended that EPA use undiscounted estimates of the SC-CO2.
Consistent with the recommendations of the Interagency Working Group (IWG) on SC-GHG,
EPA continued to apply the SC-GHG values discounted at rates of 2.5, 3.0, and 5.0 percent.
EPA identified these discount rates in the TSD Chapter 3.7, and referred to the Technical
Support Document: Technical Update of the Social Cost of Carbon for Regulatory Impact
Analysis under Executive Order 12866 (February 2010) ("2010 TSD") for a complete discussion
of the methods used to develop the estimates, including the discount rates. In sum, the 2010 TSD
concluded that arguments made under the prescriptive approach can be used to justify discount
rates between roughly 1.4 and 3.1 percent but that concerns about the most appropriate value for
r| (i.e., the parameter capturing diminishing marginal utility) make it difficult to justify rates at
the lower end of this range under the Ramsey framework. Therefore, in light of disagreement in
the literature on the appropriate market interest rate to use in this context and uncertainty about
how interest rates may change over time, the IWG used three discount rates to span a plausible
range of certainty-equivalent constant discount rates: 2.5, 3, and 5 percent per year. The IWG
further noted that these three rates reflect reasonable judgments under both descriptive and
prescriptive approaches.
Some commenters also provided constructive recommendations for potential opportunities to
improve the SC-CO2 estimates in future updates. EPA responded in TSD Chapter 3.7 that the
U.S. government is seeking input from the National Academies of Sciences, Engineering and
Medicine on how to approach future updates to ensure that the estimates continue to reflect the
best available scientific and economic information on climate change. An Academies committee,
"Assessing Approaches to Updating the Social Cost of Carbon," (Committee) will provide its
final report in early 2017 with advice on the merits of different technical approaches for
143
-------
modeling and highlight research priorities going forward. In the meantime, the IWG continues to
recommend the use of the current SC-CO2 estimates as the best scientific information on the
impacts of climate change available in a form appropriate for regulatory analysis. EPA expects
that, as in the case of the 2013 SC-CO2 update, any future updates to IWG SC-GHG guidance
based on the Academies' recommendations will apply to regulatory impact analyses going
forward.
Several commenters suggested that the vehicle Mileage Accumulation Rates (MAR) should
not be changed from values used in the FRM analysis for our Proposed Determination
assessment. EPA responded in TSD Chapter 3.1 reiterating the value of using the most current
values for the MAR rates (as found in the Annual Energy Outlook 2016).
EPA received a few public comments on the Draft TAR, relating to the omission of certain
non-GHG impacts from the analysis of the program. As we explained in TSD Chapter 3.6, and
have explained in earlier comment periods associated with the light-duty vehicle GHG program,
there are several health benefit categories that EPA is unable to quantify due to limitations
associated with using PM2.5-related benefit per-ton estimates, several of which limitations have
the possibility to be substantial (i.e. to underestimate health benefits). For example, we are not
able to quantify a number of known or suspected health benefits linked to reductions in ozone
and other criteria pollutants, as well as health benefits linked to reductions in air toxics. In
addition, we are unable to quantify a number of known welfare effects, including reduced acid
and particulate deposition damage to cultural monuments and other materials, and environmental
benefits due to reductions of impacts of eutrophication in coastal areas. As a result, the health
benefits quantified in the Draft TAR and TSD were likely underestimates of total non-GHG-
related benefits associated with the program. For this reason, for the analyses in the Proposed
Determination, EPA recognized that the omission of these benefits would not change EPA's
overall policy conclusions about the appropriateness of the existing standards. The Proposed
Determination acknowledged that their inclusion would only increase the amount by which the
quantified benefits outweigh the program's estimated costs.
Summary of Comments on the Proposed Determination
Energy Security: EPA received one comment on the energy security impacts of the Proposed
Determination. This commenter noted that the LDV standards enhance the energy security
position of the U.S. by significantly reducing U.S. imports of oil over time.
Social Cost of GHGs: EPA received two comments on the Proposed Determination that
discussed the social cost of CO2, also referred to as the social cost of GHGs (SC-GHG), to
include estimates of the social cost of non-CC>2 GHGs, such as CH4 and N2O. Of these
comments, one supported the application of the social cost of carbon to monetize CO2, CH4, and
N2O impacts. This commenter also stated that the SC-GHG values used by EPA and other
federal agencies are underestimates, due in part to omission of key climate change impacts, and
that the 3 percent discount rate applied to the SC-GHG values is too high. The second SC-GHG
comment letter stated that the precise magnitude and sign of the social cost of carbon is in
question. This second commenter also noted that they submitted a letter to the Office of
Management and Budget's separate comment solicitation on the SC-GHG (78 FR 70586;
November 26, 2013); the comment period for this separate OMB solicitation ended on February
26, 2014.
144
-------
Response to Comments on the Proposed Determination
Energy Security: EPA agrees with the commenter that the Proposed Determination will
improve the energy security position of the U.S. by reducing oil imports.
Social Cost of GHGs: EPA agrees with the comment that the application of the SC-GHG
values in benefit-cost analysis is appropriate. The comments about omitted climate change
impacts and the discount rates applied to the SC-GHG are identical to those received on the Draft
TAR. Please see the summary above, "Summary of Comments on the Draft TAR addressed in
the Proposed Determination," for details about EPA's response.
Regarding the second comment letter on SC-GHG, which made note of comments submitted
to a separate OMB solicitation on SC-GHG, EPA notes that the interagency working group
(IWG) on the SC-GHG responded to all of the comments submitted to OMB through that
separate comment solicitation on the SC-GHG (78 FR 70586; November 26, 2013). As a
member of the SC-GHG IWG, EPA carefully examined and evaluated comments submitted to
OMB's separate solicitation. EPA has determined that the IWG responses to the comments on
the OMB solicitation address the comments on the SC-GHG methodology, including the
magnitude and sign of the SC-GHG estimates, the selection of discount rates, the use of global
measures, climate sensitivity, damage functions, and other market impacts. Specifically, EPA
concurs with the IWG's response to these comments and hereby incorporates them by
reference.63 The second comment letter also noted the uncertainty of the SC-GHG estimates.
While EPA acknowledges uncertainty in the SC-GHG estimates, EPA disagrees with the
commenter's implication that the uncertainty is so great as to undermine use of the SC-GHG
estimates in regulatory impact analysis. The uncertainty in the SC-GHG estimates is fully
acknowledged and comprehensively discussed in the SC-GHG TSDs and supporting academic
literature. While uncertainty must be acknowledged and addressed in regulatory impact analyses,
even an uncertain analysis provides useful information to decision makers and the public.
After careful evaluation of the full range of comments on the SC-GHG received through the
Draft TAR, the Proposed Determination, and other comment solicitations related to SC-GHG,
EPA has determined that use of the current SC-CO2, SC-CH4, and SC-N2O estimates is
appropriate and that the current estimates continue to represent the best scientific information on
the impacts of climate change available in a form appropriate for incorporating the damages from
incremental GHG emissions changes into regulatory analysis.
63 Referred to as the "OMB Response to Comments on SC-GHG." See
https://www.whitehouse.gov/sites/default/files/omb/inforeg/scc-response-to-comments-final-july-2015.pdf. For
the issues highlighted in the comment letter submitted to the Proposed Determination, see OMB Response to
Comments on SC-GHG, pages 3, 9-10, 25-28, 29-30 (the magnitude and sign of the SC-GHG estimates,
uncertainty), pages 20-25 (the selection of discount rates), pages 30-32 (the use of global measures), pages 11-17
(climate science, including climate sensitivity), pages 6-11 (damage functions), and page 11 (other market
impacts).
145
-------
3.6 Safety
Summary of Comments on the Draft TAR addressed in the Proposed Determination
EPA received a few public comments on the mass/safety analysis contained in Chapter 8 of
the Draft TAR. Natural Resources Defense Council (NRDC) stated a strong belief that the 2025
standards can be achieved without an increased risk to safety, and that the fleet of future vehicles
can be built lighter weight, less polluting and safe. The Alliance of Automobile Manufacturers
commented that it found inconsistencies in the results "that require further physical
explanations." Tom Wenzel, of Lawrence Berkeley National Laboratory (LBNL), on behalf of
Department of Energy (DOE), recommended that the agencies should use a second set of
regression coefficients, such as those used in the "LBNL baseline"64 to run EPA's OMEGA
model, "because the estimated relationships between mass reduction and societal fatality risk are
not consistently statistically different from zero, and are sensitive to the data and variables used
in the regression models."
As we discussed in Section B.3.1 of the Proposed Determination Appendix, and more
specifically indicated at Table B.8 of that document, if we were to apply Wenzel's "LBNL
baseline" in our OMEGA model, the estimates of potential adverse safety implications would be
even lower, which might influence a choice to model greater levels of mass reduction in
assessing potential compliance pathways. We acknowledged the rationale for Wenzel's
recommendation. However, for purposes of the Proposed Determination, we believed it
appropriate to continue using the approach taken in the Draft TAR, since it was more
conservative and we wanted to ensure there are no significant adverse safety implications
associated with the 2022-2025 standards.
Summary of Comments on the Proposed Determination
NRDC and the Alliance of Automobile Manufacturers (the Alliance) commented on the
safety analysis presented in the Proposed Determination. NRDC states its belief that the analysis
is conservative because it "continues to rely on the conservative assessment from the TAR," and
recommends that future assessments of safety impacts should include inputs (e.g. regression
coefficients) that account for deficiencies identified by Tom Wenzel of Lawrence Berkeley
National Laboratory (LBNL). The Alliance claims EPA did not address in the Final
Determination its two concerns over the Draft TAR safety analysis. One concern was EPA's use
of "two highly correlated factors (mass and footprint) in a regression analysis," which was
described as having the potential to lead to non-physical results. The other concern was "the
apparent disconnect between the EPA Draft TAR safety analysis and the NHTSA 2016 VOLPE
Report," which refers to inconsistencies between Table 3-7 of NHTSA's 2016 Volpe Report and
Table 8.7 of the Draft TAR. The Alliance states that "both address the 100 lbs. reduction of the
entire fleet of light vehicles. The estimated increase reported by Volpe is 91, but the increase in
the Draft TAR is 55" and states that EPA offers no explanation for the difference.
64 Tom Wenzel, Table 5.16, "Assessment of NHTSA's Report "Relationships Between Fatality Risk, Mass, and
Footprint in Model Year 2003-2010 Passenger Cars and LTVs, Preliminary report prepared for the Office of
Energy Efficiency and Renewable Energy, U.S. Department of Energy." LBNL -1005177. (July, 2016).
146
-------
Response to Comments on the Proposed Determination
We agree with NRDC that EPA's analysis in the Proposed Determination based on the
assessment from the Draft TAR is conservative as compared to Wenzel's recommended "LBNL
baseline," which might influence a choice to model greater levels of mass reduction in assessing
potential compliance pathways. However, we believe the conservative approach taken in the
Proposed Determination is appropriate, since we want to ensure there are no significant adverse
safety implications associated with the 2022-2025 standards as we have stated previously.
The comment made by the Alliance asserting that EPA did not address its concern in the
Proposed Determination on "EPA's use of two highly correlated factors (mass and footprint) in
the regression analysis" is not accurate. First, EPA did not perform its own safety statistics
(regression) analysis. We relied on the Draft TAR's safety statistics results as inputs to the
OMEGA model in the Proposed Determination. Second, the issue of "use of two highly
correlated factors (mass and footprint) in a regression" was discussed and reviewed extensively
in Chapter 8 of the Draft TAR. We believe using two or more variables that are strongly
correlated in the same regression model (referred to as multicollinearity) can lead to inaccurate
results. However, the correlation between vehicle mass and footprint may not be strong enough
to cause serious concern. See Draft TAR, Chapter 8.2.4.6, p. 8-33. As stated at page 4, Executive
Summary in the NHTSA 2016 preliminary report,65
"NHTSA considered the near multicollinearity of mass and footprint to be a major
issue in the 2010 report and voiced concern about inaccurately estimated regression
coefficients.66... Nevertheless, multicollinearity appears to have become less of a
problem in the 2012 and 2016 analyses. The "decile" analysis comparing fatality rates
of vehicles of different mass but nearly identical footprint (modified in 2012 in
response to peer-review comments to control for factors such as driver age and
gender) largely corroborates the main regression results. Ultimately, only three of the
27 core models of fatality risk by vehicle type indicate the potential presence of effects
of multicollinearity, with estimated effects of mass and footprint reduction greater
than two percent per 100-pound mass reduction and one-square-foot footprint
reduction, respectively: passenger cars and CUVs in first-event rollovers, and CUVs in
fixed-object collisions."
The second issue raised by the Alliance, regarding their contention that there is a disconnect
between the EPA Draft TAR safety analysis and the NHTSA 2016 VOLPE Report, speaks to
inconsistencies between NHTSA's 2016 Volpe Report (Table 3-7) and Table 8.7 of the Draft
TAR. However, Table 8.7 of the Draft TAR is a table authored by NHTSA, showing results from
the CAFE model. EPA does not agree that there is any implication for the EPA GHG analysis
presented in the Draft TAR or the Proposed Determination. It is possible that the Alliance meant
to point to Draft TAR Tables 8.3 and 8.4 which do show fatality impacts per 100 lb mass
reduction. In the Draft TAR, and again in the Proposed Determination, EPA used values
65 Puckett, S.M. and Kindelberger, J.C. (2016, June). Relationships between Fatality Risk, Mass, and Footprint in
Model Year 2003-2010 Passenger Cars and LTVs - Preliminary Report. (Docket No. NHTSA-2016-0068).
Washington, DC: National Highway Traffic Safety Administration.
66 Van Auken and Green also discussed the issue in their presentations at the NHTSA Workshop on Vehicle Mass-
Size-Safety in Washington, DC on February 25, 2011, http://www.nhtsa.gov/Laws+&+Regulations/CAFE+-
+Fuel+Economy/NHTSA+Workshop+on+Vehicle+Mass-Size-Safety.
147
-------
consistent with Draft TAR Table 8.4. However, that table does not present a value of 91 or 55 as
stated by the Alliance, so we cannot determine what is the inconsistency of concern. Importantly,
Draft TAR Table 8.4 shows results of NHTSA's 2016 Preliminary Report while Table 8.3 shows
results of the 2012 Final Report. EPA did not use results from the 2012 report.
3.7 Alternative Fuel (PEV) Infrastructure
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Although the Draft TAR projected that meeting the MY2025 standards will require only a
very small fraction of PEVs in the fleet, alternative fuel vehicles such as battery electric vehicles
(BEVs), plug-in hybrid electric vehicles (PHEVs), and fuel cell electric vehicles (FCEVs) are
likely to be an essential part of any future vehicle fleet intended to meet long term climate and air
quality goals. In addition, other alternative fuels such as ethanol (E85) and compressed natural
gas (CNG) have the potential to contribute to GHG emission reductions. Chapter 9 of the Draft
TAR provided an overview of alternative fuel vehicle infrastructure, including the status, costs,
and trends in PEV charging infrastructure and hydrogen fueling infrastructure, and examined the
challenges being addressed to scale up the infrastructure as advanced vehicle sales grow in
response to market demand and for compliance with the federal standards. Chapter 9 of the
Draft TAR concluded that infrastructure does not present a barrier for the small numbers of
alternative fuel vehicles that we expect manufacturers to choose to produce as a part of their
compliance with the MY2022-2025 GHG standards.
In public comments on the Draft TAR, several stakeholders discussed the conclusions of the
Draft TAR about the sufficiency of existing and expected infrastructure development. A number
of these comments, generally from the automotive manufacturing industry, focused on the
commenters' belief that a greater degree of infrastructure development would be needed because
they expect that more of these vehicles will be needed to meet the standards. However, as
discussed in Section III.C.2 and in Appendix B.3.2 of the Proposed Determination, we continue
to conclude that only a few percent of PEVs will be needed to meet the standards and thus we
also continue to conclude that current and expected expansion of electric charging and hydrogen
fueling infrastructure, as discussed in Chapter 9 of the Draft TAR, will be sufficient to supply
that segment of the automotive fleet. EPA also responded to comments on infrastructure in TSD
Chapter 2.2.4.4.5 at p. 2-97, and also provided additional discussion of charging infrastructure
developments in TSD Chapter 2.2.4.3.2 at p. 2-70.
Summary of Comments and Responses on the Proposed Determination
Nissan commented generally that public PEV charging infrastructure is not growing quickly
enough to support consumer needs and expectations. Mercedes-Benz provided more detailed
comments on charging infrastructure, arguing that the longer charging time required for longer-
range PEVs makes 110V (Level 1) charging impractical; that the cost of upgrading electrical
service to accommodate home charging at higher rates must be considered; and that public
charging in "cities where land is at a premium can also mean higher charging rates outside of the
home for consumers." Mercedes-Benz also suggested that EPA had not addressed private
infrastructure concerns in the Proposed Determination.
The comment from Nissan concerning general adequacy of growth in charging infrastructure,
and the portions of the Mercedes-Benz comments concerning Level 1 charging, private
148
-------
infrastructure concerns, and cost of upgrading electrical service, mirror the concerns outlined in
many of the public comments received on the Draft TAR. EPA addressed those comments in
Section B.3.2 of the Proposed Determination Appendix at p. A-98 and in TSD Chapter 2.2.4.4.5
at p. 2-97, and also provided additional discussion of charging infrastructure developments in
TSD Chapter 2.2.4.3.2 at p. 2-70. Specifically, EPA pointed out several examples of charging
infrastructure initiatives that suggest public charging infrastructure will continue to grow; that
the cost of installing home charging capability was in fact included in the projected cost of BEVs
and PHEVs; and that EPA did not assume that PEV users would rely solely on Level 1 charging
but rather on a mix of Level 1 and Level 2 (weighted heavily toward Level 2) depending on the
specific PEV type. EPA disagrees that these responses did not address private infrastructure
concerns of the sort raised by Mercedes-Benz, in that they referred to the relative capabilities of
Level 1 and Level 2 and the fact that EPA had accounted for the costs necessary for installing
Level 1 or Level 2 charging service capability and equipment. EPA also notes, as mentioned in
the TSD discussion of BEVs in Chapter 2.2.4.4.5 at p. 2-70, that longer-range electric vehicles
do not render Level 1 charging useless, in that they are equally capable as shorter-range vehicles
of having a given daily mileage replenished on a nightly basis. That is, while a fully depleted
battery of a longer-range vehicle would in fact require a longer time to restore to full charge than
a shorter-range vehicle, the larger capacity of the battery means that, for a given average daily
mileage, depleting the battery completely will rarely happen as long as charge is replaced daily.
Stated another way, if the primary requirement of routine daily charging for both longer- and
shorter-range PEVs is to achieve a state of charge suitable to meet the driving needs of the next
day, then the total capacity of the battery does not inherently affect the average necessary charge
time to reach that state of charge. This is not to suggest that Level 1 charging is practical in all
situations (such as those where daily mileage is unusually large, or where owners do not wish to
charge daily), but only that the usefulness of Level 1 charging in many situations is not inversely
related to the range of the vehicle. Also as stated in the responses in the TSD and above, EPA
assumed that home installations of charging capability would be significantly weighted toward
Level 2 charging, and accounted for the cost of installation accordingly.
With regard to the Mercedes-Benz comment that public charging in "cities where land is at a
premium can also mean higher charging rates outside of the home for consumers," EPA notes
that it is commonly understood that real estate costs often have an influence on the retail pricing
of many goods and services, and that real estate costs are often higher in densely populated areas.
EPA is not aware of data supporting the suggestion that retail public charging rates for
consumers in cities should experience this effect more than other similar services, nor that retail
public charging rates being paid by consumers are systematically and significantly higher in
cities, even as much of the currently available public charging infrastructure in the U.S. is
located in cities and urbanized areas of varying density and real estate value. Much of this
infrastructure is not accessed through a retail payment mechanism, and is it not yet clear what
proportion of future infrastructure will require retail payment, nor which payment models (e.g.
flat rate per use, metered, subscription based, etc.) will be common. The existence or magnitude
of any supposed differential in charging rates in cities would therefore be very difficult to
reliably assess at this early stage of retail public charging infrastructure, and EPA believes that in
the context of this analysis any such future charging cost differential, if it were possible to
identify and measure, would amount to only a small portion of projected PEV costs and would
not affect the overall conclusions of the analysis.
149
-------
3.8 Standards Design
Summary of Comments on the Draft TAR addressed in the Proposed Determination
In the design of the MY2012-2025 GHG standards, EPA carefully considered the impact the
standards can have on vehicle utility and consumer choice such that the automotive companies
have the ability to maintain vehicle utility and consumer choice while complying with the
standards. EPA decided to use vehicle "footprint" as the attribute to determine the GHG
standards for a given automotive manufacturer's fleet (the standard being the production-
weighted average of the footprint-based targets for each vehicle produced). The light-duty
vehicle GHG standards are curves based on the footprint attribute (Section I of the Proposed
Determination shows a graphical depiction of the footprint curves). There are separate passenger
car footprint-based standards and light-truck footprint-based standards.
EPA received a variety of comments regarding use of the footprint attribute. Several
commenters stressed the importance of the footprint-based standards in ensuring consumer
choice and encouraging emissions reductions across vehicles of all sizes. Several commenters
expressed concern regarding the footprint standards, asserting that vehicle footprints are
increasing over time. Related to the comments regarding footprint, EPA received comments
supporting a backstop standard. EPA also received comment on the current light truck
definition. Section B.3.3 of the Appendix to the Proposed Determination provided an overview
of these comments as well as EPA's response. Because the Proposed Determination was that the
standards remain appropriate and the footprint attribute is a key feature of those standards, EPA
did not propose to change any aspect of the design of the standards.
Summary of Comments on the Proposed Determination
EPA received comments similar to comments received on the Draft TAR supporting the use
of the footprint-based attribute. The UAW commented that well-constructed regulations can
protect the environment while simultaneously supporting existing jobs and creating jobs in new
advanced technology sectors of the economy. The UAW further commented that the basic
design of the National Program was carefully constructed by a wide array of stakeholders and
should be kept intact and not dramatically altered. Consumers Union commented that footprint-
based standards encourage automakers to design and sell vehicles that have better fuel economy
across all vehicle classes. Consumer Federation of America commented that the attribute-based
approach ensures that the standards do not require radical changes in the available products or
the product features that will be available to consumers. ICCT commented that the footprint
standards appropriately accommodate the changing fleet mix due to market shifts. The
BlueGreen Alliance also commented that the standards are "smartly structured," adjusting with
changes in the mix of cars and trucks while ensuring that no matter how the market shifts, each
size of vehicles makes gradual but steady progress.
EPA also received comments similar to those on the Draft TAR raising concerns with the
footprint attribute. The Institute for Energy Research commented that EPA has not adequately
addressed concerns raised by Whitefoot and Skerlos in a 2012 peer reviewed publication
regarding the potential incentive for manufacturers to upsize vehicles in response to footprint-
based standards. Carnegie Mellon also referenced the Whitefoot and Skerlos paper and
recommended that EPA study recent pickup truck footprint increases as well as footprint trends
in other vehicle segments to determine if the footprint-based standards create an incentive for
150
-------
automakers to increase vehicle size. Consumers Union also raised concerns regarding "footprint
creep" where manufacturers would increase the footprint of vehicles in order to lower their fleet
standard, and recommended EPA consider backstop standards to complement the footprint-based
standards. Center for Biological Diversity also commented that a backstop standard is needed to
prevent the loss of emissions reductions from shifts in fleet mix and footprint creep. EPA also
received an anonymous comment that the footprint standards would restrict vehicle offerings,
reducing cabin sizes, and require consumers to purchase larger vehicles than they would
otherwise.
Mercedes-Benz commented that the footprint standards cannot compensate for the disparity
between large fleets with diverse product offerings and fleets that sell traditionally in the luxury
market. Mercedes recommends EPA take a separate action outside of the MTE process to extend
the Temporary Lead Time Allowance Alternative Standards (TLAAS) provisions established in
the MY2012-2016 rule. Mercedes believes EPA should include more flexibilities for
manufacturers who traditionally sell into the luxury market with modest volumes over which to
spread their compliance obligations.
EPA received comments from BMW regarding the relative stringency of the car and light
truck standards. BMW commented that the stringency gap between cars and light trucks should
be further reduced based on physics and with a view toward ever increasing overlap between the
two segments. BMW further commented that not every automaker can compensate for a
potential compliance shortfall from small passenger cars with a portion of very large light trucks
with a comparable relaxed standard.
Response to Comments on the Proposed Determination
Comments regarding the footprint-based standards providing an incentive for manufacturers
to increase vehicle footprints, or "footprint creep," and the need for a backstop standard are very
similar to comments received on the Draft TAR and addressed in Section B.3.3 of the Appendix
to the Proposed Determination. EPA understands the concerns of commenters that the program
is now projected to deliver a somewhat higher numerical fleetwide CO2 target than originally
estimated. However, EPA continues to believe that the program is operating as designed, by
accommodating shifts in consumer choice in the fleet while requiring increasingly more stringent
GHG emission reductions across all vehicle types. EPA disagrees with the Institute for Energy
Research that EPA has not adequately responded to concerns regarding footprint. EPA
previously addressed the Whitefoot and Skerlos study as part of the 2012 rule.67 EPA noted that
the authors made several assumptions for the study and changes to any of the assumptions could
yield different analytic results. Underlining the potential uncertainty, the authors obtained a
wide range of results with their analysis. EPA has monitored trends in footprint and has found
that average footprint has remained relatively flat since the standards were first established, as
discussed in Section B.3.3 of the Appendix to the Proposed Determination. EPA is not aware of
any evidence that the standards structure is motivating the shift from cars to trucks, beyond the
effect of market forces such as lower gasoline prices. EPA also does not agree with the
anonymous comment that the program is restricting vehicle choice and forcing consumers to
purchase larger vehicles than they would otherwise. Nothing in the design of the GHG program
restricts manufacturers' ability to offer a full range of vehicle sizes and features, including cabin
67 77 FR 62962, October 15, 2012.
151
-------
size. A key feature of the footprint attribute approach is that it provides an incentive for
manufacturers to reduce GHGs from vehicles of all sizes without reducing consumer choice.
Mercedes provided comments that EPA should include less stringent standards for smaller
volume manufacturers similar to the less stringent standards provided by EPA in the MY2012-
2016 rule under Temporary Lead Time Allowance Alternative Standards (TLAAS), in order to
provide smaller volume manufacturers with more lead time. EPA does not believe less stringent
standards are appropriate or necessary. The TLAAS program was established to provide
additional lead time on a temporary basis to lower volume manufacturers in the initial years of
the GHG program.68 The MY2012-2016 standards were finalized in 2010 and additional lead-
time was found to be appropriate for lower volume manufacturers. TLAAS was designed to
address two situations where EPA projected that more lead time was needed for the initial phase
of the GHG program. One situation involved manufacturers who had traditionally paid CAFE
fines instead of complying with the CAFE fleet average, and as a result at least part of their
vehicle production had significantly higher CO2 than the industry average. The other situation
involved manufacturers who had a limited line of vehicles and are therefore unable to average
emissions performance across a full line of production. EPA provided additional lead-time in the
initial years for such manufacturers to upgrade their vehicles and meet the standards. EPA did
not extend the TLAAS program in the MY2017-2025 rule, which went through a full notice and
comment period, because those standards provided significantly more lead time than the first rule
and the TLAAS was designed to be temporary.69 This is especially true for the MY2022-2025
standards, where significant lead time has been provided since the 2012 rule. EPA has
determined that the MY2022-2025 standards currently in place remain appropriate under section
202 of the Clean Air Act considering available lead time without the need for less stringent
standards for lower volume manufacturers.
Regarding BMW comments concerning relative stringency between car and truck standards,
EPA does not believe that the BMW comment is supported by the analysis conducted by EPA in
support of the determination. As discussed in the Final Determination document, the
Administrator has concluded that the record does not support a conclusion that the MY2022-
2025 standards should be revised to make them less stringent. The Administrator did consider
whether it would be appropriate to propose to amend the standards to increase their stringency.
In her view, the current record, including the current state of technology and the pace of
technology development and implementation, could support a proposal, and potentially an
ultimate decision, to adopt more stringent standards for MY2022-2025. However, she also
recognizes that regulatory certainty and consequent stability is important, and that it is important
not to disrupt the industry's long-term planning. Long lead time is needed to accommodate
significant redesigns. The Administrator consequently has concluded that it is appropriate to
provide the full measure of lead time for the MY2022-2025 standards, rather than adopting (or,
more precisely, proposing to adopt) new, more stringent standards with a shorter lead time.
3.9 Credits, Incentives, and Flexibilities
Summary of Comments on the Draft TAR addressed in the Proposed Determination
68 75 FR 25414, May 7, 2010.
69 77 FR 62795, October 15, 2012.
152
-------
The National Program provides a wide range of optional flexibilities to allow manufacturers
to maintain consumer choice, spur technology development, provide compliance flexibility, and
reduce compliance costs, while achieving significant GHG reductions. Chapter 11 of the Draft
TAR provided an overview of these provisions which include averaging, banking, and trading of
credits, air conditioning system credits, off-cycle technology credits, and advanced technology
vehicle incentives including incentives for large pickups using advanced technologies.
EPA received comments on various aspects of the credits programs. Air conditioning system
credits and related comments are discussed in Chapter 2.2.9 of the TSD (and are discussed
further in Chapter 2.5.16 of this Response to Comments (RTC) document). EPA received a
variety of comments supporting both procedural changes to expedite the off-cycle credit
approval process, and substantive changes to increase or remove caps on the amount of credits
provided without need for prior approval (the so-called menu credits in section 86.1869-12) and
to expand the eligibility criteria for receiving off-cycle credits. Other commenters expressed
concerns regarding possible changes to the off-cycle credits program. Section B.3.4.1 of the
Appendix to the Proposed Determination provided an overview and EPA's response to
comments received regarding off-cycle credits. Our conclusion at that time was that the
standards for MYs 2022-2025 remain appropriate and therefore that no rulemaking to amend the
standards was necessary. Since the Proposed Determination would leave both the standards and
all of the regulatory provisions supporting the standards unaltered, EPA did not propose any
changes to the off-cycle credit provisions. Those provisions are a part of the standards, and EPA
explicitly considered and quantified off-cycle credit usage in evaluating available compliance
paths. See also Chapters 2.5.16.1 and 2.8 of this RTC document. Consideration of off-cycle
credits under the current standards thus was a direct part of EPA's analysis in the Proposed
Determination that the MY2022-2025 standards remain appropriate.
As discussed in Section B.3.4.2 of the Appendix, EPA also received comments regarding
incentives for advanced technology vehicles including BEVs, PHEVs, and FCEVs. In response
to these comments, EPA noted that it was proposing a determination that would leave the
MY2022-2025 standards unchanged based on the existing regulatory program and therefore did
not propose any changes to these programs. Put another way, EPA proposed a determination
that took into account the existing regulatory provisions on incentives for advanced technology
vehicles (just as it proposed to find the standards appropriate considering the current regulations
regarding off-cycle credits). In addition to the request for comment on EPA's proposed
determination, EPA requested comments on the need to continue incentives for advanced
technology vehicles, including for the MY2022-2025 time frame.
As discussed in Section B.3.4.3 of the Proposed Determination Appendix, EPA also received
comments regarding incentives for flexible fuel vehicles and natural gas vehicles as well as
credits for investments in alternative fuel infrastructure. EPA responded that, as with off-cycle
credits, the Proposed Determination was that the MY2022-2025 standards remain appropriate
and that the Proposed Determination took into account the standards' flexibilities as they now
stand, including the incentives for flexible fuel vehicles and the advanced technology incentives.
Summary of Comments on the Proposed Determination
EPA received several comments on the topic of off-cycle credits. Many of the comments
were essentially the same as comments provided on the Draft TAR. Mercedes-Benz provided
several suggestions it believes would improve the program including removing the menu credit
153
-------
cap, adding technologies to the menu via guidance, allowing suppliers to participate in the
program, allowing all manufacturers to receive the same credits approved for a manufacturer
through the approval process, and streamlining the approval process for unique technologies.
Mercedes-Benz commented, also, that credits should be provided for congestion mitigation and
crash avoidance and that the agency should work with interested parties to devise an acceptable
methodology for determining these credits. Mercedes requested that EPA issue guidance or
initiate a rulemaking to improve the off-cycle credits program. BMW commented that EPA
should expand the menu of technologies, include credits for European off-cycle technologies,
and include credits for reduced traffic, faster exchange of older vehicles for new, more fuel
efficient vehicles, car sharing, and establishing charging infrastructure for electric vehicles. The
Alliance asserted that EPA dismissed their Draft TAR comments regarding off-cycle credits, in
which they suggested expanding the credit menu, removing the credit caps, and providing
additional credits for 48V mild hybrids.
The Motor and Equipment Manufacturers Association (MEMA) commented that it disagrees
with EPA's Proposed Determination finding that the standards remain appropriate with no
changes to the off-cycle credits program, asserting that "further development of and
improvement to the off-cycle credits program is necessary to provide much needed flexibilities
and allow industry to stay on track for MYs 2022-2025." MEMA believes that EPA should wait
to consider MY2015-2017 usage of off-cycle credits before determining no changes are needed
to meet the MYs 2022-2025 standards. MEMA recommended that EPA consider how off-cycle
credits can be optimized even if it is outside of the MTE process, including expanding the
predefined menu, removing the menu credit cap, streamlining the process and allowing suppliers
to directly petition for approval of new off-cycle technologies.
Denso reiterated their comments supporting changes to the off-cycle credits program to
streamline the process and encouraging EPA to develop a working group to facilitate
standardized test methods. Denso also encouraged EPA to consider future updates to the off-
cycle credits program, based on the relevancy of advanced technologies in future model years.
American Iron and Steel Institute (AISI) commented that EPA should include GHG emissions
from the material production phase to ensure that the program results in the greatest GHG
reductions. AISI commented that depending on the mix of materials used in the vehicle,
particularly in the context of vehicle light-weighting, increases in material production emissions
may be higher in magnitude than tailpipe savings. AISI commented that developing an off-cycle
credit would be an option that could help account for differences in materials production phase
emissions.
EPA also received comments, similar to Draft TAR comments, expressing concern regarding
any expansion of the off-cycle credits program. ACEEE commented that EPA's analysis
demonstrates that commenter claims that manufacturers will need large quantities of off-cycle
credits to comply with the standards is incorrect. ACEEE reiterated their Draft TAR comments
that the changes to the off-cycle credits program suggested by the automotive industry could
undermine the credibility of the program and the effectiveness of the standards. The Center for
Biological Diversity commented that the program must avoid double counting, the credits
awarded must be demonstrated to translate to actual real-world on-road improvements, and that
credits should not be approved until such studies have been completed.
154
-------
In addition to EPA's request for comments on the Proposed Determination, EPA requested
comments on the need for continued incentives for advanced technologies such as plug-in hybrid
electric vehicles, battery electric vehicles, and fuel cell vehicles. EPA received comments
supporting extending advanced technology incentives similar to comments received on the Draft
TAR, including from the Alliance, Global Automakers, Ford, Mercedes-Benz, BMW, Nissan,
Toyota, Center for Biological Diversity, and Edison Electric Institute. The commenters
supported extending the incentive multipliers and the continued use of 0 g/mile tailpipe
compliance value for electric operation with no accounting of upstream emissions to continue to
encourage advanced technology vehicles. Ford also commented in support of a credit to
incentivize HEVs, noting that sales of HEVs have been stagnant or declining since 2013.
Manufacturers reiterated comments that they should not be held responsible for upstream
emissions since they have no control over those emissions and that upstream emissions are
addressed under EPA's Clean Power Plan. Toyota also commented that EPA did not respond to
the Alliance, Toyota and other comments received on the Draft TAR regarding expanding
advance technology incentives for full size pickups. Mercedes requested that EPA initiate a
rulemaking to extend/expand incentives for advanced technology vehicles.
EPA also received comments opposing extending or expanding the advanced technology
incentives. Commenters against expanding the incentives included Manufacturers of Emission
Controls Association, Carnegie Mellon University, ACEEE, and UCS. The commenters believe
that additional incentives are not needed to encourage advanced technologies, with some
commenters noting that EPA's analysis shows the standards are feasible with low penetrations of
electric vehicles. Commenters were also concerned that expanding the incentives would reduce
the overall benefits of the program, and that advanced technology vehicles should compete in the
marketplace on the same basis as other technologies. ACEEE commented that the Clean Power
Plan is currently being litigated and therefore they do not find the Alliance's rationale for
eliminating the accounting of upstream emissions for electric operation persuasive at this time.
One commenter believed that the CAFE credit for BEVs and PHEVs lack a compelling
thermodynamic basis in terms of equitable system-level energy substitutions between oil and
electricity.
EPA received comments on the Proposed Determination supporting incentives for flexible
fuel vehicles and natural gas vehicles as well as credits for investments in alternative fuel
infrastructure similar to comments received on the Draft TAR.
Mercedes-Benz commented that EPA should allow credits to be transferred to the light-duty
fleet from heavy-duty 2B/3 class vehicles in order to provide manufacturers with additional
flexibility. Mercedes also commented that EPA should consider the potential lack of future
availability of a credits market and traded credits from one manufacturer to another.
Response to Comments on the Proposed Determination
Most of the comments on off-cycle credits were essentially the same as comments submitted
on the Draft TAR and addressed in Section B.3.4.1 of the Appendix to the Proposed
Determination. MEMA commented that the MY2022-2025 standards are not appropriate
without changes to the off-cycle credits program. EPA disagrees with this comment and is
finalizing its finding that standards remain appropriate with the off-cycle credits provisions
currently in place. In response to AISI comments, expanding the program to include credits for
potential upstream emissions reductions associated with material production would represent a
155
-------
significant expansion of the credits provisions and EPA believes it is not appropriate or
necessary as part of the MTE. EPA's analysis incorporates a modest level of off-cycle credits
that are readily available with the current off-cycle provisions including the off-cycle credits
menu. The analysis demonstrates that the standards are feasible under the Clean Air Act without
an expansion of the off-cycle credits program. Regarding the Alliance comments that EPA
dismissed its Draft TAR comments on the off-cycle credits program, EPA notes that although
EPA did not propose changes to the program in response to the comments, the comments were
addressed in Section B.3.4.1 of the Appendix to the Proposed Determination.
As discussed in Section B.3.5 of the Appendix to the Proposed Determination, the Alliance
and Global Automakers also raised issues regarding the off-cycle credits program in its June
2016 petition to EPA and NHTSA.70 EPA intends to work with the Petitioners and other
stakeholders in the future as we carefully consider the requests made in the June 2016 petition.
EPA will be taking a separate action to respond to this petition and none of the issues raised in
the petition change EPA's assessment of the appropriateness of the MY2022-2025 standards.
EPA is making a determination that the MY2022-2025 standards are still appropriate, based on
the existing regulations, including the credit provisions raised in the auto petition.
Comments on advanced technology incentives submitted on the Proposed Determination
largely overlap with the comments received on the Draft TAR, addressed in the Proposed
Determination at Section B.3.4.2 of the Appendix. With regard to Toyota's claim that EPA did
not address comments on incentives for full size pickups submitted on the Draft TAR, EPA did
consider the comments and the response to the comments is also provided in Proposed
Determination Section B.3.4.2 of the Appendix. The analyses supporting EPA's determination
show that the standards remain appropriate without changes to the incentives multiplier and large
pickup provisions. Further, EPA does not believe that expanding multipliers to additional
technologies such as hybrids, as recommended by Ford, is warranted. EPA's analysis is based on
cost-effective technologies available to meet the standards with no reliance on the multiplier or
large pickup incentives and therefore the agency finds no reason within the scope of the MTE to
revisit these provisions. Regarding the comment concerning CAFE credits for EVs and PHEVs,
EPA notes that for the GHG program, the current incentives are in place temporarily to promote
the initial commercialization of advanced technology vehicles and EPA recognizes that the
incentives, to the extent they are used by manufacturers result in a small loss of emissions
benefits.71
When establishing the standards in the 2012 rule, EPA included incentives for advanced
technologies to promote the commercialization of technologies that have the potential to
transform the light-duty vehicle sector by achieving zero or near-zero GHG emissions and oil
consumption in the longer term, but which face major near-term market barriers. As noted
above, providing temporary regulatory incentives for certain advanced technologies will
decrease the overall GHG emissions reductions associated with the program in the near term.
However, in setting the 2017-2025 standards, EPA believed it was worthwhile to forego modest
additional emissions reductions in the near term in order to lay the foundation for the potential
for much larger "game-changing" GHG emissions reductions in the longer term. EPA also
70 "Petition for Direct Final Rule with Regard to Various Aspects of the Corporate Average Fuel Economy Program
and the Greenhouse Gas Program," Auto Alliance and Global Automakers, June 20, 2016.
71 77 FR 628111-62813, October 15, 2012.
156
-------
believed that temporary regulatory incentives may help bring some technologies to market more
quickly than in the absence of incentives. See 77 FR 62811 et seq.
The Administrator notes that her determination, based on the record before her, is that the
MY2022-2025 standards currently in effect are feasible (evaluated against the criteria established
in the 2012 rule) and appropriate under Clean Air Act section 202, and do not need to be revised.
This conclusion, however, neither precludes nor prejudices the possibility of a future rulemaking
to provide additional incentives for very clean technologies or flexibilities that could assist
manufacturers with longer term planning without compromising the effectiveness of the current
program. EPA is always open to further dialogue with the manufacturers, NHTSA, CARB and
other stakeholders to explore and consider the suggestions made to date and any other ideas that
could enhance firms' incentives to move forward with and to help promote the market for very
advanced technologies, such as battery electric vehicles (BEVs), plug-in hybrid electric vehicles
(PHEVs), and fuel cell vehicles (FCEVs).
Comments on incentives for FFVs and natural gas-fueled vehicles were also addressed in the
Proposed Determination at Section B.3.4.3 of the Appendix. EPA continues to believe that the
treatment of these vehicles established in the 2012 rule remains appropriate and no changes to
the provisions are needed as part of the MTE. The MY2022-2025 standards remain appropriate
with the credit provisions currently in place. EPA does not agree that the MTE must include a
rulemaking to consider further incentives for these vehicles under the MTE regulations at 40
CFR 86.1818-12(h)(l). EPA also notes that Adsorbed Natural Gas Products (ANGP) provides
no data on consumer refueling for their natural gas fueling system to support their claim that the
ANGP system would result in at home refueling on the same level as that of PHEVs. ANGP
also does not consider possible methane upstream emissions from natural gas production in their
comment regarding their views on the equitable credit levels for PHEVs and vehicles equipped
with the ANGP natural gas fueling system.
Regarding Mercedes Benz comments that EPA should allow credit to be transferred from
heavy-duty 2b/3 class vehicles to light-duty vehicles, EPA believes allowing such credit transfers
would raise competitiveness issues since only a small portion of the light-duty vehicle
manufacturers produce vehicles in the heavy-duty category and EPA believes it is important to
maintain a level playing field for light-duty manufacturers not participating in the heavy-duty
market. Also, the heavy-duty 2b/3 standards are based on a different attribute than the standards
for light-duty vehicles. EPA is determining the MY2022-2025 standards remain appropriate
without allowing credit transfers between heavy-duty and light-duty vehicles. Regarding
Mercedes' comment that EPA should consider the potential lack of available credits through
credit trading between manufacturers, EPA notes that its analysis does not incorporate credit
trading and instead projects that manufacturers are able to comply with the standards without
credit trades. Any credit trading that occurs in the future would presumably lower compliance
costs.
3.10 Harmonization
Summary of Comments on the Draft TAR addressed in the Proposed Determination
Section B.3.5 of the Appendix to the Proposed Determination considers comments regarding
harmonization among EPA, NHTSA, and CARB programs. Separate from the Midterm
Evaluation process, on June 20, 2016, the Alliance and Global Automakers submitted a petition
157
-------
asking EPA and NHTSA to make several regulatory changes they believe would better
harmonize the programs.72 The Alliance and Global Automakers raised several of the issues
again in their comments on the Draft TAR. Commenters also provided comments regarding the
CARB regulatory process and ZEV program. Section B.3.5 of the Appendix provides an
overview of the comments and EPA's response. EPA noted that while EPA will be taking a
separate action to respond to this petition, none of the issues raised in the petition would change
EPA's assessment of the appropriateness of the MY2022-2025 standards. EPA made a proposed
determination that the MY2022-2025 standards are still appropriate, based on the existing
regulations, including the credit provisions raised in the auto petition. EPA also noted that it
intends to work with the Petitioners and other stakeholders in the future as we carefully consider
the requests made in the June 2016 petition. We also note that NHTSA has now partially granted
the portions of the petition addressed to NHTSA, and intends to conduct rulemaking addressing
the issues exclusive to NHTSA raised in the petition.73
Summary of Comments on the Proposed Determination
Several commenters raised concerns regarding how the timing of EPA's Proposed
Determination may potentially impact the harmonization of EPA's and NHTSA's programs
under the National Program. These comments are addressed in Chapter 1 of this RTC document.
Several commenters stressed their view of the importance of program harmonization and
reiterated concerns regarding harmonization in the current National Program raised in comments
on the Draft TAR. The Alliance and Global Automakers reiterated harmonization issues raised
in its petition to the agencies. EPA responded to comments regarding the petition in Section
B.3.5 of the Appendix to the Proposed Determination, noting that it will consider the petition in a
separate action. Global Automakers commented that this response is insufficient and the issues
must be addressed in order to ensure the current program's success through MY2016 which
ultimately impacts the bottom line beyond 2016. GM also commented that EPA did not
adequately respond to Draft TAR comments regarding harmonization in the Proposed
Determination. Toyota commented that one National Program has not been achieved and they
expressed disappointment that EPA and NHTSA are not addressing harmonization issues
immediately, but that Toyota looks forward to working with the agencies to resolve the matter.
The UAW, MEMA, and American Coalition for Ethanol also commented that harmonization is
important and that EPA should work with NHTSA and CARB to address program
inconsistencies.
Global Automakers commented specifically that EPA did not address their concern that some
manufacturers may not fully use GHG air conditioning leakage credits in certain vehicle models
due to the manufacturer's product plans and vehicle redesign cycles, thus making the EPA
standards more stringent than the NHTSA standards for those manufacturers. Global Automakers
suggest EPA adjust the air conditioning credits upward as a way of addressing this issue.
ICCT commented, "EPA has provided ample auto industry flexibilities through technology
credits, emission trading, smaller volume company provisions, and footprint indexed standards to
72 "Petition for Direct Final Rule with Regard to Various Aspects of the Corporate Average Fuel Economy Program
and the Greenhouse Gas Program," Auto Alliance and Global Automakers, June 20, 2016.
73 81 FR 95553, December 28, 2016. NHTSA did not grant the petitioners' request that it issue a direct final rule
and will instead address changes requested in the Petition in the course of a rulemaking proceeding.
158
-------
accommodate fleet shifts. These EPA provisions greatly assist automobile industry compliance.
Based on the well-designed EPA flexibilities, any further improvement toward a harmonized one
national program would best be addressed with adjustments in the Corporate Average Fuel
Economy, matching NHTSA's program with EPA's improved manufacturer flexibilities."
Response to Comments on the Proposed Determination
EPA responded to comments regarding program harmonization and the manufacturer's
petition in Section B.3.5 of the Appendix to the Proposed Determination, noting that EPA will
consider the petition in a separate action. As discussed in EPA's previous response, the agencies
have worked to establish a National Program subject to the differences in statutory authorities.
The differences in certain aspects of the GHG and CAFE programs existed when the MY2022-
2025 standards were first established and do not lead EPA to find that the GHG standards for
MYs 2022-2025 are no longer appropriate. EPA is making a final determination that the
MY2022-2025 standards remain appropriate based on existing regulations. While EPA will be
taking a separate action to respond to the petition, none of the issues raised in the petition would
change EPA's assessment of the appropriateness of the MY2022-2025 standards. NHTSA has
also indicated that it will consider the auto manufacturers' petition.
EPA does not agree with the Global Automakers comment that the programs are not
harmonized due to the situation that some manufacturers may not maximize air conditioning
refrigerant leakage credits. EPA does not believe this is an issue for the MTE because
manufacturers will have had significant lead time to incorporate air conditioning refrigerant
improvements including alternative refrigerants into their vehicles by MYs 2022-2025. Also, as
discussed in Chapter 2.2.9 of the Technical Support Document, under EPA's Significant New
Alternatives Policy (SNAP) program, manufacturers will be required to use lower-GWP
refrigerants beginning in MY2021 for which they will generate full refrigerant credits under
EPA's GHG program. For these reasons, EPA expects manufacturers will utilize air
conditioning leakage credits and has included their use in its assessment as part of the pathway
manufacturers may take to comply with the standards. This issue was also addressed in the 2012
rule where the agencies discussed the possibility of the scenario provided by Global Automakers.
The agencies similarly responded there that the comment "reflects a misunderstanding of the
agencies' purpose. The agencies have sought to craft harmonized standards such that
manufacturers may build a single fleet of vehicles to meet both agencies' requirements. That is
the case for these final standards. Manufacturers will have to plan their compliance strategies
considering both the NHTSA standards and the EPA standards and assure that they are in
compliance with both, but they can still build a single fleet of vehicles to accomplish that
goal."74 EPA also notes that an increase in the credits awarded for air conditioning leakage
improvements as suggested by Global Automakers, to the extent that such credits were greater
than the actual real-world emission reductions achieved through the air conditioning system
improvements, would result in an unwarranted decrease in overall GHG reductions from the
program.
74 77 FR 63054-63055, October 15, 2012.
159
-------
Chapter 4: Climate Science and the Further GHG Reductions Beyond 2025
4.1 Climate Science
Summary of Comments on the Draft TAR addressed in the Proposed Determination
EPA received no comments on the Draft TAR discussion of climate science.
Summary of Comments on the Proposed Determination
In the Proposed Determination, Section I.B, EPA presented an overview of climate change
science, as laid out in the climate change assessments from the National Academies, the U.S.
Global Change Research Program (USGCRP), and the Intergovernmental Panel on Climate
Change (IPCC). The Proposed Determination summarized the impacts to human health, to
ecosystems, and to physical systems in the United States and around the world, from heat waves
to sea level rise to disruptions of food security. Impacts to vulnerable populations such as
children, older Americans, persons with disabilities, those with low incomes, indigenous peoples,
and persons with preexisting or chronic conditions were also highlighted. The most recent
assessments noted by the Proposed Determination have confirmed and further expanded the
science that supported the 2009 Endangerment and Cause or Contribute Findings for Greenhouse
Gases Under section 202(a) of the Clean Air Act; Final Rule (74 FR 66496), as well as the more
recent 2016 Finding That Greenhouse Gas Emissions from Aircraft Cause or Contribute to Air
Pollution That May Reasonably Be Anticipated to Endanger Public Health and Welfare (81 FR
54421). As described in the Proposed Determination, the climate system continues to change: in
2015, CO2 concentrations grew by more than 2 parts per million, reaching an annual average of
401 ppm, sea level continued to rise at 3.3 mm/year since the satellite record started in 1993,
Arctic sea ice continues to decline, and glaciers continue to melt. 2015 was the warmest year in
the surface temperature record going back to 1880, surpassing the previous record set in 2014,
and available data show 2016 will exceed 2015.75
EPA received nine comments on the Proposed Determination that touched on climate science
issues. Of these comments, eight supported strong standards due to the evidence regarding the
impacts of climate change, and considering the large share of national emissions from the vehicle
sector. These commenters cite the strong evidence that climate change is real and urgent, the
need to protect the nation's children and grandchildren, and the grave threats to public health.
The commenters cite the recent USGCRP assessment on the "Impacts of Climate Change on
Human Health in the United States" as well as publications by the Asthma and Allergy
Foundation, the American Public Health Association, and the American Thoracic Society. The
commenters also highlight populations who are particularly vulnerable, including children, older
adults, Americans with chronic diseases, low income communities, outdoor workers, and Native
American tribal communities. One commenter notes that Millennials will be the primary
purchasers of the vehicles addressed by recent vehicle regulations, and that this age cohort will
be impacted by climate change and supports cutting greenhouse gases. Another commenter
discusses the risks to the economy, and therefore to investors, from unabated climate change.
In contrast, one commenter argues that the estimated reduction resulting from the MY2022-
2025 standards in 2025 would be 0.6 percent of national emissions, projected temperature
75 NOAA (2016): http://www.noaa.gov/news/november-2016-ranks-as-5th-warmest-on-record-for-globe
160
-------
reductions in the year 2100 would be less than two hundredths of a degree Celsius, and this
magnitude of benefits is "trivial" and "paltry" and does not justify the potential costs of the
regulation.
Response to Comments on the Proposed Determination
EPA has carefully considered all the comments regarding the science. The major assessments
demonstrate the continued and, for certain outcomes, increased certainty and likelihood that
GHGs impact health and welfare now and in the future. It continues to be EPA's view that the
scientific assessments of the IPCC, USGCRP, and the National Research Council (NRC)
represent the best reference materials for determining the general state of knowledge on the
scientific and technical issues before the agency in making an endangerment decision. No other
source of information provides such a comprehensive and in-depth analysis across such a large
body of scientific studies, adheres to such a high and exacting standard of peer review, and
synthesizes the resulting consensus view of a large body of scientific experts across the world.
These assessments draw synthesis conclusions across thousands of individual peer-reviewed
studies that appear in scientific journals, and the reports themselves undergo additional peer
review. Thus, the assessments reflect extremely high quality, rigorous work that has gone
through an exacting standard of peer review. This provides assurance that the Administrator is
basing her judgment on the best available, well-vetted science that reflects the consensus of the
climate science research community. For these reasons, EPA places primary and significant
weight on these assessment reports in reviewing the state of climate science.
These assessments support those claims of commenters who stated that the climate change
problem is real and urgent, with implications for public health and many vulnerable populations.
With regards to the commenter who claimed that the benefits of the rule are "trivial" and
"paltry," EPA finds that these claims are without merit. First, the commenter highlighted the
emissions reductions in 2025 due to the impacts of the rule on MY2022-2025 vehicles. Due to
the rate of vehicle turnover, only a relatively small fraction of the vehicles on the road in 2025
would be MY2022 or later. As the commenter noted, "EPA admittedly projects larger emissions
savings decades into the new standards."
Second, with regard to the temperature impacts, EPA previously responded to similar
comments in 77 FR at 62898 (from the 2012 final rule preamble). We repeat that response here:
IER and a number of private citizens asserted that the reductions in temperature
and other climate factors are too small to be meaningful. However, as has been
stated, no one rule will prevent climate change by itself. As stated in the
Endangerment and Cause or Contribute Findings for Greenhouse Gases Under
section 202(a) of the Clean Air Act; final rule (74 FR at 66543), "The commenters'
approach, if used globally, would effectively lead to a tragedy of the commons,
whereby no country or source category would be accountable for contributing to the
global problem of climate change, and nobody would take action as the problem
persists and worsens." While this rule does not singlehandedly eliminate climate
change, it is an important contribution to reducing the rate of change, and this
reduction in rate is global and long-lived. EPA appropriately placed the benefits of
reductions in context in the rule, by calculating the likely reductions in temperature
and comparing them to total projected changes in temperature over the same time
161
-------
period. In addition, EPA used the social cost of carbon methodology in order to
estimate a monetization of the benefits of these reductions (see section III.H.6), and
the net present value resulting from the CO2 reductions due to this rule (between
years 2017 and 2050) was calculated to be between tens to hundreds of billions of
dollars. As noted above, the D.C. Circuit pointedly rejected the argument that EPA
should refrain from issuing GHG standards under section 202(a) due to claimed lack
of mitigating effect on the endangerment, and further held that "the emission
standards would result in meaningful mitigation of greenhouse gas emissions" in the
form of "960 million metric tons of CChe over the lifetime of the model year 2012-
2016 vehicles." Coalition for Responsible Regulation v. EPA, 684 F. 3d 102, 128
(D.C. Cir. 2012); projected emissions reductions of this MYs 2017- 2025 rule are
projected to be approximately double those of the MYs 2012-2016 rule and thus, in
the D.C. Circuit's language, "result in meaningful mitigation of greenhouse gas
emissions."
4.2 Post-2025 Standards
Summary of Comments on the Draft TAR addressed in the Proposed Determination
While the agencies did not address post-2025 issues in the Draft TAR, the agencies did
receive comments on the need to consider long-term climate issues from a number of
organizations, including Consumers Union, The International Council on Clean Transportation,
University of Illinois Applied Environmental Law Program et al, Northeast States for
Coordinated Air Use Management, National Association of Clean Air Agencies, California Air
Resources Board, American Lung Association et al., and Fuel Freedom Foundation, among
others. In part as a response to these comments, EPA included in the Proposed Determination a
broader discussion of the need and opportunity for additional GHG reductions post-2025, as
described next.
Summary of Comments on the Proposed Determination
In Section V of the Proposed Determination, EPA discussed the need and opportunity for
substantial GHG emissions reductions from light-duty vehicles beyond 2025, as a step toward
having a dialogue with stakeholders. For illustrative purposes, Figure V.l in that document
presented a figure that showed projections for total U.S. light-duty vehicle plus upstream fuel
GHG emissions out to 2050 under three scenarios: 1) a business-as-usual scenario with no
regulatory changes after 2025, 2) a scenario where standards are reduced by 4.5 percent per year
for MY2026-2050, and 3) a scenario fitted to achieve the upper bound of the global GHG
emissions reduction range projected by the Intergovernmental Panel on Climate Change to limit
the global temperature rise to below 2°C. The section also briefly discussed the potential for
long-term transformational changes in the light-duty vehicle sector, and the possible impacts on
GHG emissions.
Five organizations specifically commented on the post-2025 discussion in the Proposed
Determination: ICCT, Environmental Defense Fund (EDF), Center for Biological Diversity
(CBD), American Council for an Energy-Efficient Economy (ACEEE), and the New York State
Department of Environmental Conservation (NYSDEC).
162
-------
All five commenters expressed support and/or appreciation for beginning the dialogue about
post-2025 issues. Two commenters provided more detailed comments on potential longer-term
goals. ICCT specifically encouraged the federal agencies to assess prospects for 2026-2030
standards that reflected annual reductions in GHG emissions and fuel consumption of 5 percent
per year. It further stated that 2030 is an appropriate time frame given that it would give the
industry similar lead time to what it initially had for the 2025 standards, and since California is
likely to begin work on 2030 climate policies in the near future. The Center for Biological
Diversity objected to the 2°C global temperature rise basis for one of the curves presented in the
Proposed Determination, and advocated for an alternative target based on the "well below 2°C"
language that was included in the Paris climate agreement.
Two of the commenters, ICCT and the Environmental Defense Fund, also observed that there
is additional technology available to automakers beyond that necessary to meet the 2025
standards, and more likely to be developed in the future due to ongoing innovation.
Response to Comments on the Proposed Determination
EPA appreciates the comments on this topic, as they reflect the very dialogue that Section V
was intended to spark. On the regulatory time frame of 2030 suggested by ICCT, EPA notes that
there are advantages and disadvantages with shorter and longer time frames—the former
provides more technological certainty but less time for innovation, while the latter entails less
technological certainty but more time for innovation. With respect to the long-term global
temperature rise goal raised by the Center for Biological Diversity, EPA notes that the curves
presented in the Proposed Determination were selected for illustrative purposes only and do not
reflect any judgment by EPA as to their sufficiency from a climate perspective or feasibility from
a technology perspective. EPA further notes that the upper bound of the range of global GHG
emissions reduction that would likely limit global temperature rise to 2°C, used in the illustrative
curve in the Proposed Determination, is directionally consistent with going "below 2°C," though
unlikely to limit the global temperature rise to 1.5°C.
Finally, EPA reiterates that its goal in including the discussion of post-2025 issues in the
Proposed Determination was simply to continue a dialogue, and we are pleased that this in fact is
occurring.
163
------- |