[Federal Register Volume 87, Number 84 (Monday, May 2, 2022)]
[Rules and Regulations]
[Pages 25710-26092]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 2022-07200]
[[Page 25709]]
Vol. 87
Monday,
No. 84
May 2, 2022
Part II
Department of Transportation
-----------------------------------------------------------------------
National Highway Traffic Safety Administration
-----------------------------------------------------------------------
49 CFR Parts 531, 533, 536, et al.
Corporate Average Fuel Economy Standards for Model Years 2024-2026
Passenger Cars and Light Trucks; Final Rule
Federal Register / Vol. 87 , No. 84 / Monday, May 2, 2022 / Rules and
Regulations
[[Page 25710]]
-----------------------------------------------------------------------
DEPARTMENT OF TRANSPORTATION
National Highway Traffic Safety Administration
49 CFR Parts 531, 533, 536, and 537
[NHTSA-2021-0053]
RIN 2127-AM34
Corporate Average Fuel Economy Standards for Model Years 2024-
2026 Passenger Cars and Light Trucks
AGENCY: National Highway Traffic Safety Administration (NHTSA).
ACTION: Final rule.
-----------------------------------------------------------------------
SUMMARY: NHTSA, on behalf of the Department of Transportation, is
finalizing revised fuel economy standards for passenger cars and light
trucks for model years (MYs) 2024-2025 that increase at a rate of 8
percent per year, and increase at a rate of 10 percent per year for MY
2026 vehicles. NHTSA currently projects that the revised standards
would require an industry fleet-wide average of roughly 49 mpg in MY
2026, and would reduce average fuel outlays over the lifetimes of
affected vehicles that provide consumers hundreds of dollars in net
savings. These standards are directly responsive to the agency's
statutory mandate to improve energy conservation and reduce the
Nation's energy dependence on foreign sources. This final rule fulfills
NHTSA's obligation to revisit the standards set forth in ``The Safer
Affordable Fuel Efficient (SAFE) Vehicles Rule for Model Years 2021-
2026 Passenger Cars and Light Trucks,'' as directed by President
Biden's January 20, 2021, Executive order ``Protecting Public Health
and the Environment and Restoring Science To Tackle the Climate
Crisis.'' The revised standards set forth in this final rule are
consistent with the policy direction in the order, to among other
things, listen to the science, improve public health and protect our
environment, and to prioritize both environmental justice and the
creation of the well paying union jobs necessary to deliver on these
goals. This final rule addresses public comments to the notice of
proposed rulemaking and also makes certain minor changes to fuel
economy reporting requirements.
DATES: This rule is effective July 1, 2022.
ADDRESSES: For access to the dockets or to read background documents or
comments received, please visit https://www.regulations.gov, and/or
Docket Management Facility, M-30, U.S. Department of Transportation,
West Building, Ground Floor, Room W12-140, 1200 New Jersey Avenue SE,
Washington, DC 20590. The Docket Management Facility is open between 9
a.m. and 4 p.m. Eastern Time, Monday through Friday, except Federal
holidays.
FOR FURTHER INFORMATION CONTACT: For technical and policy issues, Greg
Powell, CAFE Program Division Chief, Office of Rulemaking, National
Highway Traffic Safety Administration, 1200 New Jersey Avenue SE,
Washington, DC 20590; email: [email protected]. For legal issues,
Rebecca Schade, NHTSA Office of Chief Counsel, National Highway Traffic
Safety Administration, 1200 New Jersey Avenue SE, Washington, DC 20590;
email: [email protected].
SUPPLEMENTARY INFORMATION:
BILLING CODE 4910-59-P
[[Page 25711]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.000
[[Page 25712]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.001
[[Page 25713]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.002
[[Page 25714]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.003
[[Page 25715]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.004
[[Page 25716]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.005
[[Page 25717]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.006
[[Page 25718]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.007
[[Page 25719]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.008
[[Page 25720]]
Does this action apply to me?
This action affects companies that manufacture or sell new
passenger automobiles (passenger cars) and non-passenger automobiles
(light trucks) as defined under NHTSA's CAFE regulations.\1\ Regulated
categories and entities include:
---------------------------------------------------------------------------
\1\ ``Passenger car'' and ``light truck'' are defined in 49 CFR
part 523.
[GRAPHIC] [TIFF OMITTED] TR02MY22.009
BILLING CODE 4910-59-C
This list is not intended to be exhaustive, but rather provides a
guide regarding entities likely to be regulated by this action. To
determine whether particular activities may be regulated by this
action, you should carefully examine the regulations. You may direct
questions regarding the applicability of this action to the persons
listed in FOR FURTHER INFORMATION CONTACT.
Executive Summary
NHTSA, on behalf of the Department of Transportation, is amending
standards regulating corporate average fuel economy (CAFE) for
passenger cars and light trucks for MYs 2024-2026. This final rule
responds to NHTSA's statutory obligation to set CAFE standards at the
maximum feasible level that the agency determines vehicle manufacturers
can achieve in each model year, in order to improve energy
conservation. NHTSA's review of the prior standards was instigated in
response to President Biden's directive in Executive Order 13990 of
January 20, 2021, ``Protecting Public Health and the Environment and
Restoring Science To Tackle the Climate Crisis,'' that ``The Safer
Affordable Fuel-Efficient (SAFE) Vehicles Rule for Model Years 2021-
2026 Passenger Cars and Light Trucks'' (2020 final rule, SAFE rule, or
SAFE 2 final rule) (85 FR 24174, April 30, 2020) be immediately
reviewed for consistency with NHTSA's statutory obligation and our
Nation's abiding commitment to promote and protect our public health
and the environment, among other things. NHTSA undertook that review
immediately, and this final rule is the result of that review,
conducted with reference to NHTSA's statutory obligations.
The amended CAFE standards increase in stringency for both
passenger cars and light trucks, by 8 percent per year for MYs 2024-
2025, and by 10 percent per year for MY 2026. The agency calls the
amended standards Alternative 2.5. NHTSA concludes that these levels
are the maximum feasible for these model years as discussed in more
detail in Section VI. The final rule considers a range of regulatory
alternatives, consistent with NHTSA's obligations under the National
Environmental Policy Act (NEPA) and E.O. 12866. While E.O. 13990
directed the review of CAFE standards for MYs 2021-2026, statutory lead
time requirements \2\ mean that MY 2024 is the earliest model year that
can currently be amended in the CAFE program.\3\ The standards remain
vehicle-footprint-based, like the CAFE standards in effect since MY
2011. Recognizing that many readers think about CAFE standards in terms
of the miles per gallon (mpg) values that the standards are projected
to eventually require, NHTSA currently projects that the standards will
require, on an average industry fleet-wide basis, roughly 49 mpg in MY
2026. NHTSA notes both that real-world fuel economy is generally 20-30
percent lower than the estimated required CAFE level stated above, and
also that the actual CAFE standards are the footprint target curves for
passenger cars and light trucks, meaning that ultimate fleet-wide
levels will vary depending on the mix of vehicles that industry
produces for sale in those model years. Table I-1 shows the incremental
differences in stringency levels for passenger cars and light trucks,
by the different regulatory alternatives considered, in the model years
subject to regulation.
---------------------------------------------------------------------------
\2\ 49 U.S.C. 32902(a) and (g).
\3\ 49 U.S.C. 32902(a).
---------------------------------------------------------------------------
[[Page 25721]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.010
This final rule reflects a conclusion significantly different from
the conclusion that NHTSA reached in the 2020 final rule, but this is
because important facts have changed, and because NHTSA has
reconsidered how to balance the relevant statutory considerations in
light of those facts. In this document, NHTSA concludes that
significantly more stringent standards are the maximum feasible that
the agency determines that vehicle manufacturers can achieve in the
rulemaking time frame. Standards that are more stringent than those
that were finalized in 2020 appear economically practicable, based on
manageable average per-vehicle cost increases, large consumer fuel
savings, minimal effects on sales, and estimated increases in
employment, among other things. Additionally, and importantly, contrary
to the 2020 final rule, NHTSA recognizes that the need of the United
States to conserve energy must include serious consideration of the
energy security risks, as well as environmental and public health
implications, of continuing to consume oil, which more stringent fuel
economy standards can reduce. By increasing fuel economy, more
stringent standards can also protect consumers from oil market
volatility from global events outside the borders of the U.S. that can
result in rapid fuel price increases domestically. Through greater
energy conservation, more stringent standards also reduce climate
impacts to our Nation, which further benefit our national security.
NHTSA also believes that the final standards are complementary to other
motor vehicle standards of the Government that are simultaneously
applicable during MYs 2024-2026.
Moreover, at least part of the automobile industry is increasingly
demonstrating that improving fuel economy and reducing GHG emissions is
a growth market for them, and that the market rewards investment in
advanced technology. Nearly all auto manufacturers have rolled out new
higher fuel economy and electric vehicle models since MY 2020, and
continue to announce even more models forthcoming during the rulemaking
time frame. Five major manufacturers voluntarily bound themselves to
stricter GHG requirements than set forth by the U.S. Environmental
Protection Agency (EPA) in 2020 through contractual agreements with the
State of California.\4\ Some of the technologies that automakers will
deploy to meet those standards will both reduce emissions and improve
fuel economy. These companies (including both those who joined the
Framework Agreements with California and those that have not) are
sophisticated, for-profit enterprises. If they are taking these steps,
rolling out these new models, and making these announcements, NHTSA can
now be more confident than the agency was in 2020 that the market is
getting ready to make the leap to significantly higher fuel economy.
The California Framework Agreements and the clear planning by industry
to migrate toward more advanced technologies provide corroborating
evidence of the practicability of more stringent standards.
Additionally, more stringent CAFE standards can improve equity, by
encouraging industry to continue improving the fuel economy of all
vehicles, so that all Americans can benefit from higher fuel economy
and save money on fuel. While NHTSA does not consider the fuel economy
of electric vehicles in setting CAFE standards, consistent with
Congress' direction in 49 U.S.C. 32902(h), using electric vehicles to
meet the standards is a compliance option that many automakers are
pursuing. Further, NHTSA is setting these CAFE standards in the context
of a much larger conversation about the future of the U.S. light-duty
vehicle fleet, the increasing and obvious need to move away from fossil
fuels for reasons of national and energy security, and the evidence of
a changing climate that is emerging on an almost daily basis.
---------------------------------------------------------------------------
\4\ https://ww2.arb.ca.gov/news/framework-agreements-clean-cars
(accessed: March 23, 2022).
---------------------------------------------------------------------------
NHTSA concludes, as we will explain in more detail below, that
Alternative 2.5 is the maximum feasible alternative that manufacturers
can achieve for MYs 2024-2026, based on its significant fuel savings
benefits to consumers and its environmental and energy security
benefits relative to all other alternatives except Alternative 3.
Although Alternative 3 would provide greater fuel savings benefits,
NHTSA estimates that Alternative 3 would result in a large average per-
vehicle cost increase compared to the price of vehicles under
Alternative 2.5, which for many automakers could exceed $2,000. In
contrast to Alternative 3, Alternative 2.5
[[Page 25722]]
comes at a cost we believe the market can bear, and NHTSA believes it
is the appropriate choice given this record. We believe that providing
the greatest amount of lead time for the biggest stringency increase of
10 percent for MY 2026, the last of three years covered in the rule, is
reasonable and appropriate, particularly given the ongoing rapid
changes in the auto industry. Choosing Alternative 3 would require
industry to ramp up even faster, and thus provide less lead time, with
consequences for economic practicability. With relatively small sales
effects and positive effects on employment, we are confident that
Alternative 2.5 is feasible, and that industry can rise to meet these
standards.
For all of these reasons, and based on consideration of the
comments received, NHTSA concludes that Alternative 2.5, with standards
that increase at 8 percent per year for MYs 2024 and 2025, and a 10-
percent increase in MY 2026, is maximum feasible.
This action is also different from the 2020 final rule in that it
is issued by NHTSA alone, and EPA has issued a separate final rule.\5\
EPA's revised standards apply to MY 2023 as well as MYs 2024-2026.
NHTSA's 18-month lead time requirement precludes amendment of the MY
2023 CAFE standards. An important consequence of this is that EPA's
rate of stringency increase, after increasing in MY 2023, looks slower
than NHTSA's over the same time period, although collectively EPA's
standards achieve at least as stringent levels as NHTSA's Alternative
2.5 by MY 2026.\6\ NHTSA emphasizes, however, that the new standards
are what NHTSA believes best fulfill our statutory directive of energy
conservation. Additionally, in the context of the EPA standards, the
analysis we have done tackles the core question of whether compliance
with both standards should be achievable with the same vehicle fleet,
after manufacturers fully understand the requirements from both sets of
standards, and NHTSA believes that, as always, compliance with both
standards will be achievable with the same vehicle fleet. It is also
worth noting that the differences in what the two agencies' standards
require become smaller each year, until near alignment is achieved in
2026.
---------------------------------------------------------------------------
\5\ 86 FR 74434 (Dec. 30, 2021).
\6\ EPA projected a fleet average fuel economy value of about 52
mpg associated with its MY 2026 standards (assuming full use of air
conditioning refrigerant credits). See Table 4-43, ``Revised 2023
and Later Model Year Light-Duty Vehicle GHG Emissions Standards:
Regulatory Impact Analysis,'' EPA-420-R-21-028, December 2021.
---------------------------------------------------------------------------
While NHTSA recognizes that the last three CAFE standard
rulemakings have been issued jointly with EPA, and that issuing
separate rules represents a change in regulatory approach, NHTSA
coordinated with EPA to avoid inconsistencies and produce requirements
that are consistent with the agencies' respective statutory
authorities.\7\ Additionally, and importantly, NHTSA has also
considered and accounted for California's Zero Emission Vehicle (ZEV)
program (and its adoption by a number of other states) in developing
the baseline for this final rule, and has also accounted in the
baseline for the aforementioned ``Framework Agreements'' between
California and BMW, Ford, Honda, VWA, and Volvo, which are national-
level GHG emission reduction agreements to which these companies
committed for several model years. NHTSA reasonably assumes that
automakers will meet other regulatory requirements that apply to them,
and commitments that they have made through the Framework Agreements.
Reflecting these in the analysis improves the accuracy of the baseline
in reflecting the state of the world without the revised CAFE
standards, and thus the information available to the decision-makers.
---------------------------------------------------------------------------
\7\ Throughout this preamble, NHTSA uses the term ``maximum
feasible'' as shorthand to refer to the statutory directive in EPCA,
requiring the agency to exercise its discretionary authority to set
CAFE standards at the ``maximum feasible average fuel economy level
that the Secretary decides the manufacturers can achieve in that
model year.'' 49 U.S.C. 32902(a).
---------------------------------------------------------------------------
A number of other improvements and updates have been made to the
analysis since the 2020 final rule based on NHTSA analysis, new data,
and public comments to the NPRM (86 FR 49602, Sept. 3, 2021) as
described in Section III. Table I-2 summarizes these, and they are
discussed in much more detail below and in the documents accompanying
this preamble.
[[Page 25723]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.011
BILLING CODE 4910-59-C
[[Page 25724]]
NHTSA estimates that this action could reduce average fuel outlays
over the lifetimes of MY 2029 vehicles by about $1,387, while
increasing the average cost of those vehicles by about $1,087 over the
baseline described above, at a 3-percent discount rate. With the social
cost of greenhouse gases (SC-GHG) \8\ and all other benefits and costs
discounted at 3 percent, when considering the entire fleet for MYs
1981-2029, NHTSA estimates $128 billion in monetized costs and $145
billion in monetized benefits attributable to the new standards, such
that the present value of aggregate net monetized benefits to society
would be over $16 billion, not including other important unquantified
effects, such as energy security benefits, distributional effects, and
certain air quality benefits from the reduction of toxic air pollutants
and other emissions, among other things.
---------------------------------------------------------------------------
\8\ The ``social cost of greenhouse gases'' or ``SC-GHG'' refers
to the combination of the social costs of carbon dioxide
(CO2), methane (CH4), and nitrous oxide
(N2O) emissions. In this preamble, and in the TSD, FRIA,
and Final SEIS, NHTSA may occasionally use the term ``social cost of
carbon'' or ``SCC'' to refer to the SC-GHG, and means no substantive
difference between them.
---------------------------------------------------------------------------
These cost and benefit estimates are based on many different and
uncertain inputs. One of the inputs informing the benefits estimates is
the SC-GHG. In this final rule, NHTSA employed the SC-GHG values from
the Interim Revised Estimates developed by the Interagency Working
Group on the Social Cost of Greenhouse Gases (IWG), and discounted it
at values recommended by the IWG for its main analysis. Those values
are based on the best available science and economics and are the most
appropriate values to focus on in the analysis of this rule, though DOT
also affirms that, in its expert judgment, those values are
conservative estimates that likely significantly underestimate the full
benefits to social welfare of reducing greenhouse gas pollution. NHTSA
also explored in its sensitivity analyses values based on other
assumptions, including values calculated at different discount rates,
Furthermore, in light of pending litigation, NHTSA also explored an
analysis that used the same SC-GHG value employed in the 2020 final
rule. Specifically, on February 11, 2022, the United States District
Court for the Western District of Louisiana issued a preliminary
injunction that enjoined NHTSA from, among other activities,
``[a]dopting, employing, treating as binding, or relying upon any
Social Cost of Greenhouse Gas estimates based on global effects,'' as
well as from ``adopting, employing, treating as binding, or relying
upon the work product of the [IWG].'' \9\
---------------------------------------------------------------------------
\9\ Louisiana v. Biden, Order, No. 2:21-CV-01074, ECF No. 99
(W.D. La. Feb. 11, 2022).
---------------------------------------------------------------------------
Although the injunction was stayed by the United States Court of
Appeals for the Fifth Circuit on March 16, 2022,\10\ prior to the stay,
in order to comply with this prohibition, NHTSA conducted a cost-
benefit analysis based on the SC-GHG values presented in the 2020 final
rule. In DOT's judgment, those values do not reflect the best available
science and economics for estimating climate effects in the analysis of
this rule. As detailed more thoroughly elsewhere in this rule and the
supporting Technical Support Document (TSD) and Final Regulatory Impact
Analysis (FRIA), the only way to achieve an efficient allocation of
resources for greenhouse gas emissions reduction on a global basis--and
so benefit the United States and its citizens--is for all countries to
consider global estimates of climate damages. To correctly assess the
total climate damages to U.S. citizens and residents, an analysis must
account for all climate impacts that directly and indirectly affect the
welfare of U.S. citizens and residents, how U.S. greenhouse gas
mitigation activities affect mitigation activities by other countries,
and spillover effects from climate action elsewhere. The estimates used
in the 2020 rule, therefore, severely underestimate climate damages.
Nevertheless, even if NHTSA's cost-benefit analysis applied the
misleadingly low SC-GHG estimates from the 2020 rule, which severely
underestimate the impacts of climate effects on U.S. citizens, NHTSA
would still conclude in this rule that Alternative 2.5 is maximum
feasible under its statutory authority. Notably, for example, net
consumer benefits from significant fuel savings remained positive for
Alternative 2.5 independent of any estimate of climate benefits.
---------------------------------------------------------------------------
\10\ Louisiana v. Biden, Order, No. 22-30087, Doc. No.
00516242341 (5th Cir. Mar. 16, 2022).
---------------------------------------------------------------------------
Moreover, NHTSA is required to consider four statutory factors--
technological feasibility, economic practicability, the effect of other
motor vehicle standards of the Government on fuel economy, and the need
of the United States to conserve energy--to determine whether the
standards it adopts are maximum feasible,\11\ and NHTSA finds that
Alternative 2.5 is the maximum feasible on the basis of these factors,
and particularly considering the statutory mandate to improve energy
conservation and reduce the Nation's energy dependence on foreign
sources. The cost-benefit analysis is not one of those statutory
factors. While NHTSA's estimates of costs and benefits are important
considerations and are directed by E.O. 12866, again, it is the
balancing required by statute--that is, the requirement to set CAFE
standards at ``the maximum feasible average fuel economy level that the
Secretary decides the manufacturers can achieve in that model year'' 49
U.S.C. 32902(a)--that is the basis for the setting of CAFE standards.
Cost-benefit analysis provides only one informative data point in
addition to the host of considerations that NHTSA must balance by
statute when determining maximum feasible standards. As such, any
changes in the monetized climate benefit figures that resulted from
using the SC-GHG value from the 2020 final rule did not justify
disrupting the overall balance of other significant qualitative and
quantitative considerations and factors that support the selection of
the Preferred Alternative--as described at length throughout this final
rule. When the 5th Circuit stayed the injunction, NHTSA returned to
using the Interim SC-GHG developed by the IWG, discounted at 3 percent,
because we believe it to be the more accurate and reasonable value.
---------------------------------------------------------------------------
\11\ 49 U.S.C. 32902(g).
---------------------------------------------------------------------------
It is worth emphasizing that CAFE standards apply only to new
vehicles. The costs attributable to new CAFE standards are thus
``front-loaded,'' because they result primarily from the application of
fuel-saving technology to new vehicles. By contrast, the impact of new
CAFE standards on fuel consumption and energy savings, air pollution,
and greenhouse gases--and the associated benefits to society--occur
over an extended time, as drivers buy, use, and eventually scrap these
new vehicles. By accounting for many model years and extending well
into the future (2050), our analysis accounts for these differing
patterns in impacts, benefits, and costs. Given the front-loaded costs
versus longer-term benefits, it is likely that an analysis extending
even further into the future would reveal at least some additional net
present benefits. Our analysis also accounts for the potential that, by
changing new vehicle prices and fuel economy levels, CAFE standards
could indirectly impact the operation of vehicles produced before or
after the MYs 2024-2026 for which we are finalizing new CAFE standards.
This means that some of the final rule's impacts and corresponding
benefits and costs are actually attributable to indirect
[[Page 25725]]
impacts on vehicles produced before and after MYs 2024-2026.
The bulk of our analysis considers a ``model year'' perspective
that considers the lifetime impacts attributable to all vehicles
produced prior to MY 2030, accounting for the operation of these
vehicles over their entire lives (with some MY 2029 vehicles estimated
to be in service as late as 2068). This approach emphasizes the role of
MYs 2024-2026, while accounting for the potential that it may take
manufacturers a few additional years to produce fleets fully responsive
to the final MY 2026 standards,\12\ and for the potential that the
final standards could induce some changes in the operation of vehicles
produced prior to MY 2024, for example, some individuals might choose
to keep older vehicles in operation, rather than purchase new ones.
---------------------------------------------------------------------------
\12\ The fact that manufacturers have up to three model years to
``settle'' compliance for a given model year is a function of
statutory flexibilities--namely, that overcompliance credits may be
``carried back'' up to three model years--and does not in any way
imply that NHTSA believes that the MY 2026 standards are not
feasible in MY 2026.
---------------------------------------------------------------------------
Our analysis also considers a ``calendar year'' (CY) perspective
that includes the annual impacts attributable to all vehicles estimated
to be in service in each calendar year for which our analysis includes
a representation of the entire registered light-duty fleet. For this
final rule, this calendar year perspective covers each of CYs 2021-
2050, with differential impacts accruing as early as MY 2023.\13\
Compared to the ``model year'' perspective, this calendar year
perspective emphasizes model years of vehicles produced in the longer
term, beyond those model years for which standards are currently being
promulgated. Table I-3 summarizes estimates of selected impacts viewed
from each of these two perspectives, for each of the regulatory
alternatives considered in this final rule.\14\
---------------------------------------------------------------------------
\13\ For a presentation of effects by calendar year, please see
FRIA Chapter 6.5 and Chapter 6.6.
\14\ As discussed at length below, Alternative 0 is the set of
CAFE standards promulgated in 2020, and thus constitutes the ``No-
Action Alternative.'' Impacts of the four ``Action Alternatives''
are measured relative to this baseline. Alternatives 1, 2, 2.5, and
3 specify passenger car and light truck standards for each of MYs
2024-2026 that NHTSA estimates will, taken together, increase
overall CAFE requirements in MY 2026 by about 14, 22, 25, and 30
percent, respectively, although actual average requirements will
ultimately depend on the future composition of the fleet, which
NHTSA cannot predict with certainty. Above, Table I-1 shows
corresponding projected increases in average requirements for each
fleet in each model year. Below, Section IV.B discusses the specific
definitions of each of these regulatory alternatives.
---------------------------------------------------------------------------
BILLING CODE 4910-59-P
[[Page 25726]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.013
Additional important health, environmental, and energy security
benefits could not be fully quantified or monetized. Finally, for
purposes of comparing the benefits and costs of new CAFE standards to
the benefits and costs of other Federal regulations, policies, and
programs, we have computed ``annualized'' benefits and costs.
---------------------------------------------------------------------------
\15\ Climate benefits are based on reductions in CO2,
CH4, and N2O emissions and are calculated
using four different estimates of the global social cost of each
greenhouse gas (SC-GHG model average at 2.5 percent, 3 percent, and
5 percent discount rates; 95th percentile at 3 percent discount
rate), which each increase over time. For the presentational
purposes of this table and other similar summary tables, we show the
benefits associated with the average global SC-GHG at a 3 percent
discount rate, but the agency does not have a single central SC-GHG
point estimate. We emphasize the importance and value of considering
the benefits calculated using all four SC-GHG estimates. See Section
III.G.2 for more information. Where percent discount rate values are
reported in this table, the social benefits of avoided climate
damages are discounted at 3 percent. The climate benefits are
discounted at the same discount rate as used in the underlying SC-
GHG values for internal consistency.
\16\ To be clear, monetized values do not include other
important unquantified effects, such as certain climate benefits,
certain energy security benefits, distributional effects, and
certain air quality benefits from the reduction of toxic air
pollutants and other emissions, among other things.
[GRAPHIC] [TIFF OMITTED] TR02MY22.014
[[Page 25727]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.015
[GRAPHIC] [TIFF OMITTED] TR02MY22.016
[GRAPHIC] [TIFF OMITTED] TR02MY22.017
[GRAPHIC] [TIFF OMITTED] TR02MY22.018
[GRAPHIC] [TIFF OMITTED] TR02MY22.019
[[Page 25728]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.020
[GRAPHIC] [TIFF OMITTED] TR02MY22.021
Again, and as discussed in detail below, the monetized estimated
costs and benefits of this final rule are relevant to and inform the
agency's conclusion regarding which levels of CAFE standards are
maximum feasible for MYs 2024-2026, but they do not fully capture the
total benefits of the standards and are not part of the factors
contained in the governing statute. It is the balancing of the four
statutory factors (none of which expressly requires maximization of net
benefits, although NHTSA does consider net benefits pursuant to E.O.
12866) that provides the basis for setting CAFE standards. Notably,
NHTSA confirms that on the basis of its four statutory factors, and
particularly considering the statutory mandate to improve energy
conservation and reduce the Nation's energy dependence on foreign
sources, NHTSA would select Alternative 2.5 as the maximum feasible
even if the cost-benefit analysis had adopted different assumptions for
the monetization of climate benefits.
It is also worth emphasizing that, although NHTSA is prohibited
from considering the availability of certain flexibilities in making
our determination about the levels of CAFE standards that would be
maximum feasible,\17\ manufacturers have a variety of flexibilities
available to them to aid their compliance. Table I-12 through Table I-
15 below summarize available compliance flexibilities.
---------------------------------------------------------------------------
\17\ 49 U.S.C. 32902(h).
[GRAPHIC] [TIFF OMITTED] TR02MY22.022
[[Page 25729]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.023
[GRAPHIC] [TIFF OMITTED] TR02MY22.024
[GRAPHIC] [TIFF OMITTED] TR02MY22.025
BILLING CODE 4910-59-C
NHTSA recognizes that the lead time for this final rule is shorter
than some past rulemakings have provided, and that the economy and the
country are in the process of recovering from a global pandemic and the
resulting economic distress. At the same time, NHTSA also recognizes
that at least parts of the industry are nonetheless stepping up their
product offerings and releasing more and more high-fuel-economy vehicle
models, and many companies did not deviate significantly over the past
ten years from product plans established in response to the EPA and
NHTSA standards set forth in the 2012 final rule (77 FR 62624, Oct. 15,
2012) and the EPA standards confirmed by EPA in its January 2017 Final
Determination. With these and other considerations in mind, NHTSA is
amending the CAFE standards for MYs 2024-2026, and believes that
Alternative 2.5 is maximum feasible and represents the best balancing
of multiple statutory and policy goals for these model years. NHTSA,
like any other Federal agency, is afforded an opportunity to reconsider
prior views and, when warranted, to adopt new positions. Indeed, as a
matter of good governance, agencies should revisit their positions when
appropriate, especially to ensure that their actions and regulations
reflect legally sound interpretations of the agency's statutory
authority and remain consistent with the agency's policy views and
practices. As a matter of law, ``an Agency is entitled to change its
interpretation of a statute.'' \18\ Nonetheless, ``[w]hen an Agency
adopts a materially changed interpretation of a statute, it must in
addition provide a `reasoned analysis' supporting its decision to
revise its interpretation.'' \19\ The analysis presented in this
preamble and in the accompanying TSD, FRIA, Final Supplemental
Environmental Impact Statement (Final SEIS), CAFE Model Documentation,
and extensive
[[Page 25730]]
rulemaking docket fully supports the agency's decision and revised
balancing of the statutory factors for MYs 2024-2026 standards.
---------------------------------------------------------------------------
\18\ Phoenix Hydro Corp. v. FERC, 775 F.2d 1187, 1191 (D.C. Cir.
1985).
\19\ Alabama Educ. Ass'n v. Chao, 455 F.3d 386, 392 (D.C. Cir.
2006) (quoting Motor Vehicle Mfrs. Ass'n of U.S., Inc. v. State Farm
Mut. Auto. Ins. Co., 463 U.S. 29, 57 (1983)); see also Encino
Motorcars, LLC v. Navarro, 136 S. Ct. 2117, 2125 (2016) (``Agencies
are free to change their existing policies as long as they provide a
reasoned explanation for the change.'') (citations omitted).
---------------------------------------------------------------------------
II. Overview of the Final Rule
In this final rule, NHTSA is revising CAFE standards for MYs 2024-
2026. On January 20, 2021, the President signed E.O. 13990,
``Protecting Public Health and the Environment and Restoring Science To
Tackle the Climate Crisis.'' \20\ In it, the President directed that
the 2020 final rule must be immediately reviewed for consistency with
the policy commitments in that E.O., including listening to the
science; improving public health and protect our environment; ensuring
access to clean air and water; limiting exposure to dangerous chemicals
and pesticides; holding polluters accountable, including those who
disproportionately harm communities of color and low-income
communities; reducing greenhouse gas emissions; bolstering resilience
to the impacts of climate change; restoring and expanding our national
treasures and monuments; and prioritizing both environmental justice
and the creation of the well-paying union jobs necessary to deliver on
these goals.\21\ E.O. 13990 states expressly that the Administration
prioritizes listening to the science, improving public health and
protecting the environment, reducing greenhouse gas emissions, and
improving environmental justice while creating well-paying union
jobs.\22\ The E.O. thus directs that the 2020 final rule be reviewed at
once and that (in this case) the Secretary of Transportation consider
``suspending, revising, or rescinding'' it, via an NPRM, by July
2021.\23\ On September 3, 2021, NHTSA published an NPRM to revise these
requirements, which are being finalized, with changes in response to
public comments and additional analysis, in this final rule.
---------------------------------------------------------------------------
\20\ 84 FR 7037 (Jan. 25, 2021).
\21\ Id., sections 1, 2.
\22\ Id., section 1.
\23\ Id., section 2(a)(ii).
---------------------------------------------------------------------------
Section 32902(g)(1) of title 49, United States Code allows the
Secretary (by delegation to NHTSA) to prescribe regulations amending an
average fuel economy standard prescribed under 49 U.S.C. 32902(a), like
those prescribed in the 2020 final rule, if the amended standard meets
the requirements of section 32902(a). The Secretary's authority to set
fuel economy standards is delegated to NHTSA at 49 CFR 1.95(a);
therefore, NHTSA is revising fuel economy standards for MYs 2024-2026.
Section 32902(g)(2) states that when the amendment makes an average
fuel economy standard more stringent, it must be prescribed at least 18
months before the beginning of the model year to which the amendment
applies. NHTSA generally calculates the 18-month lead time requirement
as April of the calendar year prior to the start of the model year.
Thus, 18 months before MY 2023 would be April 2021, because MY 2023
begins in October 2022. Because of this lead time requirement, NHTSA is
not amending the CAFE standards for MYs 2021-2023, even though the 2020
final rule also covered those model years. For purposes of the CAFE
program, the 2020 final rule's standards for MYs 2021-2023 will remain
in effect.
For the model years for which there is statutory lead time to amend
the standards, however, NHTSA is amending the currently applicable fuel
economy standards. Although only two years have passed since the 2020
final rule, the agency believes it is reasonable and appropriate to
revisit the CAFE standards for MYs 2024-2026. In particular, the agency
has further considered the serious adverse effects on energy
conservation that the standards finalized in 2020 would cause as
compared to the final standards. The need of the U.S. to conserve
energy is greater than understood in the 2020 final rule. In addition,
informed by an updated technical analysis, standards that are more
stringent than those that were finalized in 2020 appear economically
practicable, based on manageable average per-vehicle cost increases,
minimal effects on sales, and estimated increases in employment, as
well as higher (and increasing) consumer demand for more fuel economy,
among other considerations. NHTSA also believes that the final
standards are complementary to other motor vehicle standards of the
Government that affect fuel economy that are simultaneously applicable
during MYs 2024-2026. The renewed focus on addressing energy
conservation and the industry's apparent ability to meet more stringent
standards show that a rebalancing of the EPCA factors, and a
corresponding issuance of more stringent standards, is appropriate for
MYs 2024-2026.
The following sections introduce the action in more detail.
Summary of NPRM
In the NPRM, NHTSA proposed to revise the existing CAFE standards
for MYs 2024-2026. NHTSA explained that it was proposing to revise
those standards because it had reconsidered its determination made in
2020 about what levels of CAFE stringency would be maximum feasible for
those model years, after reviewing the standards in response to the
President's direction in E.O. 13990. NHTSA discussed the differences
between the proposal and the 2020 final rule, including NHTSA's
tentative conclusion that significantly more stringent standards would
be maximum feasible, based on a reconsideration of how to balance the
relevant statutory considerations and updated technical information.
NHTSA also discussed the fact that it was issuing the proposal
independently, unlike several past rulemakings in which NHTSA and EPA
had issued joint proposals. NHTSA explained that EPA's revised
standards apply to MY 2023 as well as MYs 2024-2026, while NHTSA's 18-
month lead time requirement precluded amendment of the MY 2023 CAFE
standards. An important consequence of this was that EPA's proposed
rate of stringency increase, after taking a big leap in MY 2023, looked
slower than NHTSA's over the same time period. NHTSA emphasized,
however, that the proposed standards were what NHTSA believed best
fulfilled our statutory directive of energy conservation, and that the
agencies had worked closely together in developing their respective
proposals, and that by the end of the rulemaking time frame, alignment
would be achieved between the two agencies' standards. NHTSA also
explained that it had employed an analytical baseline for the NPRM that
included both a representation of the California ZEV program (and its
adoption in a number of states) and the California ``Framework
Agreements'' between that state and BMW, Ford, Honda, Volkswagen of
America (VWA), and Volvo. NHTSA also described other analytical
improvements made for the NPRM since the 2020 final rule.
NHTSA proposed CAFE standards for MYs 2024-2026 that would increase
at a rate of 8 percent per year, for both passenger cars and light
trucks, and also took comment on a wide range of alternatives,
including retaining the 2020 standards and returning to levels
consistent with what was set forth in the 2012 final rule. Table II-1
and Table II-2 below contain descriptions of the regulatory
alternatives on which comment was sought, and the estimated translation
of those alternatives into mpg levels, respectively, for the reader's
reference. The proposal was accompanied by a Preliminary Regulatory
Impact Analysis (PRIA), a Draft Supplemental Environmental Impact
Statement (Draft SEIS), and the
[[Page 25731]]
CAFE Model software source code and documentation, all of which were
also subject to comment in their entirety and all of which received
significant comments.
[GRAPHIC] [TIFF OMITTED] TR02MY22.026
[GRAPHIC] [TIFF OMITTED] TR02MY22.027
NHTSA also sought comment on another potential alternative, the
effects of which were not expressly quantified, under which MYs 2024-
2025 would increase at 8 percent per year, but MY 2026 would increase
at 10 percent per year. NHTSA explained that average requirements and
achieved CAFE levels would ultimately depend on manufacturers' and
consumers' responses to standards, technology developments, economic
conditions, fuel prices, and other factors. NHTSA estimated that over
the lives of vehicles produced prior to MY 2030, the proposal would
save about 50 billion gallons of gasoline and increase electricity
consumption (as the percentage of electric vehicles increased over
time) by about 275 terawatts (TWh), compared to the levels of gasoline
and electricity consumption that NHTSA projected would occur under the
baseline standards. Accounting for emissions from both vehicles and
upstream energy sector processes, NHTSA estimated that the proposal
would reduce greenhouse gas emissions by about 465 million metric tons
of carbon dioxide, about 500 thousand metric tons of methane, and about
12 thousand metric tons of nitrous oxide. NHTSA also estimated that
emissions of criteria pollutants would generally decline dramatically
over time.
In terms of economic effects, NHTSA estimated that for an average
MY 2029 vehicle subject to the proposed standards, consumers could see
a price increase of $960, but would gain lifetime fuel savings of
$1,280. With the SC-GHG discounted at 2.5 percent and other benefits
and costs discounted at 3 percent, NHTSA estimated that costs and
benefits could be approximately $120 billion and $121 billion,
respectively, such that the present value of aggregate net benefits to
society could be somewhat less than $1 billion. With the SC-GHG
discounted at 3 percent and other benefits and costs discounted at 7
percent, NHTSA estimated approximately $90 billion in costs and $76
billion in benefits, such that the present value of aggregate net costs
to society could be approximately $15 billion.
NHTSA explained that it tentatively concluded that Alternative 2
was maximum feasible for MYs 2024-2026 based on new information and a
reconsideration of how to interpret and balance the statutory factors,
as compared to the decision made in the 2020 final rule. The 2020 rule
had prioritized industry concerns and sought to reduce new vehicle
costs to consumers, based on assumptions about low consumer demand for
higher fuel economy vehicles and a discounting of the need of the U.S.
to conserve energy. In the NPRM, NHTSA recognized the importance of the
need of the U.S. to conserve energy, and tentatively concluded that
ongoing manufacturer announcements and rollouts of new higher-fuel-
economy vehicles indicated industry expectation of growing consumer
demand for those vehicles, such that more stringent standards could be
economically practicable. NHTSA underscored that ``an [a]gency is
entitled to change its interpretation of
[[Page 25732]]
a statute,'' \24\ even though ``[w]hen an [a]gency adopts a materially
changed interpretation of a statute, it must in addition provide a
`reasoned analysis' supporting its decision to revise its
interpretation.'' \25\
---------------------------------------------------------------------------
\24\ Phoenix Hydro Corp. v. FERC, 775 F.2d 1187, 1191 (D.C. Cir.
1985).
\25\ Alabama Educ. Ass'n. v. Chao, 455 F.3d 386, 392 (D.C. Cir.
2006) (quoting Motor Vehicle Mfrs. Ass'n. of U.S., Inc. v. State
Farm Mut. Auto. Ins. Co., 463 U.S. 29, 57 (1983)); see also Encino
Motorcars, LLC v. Navarro, 136 S. Ct. 2117, 2125 (2016) (``Agencies
are free to change their existing policies as long as they provide a
reasoned explanation for the change.'') (citations omitted).
---------------------------------------------------------------------------
NHTSA also addressed the question of harmonization with other motor
vehicle standards of the Government that affect fuel economy. Even
though NHTSA and EPA issued separate rather than joint notices, NHTSA
explained that it had worked closely with EPA in developing the
respective proposals, and that the agencies had sought to minimize
inconsistency between the programs where doing so was consistent with
the agencies' respective statutory mandates. NHTSA emphasized that
differences between the proposals, especially as regards programmatic
flexibilities, were not new in the proposal, and that differences were
often a result of the different statutory frameworks. NHTSA reminded
readers that since the agencies had begun regulating concurrently under
President Obama, these differences have meant that manufacturers have
had (and will have) to plan their compliance strategies considering
both the CAFE standards and the GHG standards and assure that they are
in compliance with both. NHTSA explained that it was proposing CAFE
standards that would increase at 8 percent per year over MYs 2024-2026
because that was what NHTSA had tentatively concluded was maximum
feasible during those model years, under the EPCA factors.
NHTSA was also confident that industry would still be able to build
a single fleet of vehicles to meet both the NHTSA and EPA standards,
even if it required them to be slightly more strategic than they might
otherwise have preferred. NHTSA sought comment broadly on all aspects
of the proposal.
B. Public Participation Opportunities and Summary of Comments
The NPRM was published on NHTSA's website on August 10, 2021, and
published in the Federal Register on September 3, 2021,\26\ beginning a
60-day comment period. The agency left the docket open for considering
late comments to the extent practicable. A separate Federal Register
notification, also published on September 14, 2021 (86 FR 51092),
announced a virtual public hearing taking place on October 13th and
14th of 2021. Approximately 77 individuals and organizations signed up
to participate in the hearing. The hearing started at 9:30 a.m. EDT on
October 13th and ended at approximately 5:30 p.m., completing the
entire list of participants within a single day, resulting in a 58-page
transcript.\27\ The hearing also collected many pages of comments from
participants, in addition to the hearing transcript, all of which were
submitted to the docket for the rule.
---------------------------------------------------------------------------
\26\ 86 FR 49602 (Sept. 3, 2021).
\27\ The transcript is available in the docket for this rule.
---------------------------------------------------------------------------
Besides the comments submitted as part of the public hearings,
NHTSA's docket received a total of 67,256 form letters, 1,636
individual comments from stakeholder organizations, and 693 attachments
in response to the proposal, for an overall total of 69,585
submissions. NHTSA also received several hundred comments on its Draft
SEIS to the separate Draft SEIS docket (NHTSA-2021-0054). While the
majority of individual comments were form letters, the agency received
over 6,000 pages of substantive comments on the proposal.
Many commenters generally supported the proposal. Commenters
supporting the proposal tended to cite concerns about climate change,
which are relevant to the need of the United States to conserve energy,
and the need for Federal programs to continue or expand for a carbon-
neutral, carbon-free future. Commenters also expressed the need for
NHTSA and EPA harmonization and close coordination for their respective
programs. Citizens and environmental groups demonstrated strong support
for pushing the proposed standard to Alternative 3 or beyond, while
closing potential loopholes in the program. There were mixed views on
NHTSA's inclusion of battery electric vehicles in NHTSA's modeling
analysis. Many manufacturers supported alignment with EPA's proposed
standards, while electric vehicle manufacturers such as Tesla and
Rivian supported NHTSA's Alternative 3.
In other areas, commenters expressed mixed views on the statutorily
mandated Petroleum Equivalency Factor (PEF) used to calculate mpg
values for electrified vehicles and the disclosure of credit trading
information in NHTSA's revised reporting templates.
Discussion and responses to comments can be found throughout this
preamble in areas applicable to the comment received.
Nearly every aspect of the NPRM's analysis and discussion received
some level of comment by at least one commenter. The comments received,
as a whole, were both broad and deep, and the agency appreciates the
level of engagement of commenters in the public comment process and the
information and opinions provided.
C. Changes in Light of Public Comments and New Information
Comments received to the NPRM were considered carefully, because
they are critical for understanding stakeholders' positions, as well as
for gathering additional information that can help to inform the agency
about aspects or effects of the proposal that the agency may not have
considered at the time of the proposal. The views, data, requests, and
suggestions contained in the comments help us to form solutions and
make appropriate adjustments to our proposals so that we may be better
assured that the final standards we set are, indeed, maximum feasible
for the rulemaking time frame.
For this final rule, the agency made substantive changes resulting
directly from the suggestions and recommendations from commenters, as
well as new information obtained from the time the proposal was
developed, and corrections both highlighted by commenters and
discovered internally. These changes reflect DOT's long-standing
commitment to ongoing refinement of its approach to estimating the
potential impacts of new CAFE standards. Through further consideration
and deliberation, and also in response to many public comments received
since then, NHTSA has made a number of changes to the CAFE Model since
the 2020 final rule, including those that are listed in the Executive
Summary and detailed in Section III, as well as in the TSD and FRIA
that accompany this final rule.
D. Final Standards--Stringency
NHTSA is setting CAFE standards for passenger cars and light trucks
manufactured for sale in the United States in MYs 2024-2026. Passenger
cars are generally sedans, station wagons, and two-wheel drive
crossovers and sport utility vehicles (CUVs and SUVs), while light
trucks are generally 4WD sport utility vehicles, pickups, minivans, and
passenger/cargo vans.\28\ The final standards, represented by
Alternative 2.5 in NHTSA's analysis, increase at a rate of 8 percent
per year for both cars and trucks for MYs 2024-
[[Page 25733]]
2025, and at a rate of 10 percent for MY 2026 cars and trucks. The
final standards, like the proposed standards, are defined by a
mathematical equation that represents a constrained linear function
relating vehicle footprint to fuel economy targets for both cars and
trucks.\29\
---------------------------------------------------------------------------
\28\ ``Passenger car'' and ``light truck'' are defined at 49 CFR
part 523.
\29\ Vehicle footprint is roughly measured as the rectangle that
is made by the four points where the vehicle's tires touch the
ground. Generally, passenger cars have more stringent targets than
light trucks regardless of footprint, and smaller vehicles will have
more stringent targets than larger vehicles. No individual vehicle
or vehicle model need meet its target exactly, but a manufacturer's
compliance is determined by how its average fleet fuel economy
compares to the average fuel economy of the targets of the vehicles
it manufactures.
---------------------------------------------------------------------------
The target curves for passenger cars and light trucks are as
follows; curves for MYs 2020-2023 are included in the figures for
context. NHTSA underscores that the equations and coefficients defining
the curves are, in fact, the CAFE standards, and not the mpg numbers
that the agency currently estimates could result from manufacturers
complying with the curves. Because the estimated mpg numbers are an
effect of the final standards, they are presented in Section II.E.
BILLING CODE 4910-59-P
[GRAPHIC] [TIFF OMITTED] TR02MY22.028
[[Page 25734]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.029
NHTSA has also amended the minimum domestic passenger car CAFE
standards for MYs 2024-2026. Section 32902(b)(4) of 49 U.S.C. requires
NHTSA to project the minimum standard when it promulgates passenger car
standards for a model year, so the minimum standards are established as
specific mpg values at this time. NHTSA retained the 1.9-percent offset
used in the 2020 final rule, such that the minimum domestic passenger
car standard is as shown in Table II-3.
[GRAPHIC] [TIFF OMITTED] TR02MY22.030
[[Page 25735]]
The next section describes some of the effects that NHTSA estimates
would follow from the final standards for passenger cars and light
trucks for MYs 2024-2026, including how the curves shown above
translate to estimated average mile per gallon requirements for the
industry.
Final Standards--Impacts
As for past CAFE rulemakings, NHTSA has used the CAFE Model to
estimate the effects of this final rule's CAFE standards, and of other
regulatory alternatives under consideration. Some inputs to the CAFE
Model are derived from other models, such as Argonne National
Laboratory's ``Autonomie'' vehicle simulation tool and Argonne's
``GREET'' fuel-cycle emissions analysis model, the U.S. Energy
Information Administration's (EIA's) National Energy Modeling System
(NEMS), and EPA's ``MOVES'' vehicle emissions model. Especially given
the scope of the NHTSA's analysis (through MY 2050, with driving of MY
2029 vehicles accounted for through CY 2068), these inputs involve a
multitude of uncertainties. For example, a set of inputs with
significant uncertainty could include future population and economic
growth, future gasoline and electricity prices, future petroleum market
characteristics (e.g., imports and exports), future battery costs,
manufacturers' future responses to standards and fuel prices, buyers'
future responses to changes in vehicle prices and fuel economy levels,
and future emission rates for ``upstream'' processes (e.g., refining,
finished fuel transportation, electricity generation). Considering that
all of this is, to some extent, uncertain from a current vantage point,
NHTSA underscores that all results of this analysis are, in turn,
uncertain, and simply represent the agency's best estimates based on
the information currently before us and on the agency's reasonable
judgment.
NHTSA estimates that this final rule would increase the eventual
\30\ average of manufacturers' CAFE requirements to about 49 mpg by
2026 rather than, under the No-Action Alternative (i.e., the baseline
standards issued in 2020), about 40 mpg. For passenger cars, the
average in 2026 is estimated to reach just over 59 mpg, and for light
trucks, just over 42 mpg. This compares with 47 mpg and 34 mpg for cars
and trucks, respectively, under the No-Action Alternative.
---------------------------------------------------------------------------
\30\ Here, ``eventual'' means by MY 2029, after most of the
fleet will have been redesigned under the MY 2026 standards. NHTSA
allows the CAFE Model to continue working out compliance solutions
for the regulated model years for three model years after the last
regulated model year, in recognition of the fact that manufacturers
do not comply perfectly with CAFE standards in each model year.
[GRAPHIC] [TIFF OMITTED] TR02MY22.031
Because manufacturers do not comply exactly with each standard in
each model year, but rather focus their compliance efforts when and
where it is most cost-effective to do so, ``estimated achieved'' fuel
economy levels differ somewhat from ``estimated required'' levels for
each fleet, for each year. NHTSA estimates that the industry-wide
average fuel economy achieved in MY 2029 could increase from about 44
mpg under the No-Action Alternative to 50 mpg under the final rule's
standards.
[GRAPHIC] [TIFF OMITTED] TR02MY22.032
As discussed above, NHTSA's analysis--unlike its CAFE analyses for
previous rulemakings--estimates manufacturers' potential responses to
the combined effect of CAFE standards and separate CO2
standards (including agreements some manufacturers have reached with
California), ZEV mandates, and fuel prices. Together, the
aforementioned regulatory programs are more binding (i.e., require more
of manufacturers) than any single program considered in isolation, and
this analysis, like past analyses, shows some estimated overcompliance
with the final CAFE standards, albeit by much less than what was shown
in the NPRM that preceded the 2020 final rule, and any overcompliance
is highly manufacturer-dependent.
The estimated average CO2 levels equivalent to the above
required and achieved CAFE levels (using 8,887 grams of CO2
per gallon of gasoline vehicle certification fuel) are provided in
Table II-6 and Table II-7.
[[Page 25736]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.033
[GRAPHIC] [TIFF OMITTED] TR02MY22.034
Average requirements and achieved CAFE levels would ultimately
depend on manufacturers' and consumers' responses to standards,
technology developments, economic conditions, fuel prices, and other
factors.
NHTSA estimates that over the lives of vehicles produced prior to
MY 2030, the final standards would save about 60 billion gallons of
gasoline and increase electricity consumption (as the percentage of
electric vehicles increases over time) by about 180 terawatts (TWh),
compared to levels of gasoline and electricity consumption NHTSA
projects would occur under the baseline standards (i.e., the No-Action
Alternative) as shown in Table II-8.\31\
---------------------------------------------------------------------------
\31\ While NHTSA does not consider electrification in its
analysis during the rulemaking time frame, the analysis still
reflects application of electric vehicles in the baseline fleet and
during the model years after the rulemaking time frame, such that
electrification (and thus, electricity consumption) increases in
NHTSA's analysis even though NHTSA is not considering it in our
decision-making.
[GRAPHIC] [TIFF OMITTED] TR02MY22.035
NHTSA's analysis also estimates total annual consumption of fuel by
the entire on-road fleet from CY 2020 through CY 2050. On this basis,
gasoline and electricity consumption by the U.S. light-duty vehicle
fleet evolves as shown in Figure II-3 and Figure II-4, each of which
shows projections for the No-Action Alternative (Alternative 0, i.e.,
the baseline), Alternative 1, Alternative 2, Alternative 2.5 (the
Preferred Alternative), and Alternative 3.
[[Page 25737]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.036
[[Page 25738]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.037
Accounting for emissions from both vehicles and upstream energy
sector processes (e.g., petroleum refining and electricity generation),
which are relevant to NHTSA's evaluation of the need of the United
States to conserve energy, NHTSA estimates that the final rule would
reduce greenhouse gas emissions by about 607 million metric tons of
carbon dioxide (CO2), about 733 thousand metric tons of
methane (CH4), and about 17 thousand tons of nitrous oxide
(N2O).
BILLING CODE 4910-59-P
[GRAPHIC] [TIFF OMITTED] TR02MY22.038
As for fuel consumption, NHTSA's analysis also estimates annual
emissions attributable to the entire on-road fleet from CY 2020 through
CY 2050. Also accounting for both vehicles and upstream processes,
NHTSA estimates that CO2 emissions could evolve over time as
shown in Figure II-5, which accounts for both emissions from both
vehicles and upstream processes.
[[Page 25739]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.039
BILLING CODE 4910-59-C
Estimated emissions of methane and nitrous oxides follow similar
trends. As discussed in the TSD, FRIA, and this preamble, NHTSA has
performed two types of supporting analysis. This document and FRIA
focus on the ``standard setting'' analysis, which sets aside the
potential that manufacturers could respond to standards by using
compliance credits or introducing new alternative fuel vehicle
(including BEVs) models during the ``decision years'' (for this
document, 2024, 2025, and 2026). The accompanying Final SEIS focuses on
an ``unconstrained'' analysis, which does not set aside these potential
manufacturer actions. The Final SEIS presents much more information
regarding projected GHG emissions, as well as model-based estimates of
corresponding impacts on several measures of global climate change.
Also accounting for vehicular and upstream emissions, NHTSA has
estimated annual emissions of most criteria pollutants (i.e.,
pollutants for which EPA has issued National Ambient Air Quality
Standards). NHTSA estimates that under each regulatory alternative,
annual emissions of carbon monoxide (CO), volatile organic compounds
(VOC), nitrogen oxide (NOX), and particulate matter with a
diameter equal to or less than 2.5 microns (PM2.5)
attributable to the light-duty on-road fleet will decline dramatically
between 2020 and 2050, and that emissions in any given year could be
very nearly the same under each regulatory alternative. For example,
Figure II-6 shows NHTSA's estimate of future NOX emissions
under each alternative.
BILLING CODE 4910-59-P
[[Page 25740]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.040
BILLING CODE 4910-59-C
On the other hand, as discussed in the FRIA and Final SEIS, NHTSA
projects that annual SO2 emissions attributable to the
light-duty on-road fleet could increase modestly under the action
alternatives, because, as discussed above, NHTSA projects that each of
the action alternatives could lead to greater use of electricity (for
PHEVs and BEVs). The adoption of actions--such as actions prompted by
President Biden's Executive order directing agencies to develop a
Federal Clean Electricity and Vehicle Procurement Strategy--to reduce
electricity generation emission rates beyond projections underlying
NHTSA's analysis (discussed in Chapter 5 of the TSD) could dramatically
reduce SO2 emissions under all regulatory alternatives
considered here.\32\
---------------------------------------------------------------------------
\32\ https://www.whitehouse.gov/briefing-room/presidential-actions/2021/01/27/executive-order-on-tackling-the-climate-crisis-at-home-and-abroad/ (accessed February 11, 2022).
---------------------------------------------------------------------------
For the ``standard setting'' analysis, the FRIA accompanying this
document provides additional detail regarding projected criteria
pollutant emissions and health effects, as well as the inclusion of
these impacts in this benefit-cost analysis. For the ``unconstrained''
or ``EIS'' type of analysis, the Final SEIS accompanying this document
presents much more information regarding projected criteria pollutant
emissions, as well as model-based estimates of corresponding impacts on
several measures of urban air quality and public health. As mentioned
above, these estimates of criteria pollutant emissions are based on a
complex analysis involving interacting simulation techniques and a
myriad of input estimates and assumptions. Especially extending well
past 2040, the analysis involves a multitude of uncertainties.
Therefore, actual criteria pollutant emissions could ultimately be
different from NHTSA's current estimates.
To illustrate the effectiveness of the technology added in response
to this final rule, Table II-10 presents NHTSA's estimates for
increased vehicle cost and lifetime fuel expenditures if we assumed the
behavioral response to the lower cost of driving were zero.\33\ These
numbers are presented in lieu of NHTSA's primary estimate of lifetime
fuel savings, which would give an incomplete picture of technological
effectiveness because the analysis accounts for consumers' behavioral
response to the lower cost-per-mile of driving a more fuel-efficient
vehicle.
---------------------------------------------------------------------------
\33\ While this comparison illustrates the effectiveness of the
technology added in response to this final rule, it does not
represent a full consumer welfare analysis, which would account for
drivers' likely response to the lower cost-per-mile of driving, as
well as a variety of other benefits and costs they will experience.
The agency's complete analysis of the final rule's likely impacts on
passenger car and light truck buyers appears in the FRIA, Appendix
I, Table A-23-1.
---------------------------------------------------------------------------
[[Page 25741]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.041
With the SC-GHG discounted at 3 percent and other benefits and
costs discounted at 3 percent, NHTSA estimates that monetized costs and
benefits could be approximately $128 billion and $145 billion,
respectively, such that the present value of aggregate monetized net
benefits to society could be approximately $16 billion. With the SC-GHG
discounted at 3 percent and other benefits and costs discounted at 7
percent, NHTSA estimates approximately $96 billion in monetized costs
and $100 billion in monetized benefits could be attributable to
vehicles produced prior to MY 2030 over the course of their lives, such
that the present value of aggregate net monetized benefits to society
could be approximately $4 billion.
[GRAPHIC] [TIFF OMITTED] TR02MY22.042
The following two tables provides a range of benefits and net
benefits representing varying discount rates for the social cost of
carbon with all other benefits discounted at 3 percent and 7 percent,
respectively.
BILLING CODE 4910-59-P
[GRAPHIC] [TIFF OMITTED] TR02MY22.043
[[Page 25742]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.044
BILLING CODE 4910-59-C
Model results can be viewed many different ways, and NHTSA's
rulemaking considers both ``model year'' and ``calendar year''
perspectives. The ``model year'' perspective, above, considers vehicles
projected to be produced in some range of model years, and accounts for
impacts, benefits, and costs attributable to these vehicles from the
present (from the model year's perspective, 2020) until they are
projected to be scrapped. The bulk of NHTSA's analysis considers
vehicles produced prior to MY 2030, accounting for the estimated
indirect impacts new standards could have on the remaining operation of
vehicles already in service. This perspective emphasizes impacts on
those model years nearest to those (2024-2026) for which NHTSA is
finalizing new standards. NHTSA's analysis also presents some results
focused only on MYs 2024-2026, setting aside the estimated indirect
impacts on earlier model years, and the impacts estimated to occur
during MYs 2027-2029, as some manufacturers and products ``catch up''
to the standards.
Another way to present the benefits and costs of the final rule is
the ``calendar year'' perspective shown in Table II-14, which is
similar to how EPA presents benefits and costs in its final analysis
for GHG standards. The calendar year perspective considers all vehicles
projected to be in service in each of some range of future calendar
years. NHTSA's presentation of results from this perspective considers
CYs 2021-2050, because the model's representation of the full on-road
fleet extends through 2050. Unlike the model year perspective, this
perspective includes vehicles projected to be produced during MYs 2021-
2050. This perspective emphasizes longer-term impacts that could accrue
if standards were to continue without change. Under the calendar year
perspective, net benefits for the standards are estimated to be nearly
$112 billion by 2050 at a 3 percent discount rate, and over $73 billion
by 2050 at a 7 percent discount rate.
[GRAPHIC] [TIFF OMITTED] TR02MY22.045
[[Page 25743]]
Finally, Table II-15 shows costs and benefits over the narrow
perspective of the lives of MY 2023-2026 vehicles while Table II-11
shows a wider perspective of the costs and benefits over the remaining
lives of all vehicles produced through MY 2029.
[GRAPHIC] [TIFF OMITTED] TR02MY22.046
Though based on the exact same model results, these two
perspectives provide considerably different views of estimated costs
and benefits. Because technology costs account for a large share of
overall estimated costs, and are also projected to decline over time
(as manufacturers gain more experience with new technologies), costs
tend to be ``front loaded''--occurring early in a vehicle's life and
tending to be higher in earlier model years than in later model years.
Conversely, because social benefits of standards occur as vehicles are
driven, and because both fuel prices and the social cost of
CO2 emissions are projected to increase in the future,
benefits tend to be ``back loaded.'' As a result, estimates of future
fuel savings, CO2 reductions, and net social benefits are
higher under the calendar year perspective than under the model year
perspective. On the other hand, with longer-term impacts playing a
greater role, the calendar year perspective is more subject to
uncertainties regarding, for example, future technology costs and fuel
prices.
Even though NHTSA and EPA estimate benefits, costs, and net
benefits using similar methodologies and achieve similar results,
different approaches to accounting may give the false appearance of
significant divergences. Table II-13 above presents NHTSA's results
using comparable accounting to EPA's preamble Table 4. EPA also
presents cost and benefit information in its RIA over CYs 2021 through
2050.\34\ The numbers most comparable to those presented in EPA's RIA
are those NHTSA developed to complete its Final SEIS using an identical
accounting approach. This is because the statutory limitations
constraining NHTSA's standard setting analysis, such as those in 49
U.S.C. 32902(h), do not similarly apply to its ``unconstrained''
analysis, some effects of which are used in NHTSA's Final SEIS.\35\
NHTSA's ``unconstrained'' analysis estimates $312 billion in monetized
costs, $443 billion in monetized benefits, and $132 billion in
monetized net benefits using a 3-percent discount rate over CYs 2021
through 2050, with the social cost of carbon discounted at 3
percent.\36\ NHTSA describes its cost and benefit accounting approach
in Section V of this preamble.
---------------------------------------------------------------------------
\34\ EPA's RIA is available at https://www.epa.gov/regulations-emissions-vehicles-and-engines/final-rule-revise-existing-national-ghg-emissions (accessed: March 24, 2022).
\35\ As the Final SEIS analysis contains information that NHTSA
is statutorily prevented from considering, the agency is limited on
the extent this analysis is used in regulatory decision-making.
Additionally, the Final SEIS includes no cost and benefit analysis,
and does not rely in any way on the social cost of greenhouse gas
emissions.
\36\ See FRIA Chapter 6.5 for more information regarding NHTSA's
estimates of annual benefits and costs using NHTSA's standard
setting analysis. See Tables B-7-25 through B-7-30 in Appendix II of
the FRIA for a more detailed breakdown of NHTSA's Final SEIS
analysis.
---------------------------------------------------------------------------
Final Standards Are the Maximum Feasible
NHTSA's conclusion, after consideration of the factors described
below and information in the administrative record for this action, is
that 8-percent increases in stringency for MYs 2024-2025 and a 10-
percent increase for MY 2026 for both passenger cars and light trucks
(Alternative 2.5 of this analysis) are maximum feasible. The Department
of Transportation is deeply committed to working aggressively to
improve energy conservation and reduce environmental harms and economic
and security risks associated with energy use. NHTSA agrees with many
public comments suggesting that the need of the United States to
conserve energy and protect the environment compels more stringent
standards than those set in 2020 if they appear to be consistent with
the other factors that NHTSA must consider. NHTSA has concluded that
Alternative 2.5 is technologically feasible, is economically
practicable (based on manageable average per-vehicle cost increases,
minimal effects on sales, and estimated increases in employment, among
other considerations), and is complementary to other motor vehicle
standards of the Government on fuel economy that are simultaneously
applicable during MYs 2024-2026, as described in more detail below.
Despite only 2 years having passed since the 2020 final rule, enough
has changed in the United States and the world, including as reflected
in the technical analysis, that revisiting the CAFE standards for MYs
2024-2026, and raising their stringency considerably, is both
appropriate and reasonable.
The 2020 final rule set CAFE standards that increased at 1.5
percent per year for cars and trucks for MYs 2021-2026, in large part
because it prioritized industry concerns and reducing upfront costs to
consumers and manufacturers--even at the expense of longer-term net
savings to consumers. This final rule reflects greater emphasis on the
statutory priority of energy conservation, while also taking into
account other statutory requirements. Moreover, NHTSA is also legally
required to consider the environmental implications of this action
under NEPA, and while the 2020 final rule did undertake a NEPA
analysis, it did not prioritize the environmental
[[Page 25744]]
considerations encompassed within the statutory mandate to set
``maximum feasible'' fuel economy standards to conserve energy. This
rule also reflects NHTSA's updated technical analysis.
NHTSA recognizes that the amount of lead time available before MY
2024 is less than what was provided in the 2012 rule. The amount of
lead time is nevertheless consistent with the agency's statutory
requirements. As will be discussed further in Section VI, NHTSA
believes that the evidence suggests that the final standards are
economically practicable as explained above and as discussed in Section
VI.
We note further that while this final rule is different from the
2020 final rule (and also from the 2012 final rule), NHTSA, like any
other Federal agency, is afforded an opportunity to reconsider prior
views and, when warranted, to adopt new positions. Indeed, as a matter
of good governance, agencies should revisit their positions when
appropriate, especially to ensure that their actions and regulations
reflect legally sound interpretations of the agency's statutory
authority and remain consistent with the agency's policy views and
practices. As a matter of law, ``an [a]gency is entitled to change its
interpretation of a statute.'' \37\ Nonetheless, ``[w]hen an [a]gency
adopts a materially changed interpretation of a statute, it must in
addition provide a `reasoned analysis' supporting its decision to
revise its interpretation.'' \38\ This preamble and the accompanying
TSD and FRIA all provide extensive detail on the agency's updated
analysis, and Section VI contains the agency's explanation of how the
agency has considered that analysis and other relevant information in
determining that the standards represented by Alternative 2.5 are
maximum feasible for MY 2024-2026 passenger cars and light trucks.
---------------------------------------------------------------------------
\37\ Phoenix Hydro Corp. v. FERC, 775 F.2d 1187, 1191 (D.C. Cir.
1985).
\38\ Alabama Educ. Ass'n v. Chao, 455 F.3d 386, 392 (D.C. Cir.
2006) (quoting Motor Vehicle Mfrs. Ass'n of U.S., Inc. v. State Farm
Mut. Auto. Ins. Co., 463 U.S. 29, 57 (1983)); see also Encino
Motorcars, LLC v. Navarro, 136 S. Ct. 2117, 2125 (2016) (``Agencies
are free to change their existing policies as long as they provide a
reasoned explanation for the change.'') (citations omitted).
---------------------------------------------------------------------------
Final Standards Are Feasible in the Context of EPA's Final Standards
and California's Programs
The NHTSA and EPA final rules remain coordinated despite being
issued as separate regulatory actions. Because NHTSA and EPA are
regulating the exact same vehicles and manufacturers will use many of
the same technologies to meet both sets of standards, NHTSA coordinated
with EPA during the development of each agency's independent rulemaking
to revise their respective standards set forth in the 2020 final rule.
The NHTSA CAFE and EPA CO2 standards for MY 2026 represent
roughly equivalent levels of stringency. While the rates of increase
for the final CAFE and CO2 standards for MYs 2024-2026 are
different, the specific differences in what the two agencies' standards
require become smaller each year, until near alignment is achieved in
2026. NHTSA nevertheless coordinated closely with EPA to minimize
inconsistency between the programs while still ensuring that NHTSA's
standards were maximum feasible for MYs 2024-2026.
While NHTSA's and EPA's programs differ in certain other respects,
like programmatic flexibilities, those differences are not new in this
final rule. Some parts of the programs are harmonized, and others
differ, often as a result of the respective statutory frameworks. Since
NHTSA and EPA began coordinating their regulations under President
Obama, differences in programmatic flexibilities have meant that
manufacturers have had (and will have) to plan their compliance
strategies considering both the CAFE standards and the GHG standards
and assure that they are in compliance with both. NHTSA is finalizing
CAFE standards that increase at 8 percent per year over MYs 2024-2025
and at 10 percent per year for MY 2026 because that is what NHTSA has
concluded is maximum feasible in those model years, under the EPCA
factors. Auto manufacturers are extremely sophisticated companies, well
able to manage compliance strategies that account for multiple
regulatory programs concurrently. Past experience with these programs
indicates that each manufacturer will optimize its compliance strategy
around whichever standard is most binding for its fleet of vehicles. If
different agencies' standards are more binding for some companies in
certain years, this does not mean that manufacturers must build
multiple fleets of vehicles, simply that they will have to be more
strategic about how they build their fleet. NHTSA discusses this issue
in greater detail in Section VI.A of this preamble. Critically, NHTSA
has concluded that it is feasible for manufacturers to meet both the
EPA and the NHTSA standards.\39\
---------------------------------------------------------------------------
\39\ This is consistent with NHTSA's and EPA joint finding in
the 2012 final rule, as discussed further in Section VI below.
---------------------------------------------------------------------------
NHTSA has also considered and accounted for California's ZEV
mandate (and its adoption by a number of other states) in developing
the baseline for this final rule, as additional legal obligations that
automakers will be meeting during this time frame, and has also
accounted for the Framework Agreements between California and BMW,
Ford, Honda, VWA, and Volvo, as those companies have committed to
meeting those Agreements. NHTSA believes that it is appropriate to
include ZEV in the baseline for this final rule because EPA has granted
a waiver of Clean Air Act preemption to California for its Clean Cars
Program,\40\ and it is appropriate for the baseline to reflect other
legal obligations that automakers will be meeting during this time
period. The baseline should reflect the state of the world without the
CAFE standards so that the regulatory analysis can identify the
distinct effects of the CAFE standards. In addition, according to
information provided by California,\41\ there has been extensive
industry overcompliance with the ZEV standards, which suggests that
regardless of the waiver, many companies intend to produce ZEVs in
volumes comparable to what the current ZEV mandate would require. Thus,
including state ZEV mandates in the regulatory baseline for this final
rule is consistent with guidance in OMB Circular A-4 directing agencies
to develop analytical baselines that are as accurate as possible
regarding the state of the world in the absence of the regulatory
action being evaluated. However, because modeling a subnational fleet
is not currently an analytical option for NHTSA, NHTSA has not
expressly accounted for California GHG standards in the analysis for
this final rule. Chapter 6 of the accompanying FRIA shows the estimated
effects of all of these programs simultaneously.
---------------------------------------------------------------------------
\40\ 87 FR 14332 (Mar. 14, 2022).
\41\ See, e.g., https://ww2.arb.ca.gov/sites/default/files/2020-01/appendix_a_minimum_zev_regulation_compliance_scenarios_formatted_ac.pdf (accessed: March 24, 2022) (stating that ``Since the 2012
adoption of the ACC requirements, vehicle technology has advanced
faster and developed more broadly than originally anticipated, and
the assumptions used in the original rulemaking scenario no longer
reflect vehicles expected in the 2018 through 2025 timeframe.'').
---------------------------------------------------------------------------
III. Technical Foundation for Final Rule Analysis
Why does NHTSA conduct this analysis?
NHTSA is establishing revised CAFE standards for passenger cars and
light trucks produced for MYs 2024-2026. NHTSA establishes CAFE
standards under the Energy Policy and Conservation Act, as amended, and
this final rule is undertaken pursuant to that authority. This final
rule would require
[[Page 25745]]
CAFE stringency for both passenger cars and light trucks to increase at
a rate of 8 percent, 8 percent, and 10 percent per year annually during
MY 2024, MY 2025, and MY 2026, respectively. NHTSA estimates that over
the useful lives of vehicles produced prior to MY 2030, these standards
would save about 60 billion gallons of gasoline and increase
electricity consumption by about 180 TWh. Accounting for emissions from
both vehicles and upstream energy sector processes (e.g., petroleum
refining and electricity generation), NHTSA estimates that these
standards would reduce greenhouse gas emissions by about 605 million
metric tons of carbon dioxide (CO2), about 730 thousand
metric tons of methane (CH4), and about 17 thousand tons of
N2O.
When NHTSA promulgates new regulations, it generally presents an
analysis that estimates the impacts of such regulations, and the
impacts of other regulatory alternatives. These analyses derive from
statutes such as the Administrative Procedure Act (APA), National
Environmental Policy Act (NEPA), Executive orders (such as E.O. 12866
and E.O. 13653), and from other administrative guidance (e.g., Office
of Management Budget Circular A-4). For CAFE, the Energy Policy and
Conservation Act (EPCA), as amended by the Energy Independence and
Security Act (EISA), contains a variety of provisions that require
NHTSA to consider certain compliance elements in certain ways and avoid
considering other things, in determining maximum feasible CAFE
standards. Collectively, capturing all of these requirements and
guidance elements analytically means that, at least for CAFE, NHTSA
presents an analysis that spans a meaningful range of regulatory
alternatives, that quantifies a range of technological, economic, and
environmental impacts, and that does so in a manner that accounts for
EPCA's express requirements for the CAFE program (e.g., passenger cars
and light trucks are regulated separately, and the standard for each
fleet must be set at the maximum feasible level in each model year).
NHTSA's decision regarding the final standards is thus supported by
extensive analysis of potential impacts of the regulatory alternatives
under consideration. Along with this preamble, a TSD, a FRIA, and a
Final SEIS, together provide an extensive and detailed enumeration of
related methods, estimates, assumptions, and results. These additional
analyses can be found in the rulemaking docket for this final rule \42\
and on NHTSA's website.\43\ NHTSA's analysis has been constructed
specifically to reflect various aspects of governing law applicable to
CAFE standards and has been expanded and improved in response to
comments received to the prior rulemaking and to the proposal, as well
as additional work conducted over the last year or two. Further
improvements may be made in the future based on comments received to
the proposal, which were either out of scope for this rulemaking or for
which the improvements were too extensive or complex to implement in
the available time, on the 2021 NAS Report,\44\ and on other additional
work generally previewed in these rulemaking documents. The analysis
for this final rule aided NHTSA in implementing its statutory
obligations, including the weighing of various considerations, by
reasonably informing decision-makers about the estimated effects of
choosing different regulatory alternatives.
NHTSA's analysis makes use of a range of data (i.e., observations
of things that have occurred), estimates (i.e., things that may occur
in the future), and models (i.e., methods for making estimates). Two
examples of data include (1) records of actual odometer readings used
to estimate annual mileage accumulation at different vehicle ages and
(2) CAFE compliance data used as the foundation for the ``analysis
fleet'' containing, among other things, production volumes and fuel
economy levels of specific configurations of specific vehicle models
produced for sale in the U.S. Two examples of estimates include (1)
forecasts of future GDP growth used, with other estimates, to forecast
future vehicle sales volumes and (2) the ``retail price equivalent''
(RPE) factor used to estimate the ultimate cost to consumers of a given
fuel-saving technology, given accompanying estimates of the
technology's ``direct cost,'' as adjusted to account for estimated
``cost learning effects'' (i.e., the tendency that it will cost a
manufacturer less to apply a technology as the manufacturer gains more
experience doing so).
---------------------------------------------------------------------------
\42\ Docket No. NHTSA-2021-0053, which can be accessed at
https://www.regulations.gov.
\43\ See https://www.nhtsa.gov/laws-regulations/corporate-average-fuel-economy.
\44\ National Academies of Sciences, Engineering, and Medicine,
2021. Assessment of Technologies for Improving Fuel Economy of
Light-Duty Vehicles--2025-2035, Washington, DC: The National
Academies Press (hereafter, ``2021 NAS Report''). Available at
https://www.nationalacademies.org/our-work/assessment-of-technologies-for-improving-fuel-economy-of-light-duty-vehicles-phase-3 (accessed: February 11, 2022) and for hard-copy review at
DOT headquarters.
---------------------------------------------------------------------------
NHTSA uses the CAFE Compliance and Effects Modeling System (usually
shortened to the ``CAFE Model'') to estimate manufacturers' potential
responses to new CAFE and CO2 standards and to estimate
various impacts of those responses. DOT's Volpe National Transportation
Systems Center (often simply referred to as the ``Volpe Center'')
develops, maintains, and applies the model for NHTSA. NHTSA has used
the CAFE Model to perform analyses supporting every CAFE rulemaking
since 2001. The 2016 rulemaking regarding heavy-duty pickup and van
fuel consumption and CO2 emissions also used the CAFE Model
for analysis.
The basic design of the CAFE Model is as follows: The system first
estimates how vehicle manufacturers might respond to a given regulatory
scenario, and from that potential compliance solution, the system
estimates what impact that response will have on fuel consumption,
emissions, and economic externalities. In a highly summarized form,
Figure III-1 shows the basic categories of CAFE Model procedures and
the sequential flow between different stages of the modeling. The
diagram does not present specific model inputs or outputs, as well as
many specific procedures and model interactions. The model
documentation accompanying this preamble presents these details, and
Chapter 1 of the TSD contains a more detailed version of this flow
diagram for readers who are interested.
BILLING CODE 4910-59-P
[[Page 25746]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.047
BILLING CODE 4910-59-C
More specifically, the model may be characterized as an integrated
system of models. For example, one model estimates manufacturers'
responses, another estimates resultant changes in total vehicle sales,
and still another estimates resultant changes in fleet turnover (i.e.,
scrappage). Additionally, and importantly, the model does not determine
the form or stringency of the standards. Instead, the model applies
inputs specifying the form and stringency of standards to be analyzed
and produces outputs showing the impacts of manufacturers working to
meet those standards, which become the basis for comparing between
different potential stringencies. A regulatory scenario, meanwhile,
involves specification of the form, or shape, of the standards (e.g.,
flat standards, or linear or logistic attribute-based standards), scope
of passenger car and truck regulatory classes, and stringency of the
CAFE standards for each model year to be analyzed. For example, a
regulatory scenario may define CAFE standards that increase in
stringency by a given percent per year for a given number of
consecutive years.
Manufacturer compliance simulation and the ensuing effects
estimation, collectively referred to as compliance modeling, encompass
numerous subsidiary elements. Compliance simulation begins with a
detailed user-provided initial forecast of the vehicle models offered
for sale during the simulation period.\45\ The compliance simulation
then attempts to bring each manufacturer into compliance with the
standards defined by the regulatory
[[Page 25747]]
scenario contained within an input file developed by the user.\46\
---------------------------------------------------------------------------
\45\ Because the CAFE Model is publicly available, anyone can
develop their own initial forecast (or other inputs) for the model
to use. The DOT-developed Market Data file that contains the
forecast used for this final rule is available on NHTSA's website at
https://www.nhtsa.gov/corporate-average-fuel-economy/cafe-compliance-and-effects-modeling-systems. (Accessed: March 22, 2022).
\46\ With appropriate inputs, the model can also be used to
estimate impacts of manufacturers' potential responses to new
CO2 standards and to California's ZEV program.
---------------------------------------------------------------------------
Estimating impacts involves calculating resultant changes in new
vehicle costs, estimating a variety of costs (e.g., for fuel) and
effects (e.g., CO2 emissions from fuel combustion) occurring
as vehicles are driven over their lifetimes before eventually being
scrapped, and estimating the monetary value of these effects.
Estimating impacts also involves consideration of consumer responses--
e.g., the impact of vehicle fuel economy, operating costs, and vehicle
price on consumer demand for passenger cars and light trucks. Both
basic analytical elements involve the application of many analytical
inputs. Many of these inputs are developed outside of the model and not
by the model. For example, the model applies fuel prices; it does not
estimate fuel prices.
NHTSA also uses EPA's MOVES model to estimate ``tailpipe'' (a.k.a.
``vehicle'' or ``downstream'') emission factors for criteria
pollutants,\47\ and uses four DOE and DOE-sponsored models to develop
inputs to the CAFE Model, including three developed and maintained by
DOE's Argonne National Laboratory. The agency uses the DOE Energy
Information Administration's (EIA's) National Energy Modeling System
(NEMS) to estimate fuel prices,\48\ and uses Argonne's Greenhouse
gases, Regulated Emissions, and Energy use in Transportation (GREET)
model to estimate emissions rates from fuel production and distribution
processes.\49\ DOT also sponsored DOE/Argonne to use Argonne's
Autonomie full-vehicle modeling and simulation system to estimate the
fuel economy impacts for over a million combinations of technologies
and vehicle types.50 51 The TSD and FRIA describe details of
the agency's use of these models. In addition, as discussed in the
Final SEIS accompanying this final rule, DOT relied on a range of
climate models to estimate impacts on climate, air quality, and public
health. The Final SEIS discusses and describes the use of these models.
---------------------------------------------------------------------------
\47\ See https://www.epa.gov/moves. This final rule uses version
MOVES3, available at https://www.epa.gov/moves/latest-version-motor-vehicle-emission-simulator-moves. (Accessed: February 16, 2022).
\48\ See https://www.eia.gov/outlooks/archive/aeo21. (Accessed:
February 16, 2022) This final rule uses fuel prices estimated using
the Annual Energy Outlook (AEO) 2021 version of NEMS (see https://www.eia.gov/outlooks/aeo/pdf/02%20AEO2021%20Petroleum.pdf).
(Accessed: February 16, 2022).
\49\ Information regarding GREET is available at https://greet.es.anl.gov/index.php. (Accessed: February 16, 2022) This final
rule uses the 2021 version of GREET.
\50\ As part of the Argonne simulation effort, individual
technology combinations simulated in Autonomie were paired with
Argonne's BatPaC model to estimate the battery cost associated with
each technology combination based on characteristics of the
simulated vehicle and its level of electrification. Information
regarding Argonne's BatPaC model is available at https://www.anl.gov/cse/batpac-model-software. (Accessed: February 16,
2022).
\51\ In addition, the impact of engine technologies on fuel
consumption, torque, and other metrics was characterized using GT-
POWER simulation modeling in combination with other engine modeling
that was conducted by IAV Automotive Engineering, Inc. (IAV). The
engine characterization ``maps'' resulting from this analysis were
used as inputs for the Autonomie full-vehicle simulation modeling.
Information regarding GT-POWER is available at https://www.gtisoft.com/gt-suite-applications/propulsion-systems/gt-power-engine-simulation-software. (Accessed: February 16, 2022).
---------------------------------------------------------------------------
To prepare for analysis supporting this final rule, DOT has refined
and expanded the CAFE Model through ongoing development. Examples of
such changes, some informed by past external comments, made since early
2020 include:
Inclusion of 400- and 500-mile BEVs;
Inclusion of high compression ratio (HCR) engines with
cylinder deactivation;
Accounting for manufacturers' responses to both CAFE and
CO2 standards jointly (rather than only separately);
Accounting for the ZEV mandates applicable in California
and the ``Section 177'' states;
Accounting for some vehicle manufacturers' (BMW, Ford,
Honda, VW, and Volvo) voluntary agreement with the state of California
to continued annual national-level reductions of vehicle greenhouse gas
emissions through MY 2026, with greater rates of electrification than
would have been required under the 2020 final rule; \52\
---------------------------------------------------------------------------
\52\ For more information on the Framework Agreements for Clean
Cars, including the specific agreements signed by individual
manufacturers, see https://ww2.arb.ca.gov/news/framework-agreements-clean-cars. (Accessed: February 16, 2022).
---------------------------------------------------------------------------
Inclusion of CAFE civil penalties in the ``effective
cost'' metric used when simulating manufacturers' potential application
of fuel-saving technologies;
Refined procedures to estimate health effects and
corresponding monetized damages attributable to criteria pollutant
emissions;
New procedures to estimate the impacts and corresponding
monetized damages of highway vehicle crashes that do not result in
fatalities;
Procedures to ensure that modeled technology application
and production volumes are the same across all regulatory alternatives
in the earliest model years; and
Procedures to more precisely focus application of the
EPCA's ``standard setting constraints'' (i.e., regarding the
consideration of compliance credits and additional dedicated
alternative fueled vehicles) to only those model years for which NHTSA
is proposing or finalizing new standards.
These changes reflect DOT's long-standing commitment to ongoing
refinement of its approach to estimating the potential impacts of new
CAFE standards. Following the proposal preceding this document, NHTSA
made several further changes to the CAFE Model, including:
New options for applying a dynamic fleet share model (of
the relative shares passenger cars and light trucks comprise of the
total U.S. new vehicle market);
Provisions allowing direct input of the number of miles to
be included when valuing avoided fuel outlays in the models used to
estimate impacts on the total sales of new vehicles and the scrappage
of used vehicles;
Expanded model output reporting to include all estimates
(for this analysis) of the social cost of carbon dioxide emissions
(i.e., the SCC) when reporting total and net benefits to society;
Procedures to calculate and report the value of miles
reallocated between new and used vehicles (when holding overall travel
demand before accounting for the rebound effect constant between
regulatory alternatives);
Adjustments to reduce exclude finance costs from reported
incremental costs to consumers, and reduce reported insurance costs by
20 percent (to prevent double-counting of the costs to replace totaled
vehicles); and
Revisions to allow direct specification of total VMT even
in years for which the CAFE Model estimates new vehicle sales (in
particular, for this analysis, 2021, to account for VMT recovering
rapidly following the decline in the early months of the COVID-19
pandemic.
The TSD accompanying this document elaborates on these changes to
the CAFE Model, as well as changes to input to the model for this
analysis.
NHTSA underscores that this analysis exercises the CAFE Model in a
manner that explicitly accounts for the fact that in producing a single
fleet of vehicles for sale in the United States, manufacturers face the
combination of CAFE standards, EPA CO2 standards,
[[Page 25748]]
and ZEV mandates, and for five manufacturers, the voluntary agreement
with California to more stringent GHG reduction requirements (also
applicable to these manufacturers' total production for the U.S.
market) through MY 2026. These regulations and contracts have important
structural and other differences that affect the strategy a
manufacturer could use to comply with each of the above.
As explained, the analysis is designed to reflect a number of
statutory and regulatory requirements applicable to CAFE and tailpipe
CO2 standard-setting. EPCA contains a number of requirements
governing the scope and nature of CAFE standard setting. Among these,
some have been in place since EPCA was first signed into law in 1975,
and some were added in 2007, when Congress passed EISA and amended
EPCA. EPCA/EISA requirements regarding the technical characteristics of
CAFE standards and the analysis thereof include, but are not limited
to, the following, and the analysis reflects these requirements as
summarized:
Corporate Average Standards: Section 32902 of 49 U.S.C. requires
standards that apply to the average fuel economy levels achieved by
each corporation's fleets of vehicles produced for sale in the U.S.\53\
The CAFE Model calculates the CAFE and CO2 levels of each
manufacturer's fleets based on estimated production volumes and
characteristics, including fuel economy levels, of distinct vehicle
models that could be produced for sale in the U.S.
---------------------------------------------------------------------------
\53\ This differs from safety standards and traditional
emissions standards, which apply separately to each vehicle. For
example, every vehicle produced for sale in the U.S. must, on its
own, meet all applicable Federal motor vehicle safety standards
(FMVSS), but no vehicle produced for sale must, on its own, meet
Federal fuel economy standards. Rather, each manufacturer is
required to produce a mix of vehicles that, taken together, achieve
an average fuel economy level no less than the applicable minimum
level.
---------------------------------------------------------------------------
Separate Standards for Passenger Cars and Light Trucks: Section
32902 of 49 U.S.C. requires the Secretary of Transportation to set CAFE
standards separately for passenger cars and light trucks. The CAFE
Model accounts separately for passenger cars and light trucks when it
analyzes CAFE or CO2 standards, including differentiated
standards and compliance.
Attribute-Based Standards: Section 32902 of 49 U.S.C. requires the
Secretary of Transportation to define CAFE standards as mathematical
functions expressed in terms of one or more vehicle attributes related
to fuel economy. This means that for a given manufacturer's fleet of
vehicles produced for sale in the U.S. in a given regulatory class and
model year, the applicable minimum CAFE requirement (i.e., the
numerical value of the requirement) is computed based on the applicable
mathematical function, and the mix and attributes of vehicles in the
manufacturer's fleet. The CAFE Model accounts for such functions and
vehicle attributes explicitly.
Separately Defined Standards for Each Model Year: Section 32902 of
49 U.S.C. requires the Secretary to set CAFE standards (separately for
passenger cars and light trucks \54\) at the maximum feasible levels in
each model year. The CAFE Model represents each model year explicitly,
and accounts for the production relationships between model years.\55\
---------------------------------------------------------------------------
\54\ Chapter 329 of title 49 of the U.S. Code uses the term
``non-passenger automobiles,'' while NHTSA uses the term ``light
trucks'' in its CAFE regulations. The terms' meanings are identical.
\55\ For example, a new engine first applied to given vehicle
model/configuration in MY 2020 will most likely be ``carried
forward'' to MY 2021 of that same vehicle model/configuration, in
order to reflect the fact that manufacturers do not apply brand-new
engines to a given vehicle model every single year. The CAFE Model
is designed to account for these real-world factors.
---------------------------------------------------------------------------
Separate Compliance for Domestic and Imported Passenger Car Fleets:
Section 32904 of 49 U.S.C. requires the EPA Administrator to determine
CAFE compliance separately for each manufacturers' fleets of domestic
passenger cars and imported passenger cars, which manufacturers must
consider as they decide how to improve the fuel economy of their
passenger car fleets. The CAFE Model accounts explicitly for this
requirement when simulating manufacturers' potential responses to CAFE
standards, and combines any given manufacturer's domestic and imported
cars into a single fleet when simulating that manufacturer's potential
response to CO2 standards (because EPA does not have
separate standards for domestic and imported passenger cars).
Minimum CAFE Standards for Domestic Passenger Car Fleets: Section
32902 of 49 U.S.C. requires that domestic passenger car fleets meet a
minimum standard, which is calculated as 92 percent of the industry-
wide average level required under the applicable attribute-based CAFE
standard, as projected by the Secretary at the time the standard is
promulgated. The CAFE Model accounts explicitly for this requirement
for CAFE standards and sets this requirement aside for CO2
standards.
Civil Penalties for Noncompliance: Section 32912 of 49 U.S.C. (and
implementing regulations) prescribes a rate (in dollars per tenth of a
mpg) at which the Secretary is to levy civil penalties if a
manufacturer fails to comply with a CAFE standard for a given fleet in
a given model year, after considering available credits. Some
manufacturers have historically demonstrated a willingness to pay civil
penalties rather than achieving full numerical compliance across all
fleets. The CAFE Model calculates civil penalties (adjusted for
inflation) for CAFE shortfalls and provides means to estimate that a
manufacturer might stop adding fuel-saving technologies once continuing
to do so would be effectively more ``expensive'' (after accounting for
fuel prices and buyers' willingness to pay for fuel economy) than
paying civil penalties. The CAFE Model does not allow civil penalty
payment as an option for CO2 standards.
Dual-Fueled and Dedicated Alternative Fuel Vehicles: For purposes
of calculating CAFE levels used to determine compliance, 49 U.S.C.
32905 and 32906 specify methods for calculating the fuel economy levels
of vehicles operating on alternative fuels to gasoline or diesel
through MY 2020. After MY 2020, methods for calculating alternative
fuel vehicle (AFV) fuel economy are governed by regulation. The CAFE
Model is able to account for these requirements explicitly for each
vehicle model. However, 49 U.S.C. 32902 prohibits consideration of the
fuel economy of dedicated alternative fuel vehicle (AFV) models when
NHTSA determines what levels of CAFE standards are maximum feasible.
The CAFE Model therefore has an option to be run in a manner that
excludes the additional application of dedicated AFV technologies in
model years for which maximum feasible standards are under
consideration. As allowed under NEPA for analysis appearing in EISs
informing decisions regarding CAFE standards, the CAFE Model can also
be run without this analytical constraint. The CAFE Model does account
for dual- and alternative fuel vehicles when simulating manufacturers'
potential responses to CO2 standards. For natural gas
vehicles, both dedicated and dual-fueled, EPA has a multiplier of 2.0
for MY 2022.\56\
---------------------------------------------------------------------------
\56\ That said, the CAFE Model reflects the EPA regulatory
flexibilities in place when the NHTSA began work on this rulemaking
to reconsider CAFE standards previously issued for MYs 2024-2026,
including a multiplier of 2.0 for natural gas vehicles, both
dedicated and dual-fueled, for MYs 2022-2026, although EPA's recent
final rule eliminated this multiplier after MY 2022. As explained
elsewhere in this preamble, the effect of this particular difference
between the modeling and EPA's final requirements is not
significant, given the lack of NGVs in the analysis.
---------------------------------------------------------------------------
[[Page 25749]]
ZEV Mandates: The CAFE Model can simulate manufacturers' compliance
with ZEV mandates applicable in California and ``Section 177'' \57\
states. The approach involves identifying specific vehicle model/
configurations that could be replaced with PHEVs or BEVs, and
immediately making these changes in each model year, before beginning
to consider the potential that other technologies could be applied
toward compliance with CAFE or CO2 standards.
---------------------------------------------------------------------------
\57\ The term ``Section 177'' states refers to states which have
elected to adopt California's standards in lieu of Federal
requirements, as allowed under Section 177 of the CAA.
---------------------------------------------------------------------------
Creation and Use of Compliance Credits: Section 32903 of 49 U.S.C.
provides that manufacturers may earn CAFE ``credits'' by achieving a
CAFE level beyond that required of a given fleet in a given model year,
and specifies how these credits may be used to offset the amount by
which a different fleet falls short of its corresponding requirement.
These provisions allow credits to be ``carried forward'' and ``carried
back'' between model years, transferred between regulated classes
(domestic passenger cars, imported passenger cars, and light trucks),
and traded between manufacturers. However, credit use is also subject
to specific statutory limits. For example, CAFE compliance credits can
be carried forward a maximum of five model years and carried back a
maximum of three model years. Also, EPCA/EISA caps the amount of credit
that can be transferred between passenger car and light truck fleets
and prohibits manufacturers from applying traded or transferred credits
to offset a failure to achieve the applicable minimum standard for
domestic passenger cars. The CAFE Model explicitly simulates
manufacturers' potential use of credits carried forward from prior
model years or transferred from other fleets.\58\ Section 32902 of 49
U.S.C. prohibits consideration of manufacturers' potential application
of CAFE compliance credits when setting maximum feasible CAFE
standards. The CAFE Model can be operated in a manner that excludes the
application of CAFE credits for a given model year under consideration
for standard setting. For modeling CO2 standards, the CAFE Model does
not limit transfers. Insofar as the CAFE Model can be exercised in a
manner that simulates trading of CO2 compliance credits, such
simulations treat trading as unlimited.\59\
---------------------------------------------------------------------------
\58\ The CAFE Model does not explicitly simulate the potential
that manufacturers would carry CAFE or CO2 credits back
(i.e., borrow) from future model years, or acquire and use CAFE
compliance credits from other manufacturers. At the same time,
because EPA has currently elected not to limit credit trading, the
CAFE Model can be exercised in a manner that simulates unlimited
(a.k.a. ``perfect'') CO2 compliance credit trading
throughout the industry (or, potentially, within discrete trading
``blocs''). NHTSA believes there is significant uncertainty in how
manufacturers may choose to employ these particular flexibilities in
the future: for example, while it is reasonably foreseeable that a
manufacturer who over-complies in one year may ``coast'' through
several subsequent years relying on those credits rather than
continuing to make technology improvements, it is harder to assume
with confidence that manufacturers will rely on future technology
investments to offset prior-year shortfalls, or whether/how
manufacturers will trade credits with market competitors rather than
making their own technology investments. Historically, carry-back
and trading have been much less utilized than carry-forward, for a
variety of reasons including higher risk and preference not to `pay
competitors to make fuel economy improvements we should be making'
(to paraphrase one manufacturer), although NHTSA recognizes that
carry-back and trading are used more frequently when standards
increase in stringency more rapidly. Given the uncertainty just
discussed, and given also the fact that the agency has yet to
resolve some of the analytical challenges associated with simulating
use of these flexibilities, the agency considers borrowing and
trading to involve sufficient risk that it is prudent to support
this final rule with analysis that sets aside the potential that
manufacturers could come to depend widely on borrowing and trading.
While compliance costs in real life may be somewhat different from
what is modeled in this document as a result of this analytical
decision, that is broadly true no matter what, and the agency does
not believe that the difference would be so great that it would
change the policy outcome. Furthermore, a manufacturer employing a
trading strategy would presumably do so because it represents a
lower-cost compliance option. Thus, the estimates derived from this
modeling approach are likely to be conservative in this respect,
with real-world compliance costs possibly being lower.
\59\ To avoid making judgments about possible future trading
activity, the model simulates trading by combining all manufacturers
into a single entity, so that the most cost-effective choices are
made for the fleet as a whole.
---------------------------------------------------------------------------
Statutory Basis for Stringency: Section 32902 of 49 U.S.C. requires
the Secretary to set CAFE standards at the maximum feasible levels,
considering technological feasibility, economic practicability, the
need of the United States to conserve energy, and the impact of other
motor vehicle standards of the Government on fuel economy. EPCA/EISA
authorizes the Secretary to interpret these factors, and as the
Department's interpretation has evolved, NHTSA has continued to expand
and refine its qualitative and quantitative analysis to account for
these statutory factors. For example, one of the ways that economic
practicability considerations are incorporated into the analysis is
through the technology effectiveness determinations: the Autonomie
simulations reflect the agency's judgment that it would not be
economically practicable for a manufacturer to ``split'' an engine
shared among many vehicle model/configurations into myriad versions
each optimized to a single vehicle model/configuration.
National Environmental Policy Act: In addition, NEPA requires the
Secretary to issue an EIS that documents the estimated impacts of
regulatory alternatives under consideration. The Final SEIS
accompanying this final rule documents changes in emission inventories
as estimated using the CAFE Model, but also documents corresponding
estimates--based on the application of other models documented in the
Final SEIS, of impacts on the global climate, on tropospheric air
quality, and on human health.
Other Aspects of Compliance: Beyond these statutory requirements
applicable to DOT, EPA, or both are a number of specific technical
characteristics of CAFE and/or CO2 regulations that are also
relevant to the construction of this analysis. For example, EPA has
defined procedures for calculating average CO2 levels, and
has revised procedures for calculating CAFE levels, to reflect
manufacturers' application of ``off-cycle'' technologies that increase
fuel economy (and reduce CO2 emissions). Although too little
information is available to account for these provisions explicitly in
the same way that the agency has accounted for other technologies, the
CAFE Model includes and makes use of inputs reflecting the agency's
expectations regarding the extent to which manufacturers may earn such
credits, along with estimates of corresponding costs. Similarly, the
CAFE Model includes and makes use of inputs regarding credits EPA has
elected to allow manufacturers to earn toward CO2 levels
(not CAFE) based on the use of air conditioner refrigerants with lower
global warming potential (GWP), or on the application of technologies
to reduce refrigerant leakage. In addition, the CAFE Model accounts for
EPA ``multipliers'' for certain alternative fueled vehicles, based on
current regulatory provisions or on alternative approaches. Although
these are examples of regulatory provisions that arise from the
exercise of discretion rather than specific statutory mandate, they can
materially impact outcomes.
Besides the updates to the model described above, any analysis of
regulatory actions that will be implemented several years in the
future, and whose benefits and costs accrue over decades, requires a
large number of assumptions. Over such time horizons, many, if not
most, of the relevant assumptions in such an analysis are inevitably
uncertain. Each successive CAFE analysis seeks to update assumptions to
reflect better the current
[[Page 25750]]
state of the world and the best current estimates of future conditions.
A number of assumptions have been updated since the 2020 final rule
for this final rule, and some of these assumptions have been further
updated since the proposal preceding this document. As discussed below,
NHTSA has updated its ``analysis fleet'' from a MY 2017 reference to a
MY 2020 reference, updated estimates of manufacturers' compliance
credit ``holdings,'' updated fuel price projections to reflect the U.S.
Energy Information Administration's (EIA's) 2021 Annual Energy Outlook
(AEO), updated projections of GDP and related macroeconomic measures,
and updated projections of future highway travel. While NHTSA would
have made these updates as a matter of course, we note that that the
COVID-19 pandemic impacted major analytical inputs such as fuel prices,
gross domestic product (GDP), vehicle production and sales, and highway
travel. However, while NHTSA was able to further update forecasts of
GDP and related macroeconomic measures after the 2021 proposal to
reflect a more rapid economic recovery from the pandemic than
anticipated in early 2021, EIA did not publish AEO 2022 early enough
for NHTSA to include a correspondingly updated fuel price forecast in
this analysis, so this analysis retains the fuel price forecasts from
AEO 2021. E.O. 13990 required the formation of an Interagency Working
Group (IWG) on the Social Cost of Greenhouse Gases and charged this
body with updating estimates of the social costs of carbon, nitrous
oxide, and methane. As discussed in the TSD, NHTSA has followed DOT's
determination that the values developed in the IWG's interim guidance
are the most consistent with the best available science and economics
and are the most appropriate estimates to use in the analysis of this
rule. Those estimates of costs per ton of emissions (or benefits per
ton of emissions reductions) are considerably greater than those
applied in the analysis supporting the 2020 final rule. Even still, the
estimates NHTSA is now using are not able to fully quantify and
monetize a number of important categories of climate damages; because
of those omitted damages and other methodological limits, DOT believes
its values for SC-GHG are conservative underestimates. These and other
updated analytical inputs are discussed in detail in the TSD. NHTSA
addresses comments about these assumptions later in this preamble.
What is NHTSA analyzing?
As in the CAFE and CO2 rulemakings in 2010, 2012, and
2020, NHTSA is establishing attribute-based CAFE standards defined by a
mathematical function of vehicle footprint, which has observable
correlation with fuel economy. EPCA, as amended by EISA, expressly
requires that CAFE standards for passenger cars and light trucks be
based on one or more vehicle attributes related to fuel economy and be
expressed in the form of a mathematical function.\60\ Thus, the final
standards (and regulatory alternatives) take the form of fuel economy
targets expressed as functions of vehicle footprint (the product of
vehicle wheelbase and average track width) that are separate for
passenger cars and light trucks. Chapter 1.2.3 of the TSD discusses in
detail NHTSA's continued reliance on footprint as the relevant
attribute on which these standards are based.
---------------------------------------------------------------------------
\60\ 49 U.S.C. 32902(a)(3)(A).
---------------------------------------------------------------------------
Under the footprint-based standards, the function defines a fuel
economy performance target for each unique footprint combination within
a car or truck model type. Using the functions, each manufacturer thus
will have a CAFE average standard for each year that is almost
certainly unique to each of its fleets,\61\ based upon the footprints
and production volumes of the vehicle models produced by that
manufacturer. A manufacturer will have separate footprint-based
standards for cars and for trucks, consistent with 49 U.S.C. 32902(b)'s
direction that NHTSA must set separate standards for cars and for
trucks. The functions are mostly sloped, so that generally, larger
vehicles (i.e., vehicles with larger footprints) will be subject to
lower mpg targets than smaller vehicles. This is because, generally
speaking, smaller vehicles are more capable of achieving higher levels
of fuel economy, mostly because they tend not to have to work as hard
(and therefore require as much energy) to perform their driving task.
Although a manufacturer's fleet average standards could be estimated
throughout the model year based on the projected production volume of
its vehicle fleet (and are estimated as part of EPA's certification
process), the standards with which the manufacturer must comply are
determined by its final model year production figures. A manufacturer's
calculation of its fleet average standards, as well as its fleets'
average performance at the end of the model year, will thus be based on
the production-weighted average target and performance of each model in
its fleet.\62\
---------------------------------------------------------------------------
\61\ EPCA/EISA requires NHTSA and EPA to separate passenger cars
into domestic and import passenger car fleets for CAFE compliance
purposes (49 U.S.C. 32904(b)), whereas EPA combines all passenger
cars into one fleet for GHG compliance purposes.
\62\ As discussed in prior rulemakings, a manufacturer may have
some vehicle models that exceed their target and some that are below
their target. Compliance with a fleet average standard is determined
by comparing the fleet average standard (based on the production-
weighted average of the target levels for each model) with fleet
average performance (based on the production-weighted average of the
performance of each model).
---------------------------------------------------------------------------
For passenger cars, consistent with prior rulemakings, NHTSA is
defining fuel economy targets as shown in Equation III-1.
BILLING CODE 4910-59-P
[GRAPHIC] [TIFF OMITTED] TR02MY22.048
Where:
TARGETFE is the fuel economy target (in mpg) applicable to a
specific vehicle model type with a unique footprint combination,
a is a minimum fuel economy target (in mpg),
b is a maximum fuel economy target (in mpg),
c is the slope (in gallons per mile per square foot, or gpm, per
square foot) of a line relating fuel consumption (the inverse of
fuel economy) to footprint, and
d is an intercept (in gpm) of the same line.
[[Page 25751]]
Here, MIN and MAX are functions that take the minimum and maximum
values, respectively, of the set of included values. For example,
MIN[40, 35] = 35 and MAX(40, 25) = 40, such that MIN[MAX(40, 25), 35] =
35.
For the Preferred Alternative, this equation is represented
graphically as the curves in Figure III-2.
[GRAPHIC] [TIFF OMITTED] TR02MY22.049
For light trucks, also consistent with prior rulemakings, NHTSA is
defining fuel economy targets as shown in Equation III-2.
[GRAPHIC] [TIFF OMITTED] TR02MY22.050
[[Page 25752]]
Where:
TARGETFE is the fuel economy target (in mpg) applicable to a
specific vehicle model type with a unique footprint combination,
a, b, c, and d are as for passenger cars, but taking values specific
to light trucks,
e is a second minimum fuel economy target (in mpg),
f is a second maximum fuel economy target (in mpg),
g is the slope (in gpm per square foot) of a second line relating
fuel consumption (the inverse of fuel economy) to footprint, and
h is an intercept (in gpm) of the same second line.
For the Preferred Alternative, this equation is represented
graphically as the curves in Figure III-3.
[GRAPHIC] [TIFF OMITTED] TR02MY22.051
Although the general model of the target function equation is the
same for each vehicle category (passenger cars and light trucks) and
each model year, the parameters of the function equation differ for
cars and trucks. The actual parameters for both the Preferred
Alternative and the other regulatory alternatives are presented in
Section IV.B of this preamble.
As has been the case since NHTSA began establishing attribute-based
standards, no vehicle need meet the specific applicable fuel economy
target, because compliance with CAFE standards is determined based on
corporate average fuel economy. In this respect, CAFE standards are
unlike, for example, Federal Motor Vehicle Safety Standards (FMVSS) and
certain vehicle criteria pollutant emissions standards where each car
must meet the requirements. CAFE standards apply to the average fuel
economy levels achieved by manufacturers' entire fleets of vehicles
produced for sale in the U.S. Safety standards apply on a vehicle-by-
vehicle basis, such that every single vehicle produced for sale in the
U.S. must, on its own, comply with minimum FMVSS. When first mandating
CAFE standards in the 1970s, Congress specified a more flexible
averaging-based approach that inherently allows some vehicles to
``under comply'' (i.e., fall short of the overall flat standard, or
fall short of their target under attribute-based standards), as long as
a manufacturer's overall fleet is in compliance.
[[Page 25753]]
The required CAFE level applicable to a given fleet in a given
model year is determined by calculating the production-weighted
harmonic average of fuel economy targets applicable to specific vehicle
model configurations in the fleet, as shown in Equation III-3.
[GRAPHIC] [TIFF OMITTED] TR02MY22.052
BILLING CODE 4910-59-C
Where:
CAFErequired is the CAFE level the fleet is required to achieve,
i refers to specific vehicle model/configurations in the fleet,
PRODUCTIONi is the number of model configuration i produced for sale
in the U.S., and
TARGETFE,I is the fuel economy target (as defined above) for model
configuration i.
Chapter 1 of the TSD describes the use of attribute-based
standards, generally, and explains the specific decision, in past rules
and for the current rule, to continue to use vehicle footprint as the
attribute over which to vary stringency. That chapter also discusses
the policy in selecting the specific mathematical function; the
methodologies used to develop the current attribute-based standards;
and methodologies previously used to reconsider the mathematical
function for CAFE standards. NHTSA refers readers to the TSD for a full
discussion of these topics.
Several commenters supported the continued use of footprint as the
attribute on which to base fuel economy standards. Consumer
Reports,\63\ Alliance for Automotive Innovation (Auto Innovators),\64\
the Aluminum Association,\65\ and National Automobile Dealers
Association (NADA) \66\ all agreed that footprint-based standards
continue to incentivize improvements in fuel economy across all
companies and across all market segments/vehicle classes. Auto
Innovators pointed to the most recent EPA Trends Report as indicating
that any change in average vehicle footprint has been minimal at the
industry level, implying that footprint-based standards are not leading
to ``gaming'' by manufacturers seeking a less-stringent standard by
increasing their vehicles' footprints.\67\ The Aluminum Association
suggested that footprint-based standards could be beneficial for
safety, because they incentivize weight reduction in larger footprint
vehicles, which make up an increasing portion of the fleet.\68\ NADA
\69\ and International Union, United Automobile, Aerospace &
Agricultural Implement Workers of America (UAW) \70\ both stated that
footprint-based standards supported manufacturers continuing to provide
a wide range of vehicles from which consumers could choose, with UAW
stating that ``[s]imply put, to do otherwise undermines domestic
manufacturing, workers' living standards, and communities well-being.
All vehicles do not have the same function and surely our rules need to
continue to reflect this reality.'' \71\
---------------------------------------------------------------------------
\63\ Consumer Reports, Docket No. NHTSA-2021-0053-1576-A9, at p.
7.
\64\ Auto Innovators, Docket No. NHTSA-2021-0053-1492, at p. 47.
\65\ The Aluminum Association (Aluminum Association), Docket No.
NHTSA-2021-0053-1518, at p. 3; Arconic Corporation (Arconic), Docket
No. NHTSA-2021-0053-1560, at p. 2 (Arconic, an individual aluminum
producer, also supported footprint-based standards).
\66\ NADA, Docket No. NHTSA-2021-0053-1471, at p. 3.
\67\ Auto Innovators, at p. 48.
\68\ Aluminum Association, at p. 3.
\69\ NADA, at p. 3.
\70\ UAW, Docket No. NHTSA-2021-0053-0931, at p. 2.
\71\ UAW, at p. 4.
---------------------------------------------------------------------------
One citizen commenter, Doug Peterson (Peter Douglas), objected to
the use of footprint as the attribute on which to base fuel economy
standards, stating that a consequence of using footprint is that
``[w]asteful models are simply compensated for by more efficient models
that outperform their footprint targets, and this will become a huge
problem as more and more ZEVs enter the marketplace.'' \72\ Mr. Douglas
further commented that discouraging vehicle downsizing (as footprint-
based standards can do) was an inappropriate policy goal, because
downsizing can be a good way to reduce fuel consumption and the current
upsizing trend in the fleet is not mitigated by footprint-based
standards. He also commented that the safety concern that footprint-
based standards can address is in fact misplaced, because ``[l]arge
vehicles provide safety benefits to their occupants at the expense of
people occupying small vehicles.'' \73\
---------------------------------------------------------------------------
\72\ Peter Douglas, Docket No. NHTSA-2021-0053-0085, at pp. 12-
13, p. 19.
\73\ Id.
---------------------------------------------------------------------------
NHTSA appreciates these comments but is continuing to rely on
footprint as the attribute for the final standards for MYs 2024-2026.
NHTSA notes that the first issue that Mr. Douglas raised is due to the
fact that the standards are, by law, corporate average standards, and
that ``wasteful models [being] compensated for by more efficient
models'' is difficult to avoid when standards are corporate averages--
by their nature, they enable averaging across a manufacturer's fleet.
The comments from the Aluminum Association comments, Auto Innovators,
and Mr. Douglas' further comments on the topic of footprint seem to
address one another. As Auto Innovators notes, the most recent EPA
Trends Report appears to suggest that, on average, vehicle upsizing has
been minimal at the industry (fleet) level. While footprint may not
encourage vehicle downsizing, it does reward vehicle downweighting,
which NHTSA typically refers to as ``mass reduction.'' A lighter
vehicle saves fuel compared to a heavier vehicle of the same footprint,
and thus performs better against its footprint target. NHTSA addresses
safety comments in Section V of this preamble.
While Chapter 1 of the TSD explains why the final standards for MYs
2024-2026 continue to be footprint-based, the question has arisen
periodically of whether NHTSA should instead consider multi-attribute
standards, such as those that also depend on weight, torque, power,
towing capability, off-road capability, or a combination of such
attributes. To date, every time NHTSA has considered options for which
attribute(s) to select, the agency has concluded that a properly
designed footprint-based approach provides the best means of achieving
the basic policy goals (i.e., by increasing the likelihood of improved
fuel economy across the
[[Page 25754]]
entire fleet of vehicles, as noted by commenters) involved in applying
an attribute-based standard. At the same time, footprint-based
standards need also to be structured in a way that furthers the energy
and environmental policy goals of EPCA without creating inappropriate
incentives to increase vehicle size in ways that could increase fuel
consumption or compromise safety. That said, as NHTSA moves forward
with the CAFE program, and continues to refine our understanding of the
light-duty vehicle market and trends in vehicle and highway safety,
NHTSA will also continue to revisit whether other approaches (or other
ways of applying the same basic approaches) could provide better means
of achieving policy goals.
For example, in the 2021 NAS Report, the committee recommended that
if Congress does not act to remove the prohibition at 49 U.S.C.
32902(h) on considering the fuel economy of dedicated alternative fuel
vehicles (like BEVs) in determining maximum feasible CAFE standards,
then NHTSA should account for the fuel economy benefits of ZEVs by
``setting the standard as a function of a second attribute in addition
to footprint--for example, the expected market share of ZEVs in the
total U.S. fleet of new light-duty vehicles--such that the standards
increase as the share of ZEVs in the total U.S. fleet increases.'' \74\
DOE seconded this suggestion in its comments during interagency review
of the proposal. NHTSA sought comment on whether and how NHTSA might
consider adding electrification as an attribute on which to base CAFE
standards, and specifically on the NAS committee recommendation.
---------------------------------------------------------------------------
\74\ 2021 NAS Report, at Summary Recommendation p. 5.
---------------------------------------------------------------------------
Two electric vehicle manufacturers supported the addition of
electrification as an attribute on which fuel economy standards could
be based. Lucid USA, Inc. (Lucid) stated that, in setting standards
based on electrification as well as footprint, NHTSA should ``consider
the battery efficiency of the electric vehicles manufactured by each
automaker, as well as the market penetration of electric vehicles in
the fleet.'' \75\ Rivian Automotive, LLC (Rivian) stated that such
``[a]pproaches . . . merit further study and eventual implementation.''
\76\ With regard to the timing of making such a change, a question on
which NHTSA specifically sought comment, Rivian commented that ``[i]t
is likely infeasible and inappropriate to implement such a change in
time for any of the model years subject to this rulemaking, but Rivian
believes development, review, and implementation of a newly conceived
multi-attribute function could take effect in the second half of this
decade, coinciding with a post-MY 2027 rule, and provide industry with
appropriate lead-time given typical product development lifecycles.''
\77\
---------------------------------------------------------------------------
\75\ Lucid, Docket No. NHTSA-2021-0053-1584, at p. 5.
\76\ Rivian, Docket No. NHTSA-2021-0053-1562, at p. 5.
\77\ Id.
---------------------------------------------------------------------------
Other commenters disagreed with adding electrification as an
attribute. Several opined that adding electrification as an attribute
seemed impermissible under 49 U.S.C. 32902(h).\78\ Auto Innovators
argued that it could create battery supply chain risks as an unintended
consequence, and that ``. . . including electrification as a fuel
economy attribute could be solidifying a dependence on foreign supply
chains that might not be reliable or have shared interests with our
country.'' \79\ American Honda Motor Co., Inc. (Honda) \80\ and Kia
Corporation (Kia) \81\ also raised the possibility of unintended
consequences and externalities. Kia further suggested that ``[i]n the
same manner that the footprint curves include many of the weight,
technology cost, and engineering analyses that go in to bringing these
vehicles online, electrification would need to have similar
considerations accounted for in the modeling assumptions,'' \82\ while
Honda stated that the agency should provide ``more than a full product
cycle (5-6 year[s]) of lead time'' to give industry time to plan for
any changes.\83\ Auto Innovators commented that it could be permissible
to limit consideration of electrification to HEVs, but ``[t]he existing
approach with footprint-based curves does not need to be modified if
one simply wants to require a more efficient gasoline-powered fleet--
whether through increased electrification or some other means.'' \84\
Jaguar Land Rover NA, LLC (JLR) offered a similar comment.\85\
---------------------------------------------------------------------------
\78\ Auto Innovators, at 48; Stellantis, Docket No. NHTSA-2021-
0053-1527, at 12; NADA, at p. 4; Valero Energy Corporation (Valero),
Docket No. NHTSA-2021-0053-1541, at pp. 3-4; Peter Douglas, at p.
25.
\79\ Auto Innovators, at p. 50.
\80\ Honda, Docket No. NHTSA-2021-0053-1501, at p. 4.
\81\ Kia, Docket No. NHTSA-2021-0053-1525, at p. 10.
\82\ Id.
\83\ Honda, at p. 4.
\84\ Auto Innovators, at p. 50.
\85\ JLR, Docket No. NHTSA-2021-0053-1505, at p. 4.
---------------------------------------------------------------------------
Stellantis commented that ``the `percent of work' metric as
ultimately applied in the proposal is a fleet level of electrification
selected as a policy goal rather than an attribute of a particular
vehicle (like footprint) as intended by the statute.'' \86\ NADA argued
that ``[f]leet-wide standards should be technologically neutral and set
at levels that are achievable without ZEVs so as not to penalize those
OEMs (and their dealers) that choose not to aggressively develop,
produce, and push ZEVs to market.'' \87\ And finally, Securing
America's Future Energy commented that adding electrification as an
attribute just makes the program more complicated, and NHTSA should be
looking for ways to simplify it instead, perhaps via a legislative
solution.\88\
---------------------------------------------------------------------------
\86\ Stellantis, at p. 12.
\87\ NADA, at pp. 3-4.
\88\ Securing America's Future Energy, Docket No. NHTSA-2021-
0053-1513, at pp. 18-19.
---------------------------------------------------------------------------
As explained above, for this final rule, NHTSA is continuing to
base the MY 2024-2026 standards on footprint. NHTSA is not adding
electrification as an attribute at this time, based in part on comments
that raised concerns with how to implement such an approach
practically, in a way that would further EPCA's overarching goal of
energy conservation, while providing industry with appropriate lead
time to make changes to their fleet. NHTSA is also mindful of
introducing further uncertainty to the standards during this time of
rapid change in the stringency of the standards. Therefore, while NHTSA
agrees with comments suggesting that the recommendation from the NAS
committee merits further consideration, NHTSA also agrees with other
commenters who suggested that this rulemaking is not the proper one in
which to implement such a change, given the available lead time for
manufacturers to adjust their compliance approaches.
C. What inputs does the compliance analysis require?
The CAFE Model applies various technologies to different vehicle
models in each manufacturer's product line to simulate how each
manufacturer might make progress toward compliance with the specified
standard. Subject to a variety of user-controlled constraints, the
model applies technologies based on their relative cost-effectiveness,
as determined by several input assumptions regarding the cost and
effectiveness of each technology, the cost of compliance (determined by
the change in CAFE or CO2 credits, CAFE-related civil
penalties, or value of CO2 credits, depending on the
compliance
[[Page 25755]]
program being evaluated), and the value of avoided fuel expenses. For a
given manufacturer, the compliance simulation algorithm applies
technologies either until the manufacturer runs out of cost-effective
technologies,\89\ until the manufacturer exhausts all available
technologies, or, if the manufacturer is assumed to be willing to pay
civil penalties or acquire credits from another manufacturer, until
paying civil penalties or purchasing credits becomes more cost-
effective than increasing vehicle fuel economy. At this stage, the
system assigns an incurred technology cost and updated fuel economy to
each vehicle model, as well as any civil penalties incurred/credits
purchased by each manufacturer. This compliance simulation process is
repeated for each model year included in the study period (through MY
2050 in this analysis).
---------------------------------------------------------------------------
\89\ Generally, the model considers a technology cost-effective
if it pays for itself in fuel savings within a ``payback period''
specified as a model input (for this analysis, 30 months). Depending
on the settings applied, the model can continue to apply
technologies that are not cost-effective rather than choosing other
compliance options; if it does so, it will apply those additional
technologies in order of cost-effectiveness (i.e., most cost-
effective first).
---------------------------------------------------------------------------
At the conclusion of the compliance simulation for a given
regulatory scenario, the system transitions between compliance
simulation and effects calculations. This is the point where the system
produces a full representation of the registered light-duty vehicle
population in the United States. The CAFE Model then uses this fleet to
generate estimates of the following (for each model year and calendar
year included in the analysis): Lifetime travel, fuel consumption,
carbon dioxide and criteria pollutant emissions, the magnitude of
various economic externalities related to vehicular travel (e.g.,
congestion and noise), and energy consumption (e.g., the economic costs
of short-term increases in petroleum prices, or social damages
associated with GHG emissions). The system then uses these estimates to
measure the benefits and costs associated with each regulatory
alternative (relative to the No-Action Alternative).
To perform this analysis, the CAFE Model uses millions of data
points contained in several input files that have been populated by
engineers, economists, and safety and environmental program analysts at
both NHTSA and the DOT's Volpe National Transportations Systems Center
(Volpe). In addition, some of the input data come from modeling and
simulation analysis performed by experts at Argonne National Laboratory
using their Autonomie full vehicle simulation model and BatPaC battery
cost model. Other inputs are derived from other models, such as the
U.S. Energy Information Administration's (EIA's) National Energy
Modeling System (NEMS), Argonne's ``GREET'' fuel-cycle emissions
analysis model, and U.S. EPA's ``MOVES'' vehicle emissions analysis
model. As NHTSA and Volpe are both organizations within DOT, we use DOT
throughout these sections to refer to the collaborative work performed
for this analysis.
This section and Section III.D describe the inputs that the
compliance simulation requires, including an in-depth discussion of the
technologies used in the analysis, how they are defined in the CAFE
Model, how they are characterized for vehicles that already exist in
the market, and how they can be applied to realistically simulate
manufacturers' decisions, their effectiveness, and their cost. The
inputs and analyses for the effects calculations, including economic,
safety, and environmental effects, are discussed later in Sections
III.C through III.H.
1. Overview of Inputs to the Analysis
As discussed above, the current analysis involves estimating four
major swaths of effects. First, the analysis estimates how the
application of various combinations of technologies could impact
vehicles' costs and fuel economy levels (and CO2 emission
rates). Second, the analysis estimates how vehicle manufacturers might
respond to standards by adding fuel-saving technologies to new
vehicles. Third, the analysis estimates how changes in new vehicles
might impact vehicle sales and operation. Finally, the analysis
estimates how the combination of these changes might impact national-
scale energy consumption, emissions, highway safety, and public health.
There are several CAFE Model input files important to the
discussion of these first two steps, and these input files are
discussed in detail later in this section and in Section III.D. The
Market Data file contains the detailed description of the vehicle
models and model configurations each manufacturer produces for sale in
the United States. The file also contains a range of other inputs that,
though not specific to individual vehicle models, may be specific to
individual manufacturers. The Technologies file identifies about six
dozen technologies to be included in the analysis, indicates when and
how widely each technology can be applied to specific types of
vehicles, provides most of the inputs involved in estimating what costs
will be incurred, and provides some of the inputs involved in
estimating impacts on vehicle fuel consumption and weight.
The CAFE Model also makes use of databases of estimates of fuel
consumption impacts and, as applicable, battery costs for different
combinations of fuel-saving technologies.\90\ These databases are
termed the FE1 and FE2 Adjustments databases (the main database and the
database specific to plug-in hybrid electric vehicles, applicable to
those vehicles' operation on electricity) and the Battery Costs
database. DOT developed these databases using a large set of full
vehicle and accompanying battery cost model simulations developed by
Argonne National Laboratory. The Argonne simulation outputs, battery
costs, and other reference materials are also discussed in the
following sections.\91\
---------------------------------------------------------------------------
\90\ To be used as files provided separately from the model and
loaded every time the model is executed, these databases are
prohibitively large, spanning more than a million records and more
than half a gigabyte. To conserve memory and speed model operation,
DOT has integrated the databases into the CAFE Model executable
file. When the model is run, however, the databases are extracted
and placed in an accessible location on the user's disk drive.
\91\ The Argonne workbooks included in the docket for this
notice include 10 databases that contain the outputs of the
Autonomie full vehicle simulations, two summary workbooks of
assumptions used for the full vehicle simulations, a data
dictionary, and the lookup tables for battery costs generated using
the BatPaC battery cost model.
---------------------------------------------------------------------------
The following discussion in this section and in Section III.D
expands on the inputs used in the compliance analysis. Further detail
is included in Chapters 2 and 3 of the TSD accompanying this notice,
and all input values relevant to the compliance analysis can be seen in
the Market Data, Technologies, fuel consumption and battery cost
database files, and Argonne summary files included in the docket for
this notice. As previously mentioned, other model input files underlie
the effects analysis, and these are discussed in detail in Sections
III.C through III.H.
2. The Market Data File
The Market Data file contains the detailed description of the
vehicle models and model configurations each manufacturer produces for
sale in the U.S. This snapshot of the recent light duty vehicle market,
termed the analysis fleet, or baseline fleet, is the starting point for
the evaluation of different stringency levels for future fuel economy
standards. The analysis fleet provides a reference from which to
project how manufacturers could apply additional technologies to
vehicles to
[[Page 25756]]
cost-effectively improve vehicle fuel economy, in response to
regulatory action and market conditions.\92\ For this analysis, the MY
2020 light duty fleet was selected as the baseline for further
evaluation of the effects of different fuel economy standards. The
Market Data file also contains a range of other inputs that, though not
specific to individual vehicle models, may be specific to individual
manufacturers.
---------------------------------------------------------------------------
\92\ The CAFE Model does not generate compliance paths a
manufacturer should, must, or will deploy. It is intended as a tool
to demonstrate a compliance pathway a manufacturer could choose. It
is almost certain all manufacturers will make compliance choices
differing from those projected by the CAFE Model.
---------------------------------------------------------------------------
The Market Data file is an Excel spreadsheet that contains five
worksheets. Three worksheets, the Vehicles worksheet, Engines
worksheet, and Transmissions worksheet, characterize the baseline fleet
for this analysis. The three worksheets contain a characterization of
every vehicle sold in MY 2020 and their relevant technology content,
including the engines and transmissions that a manufacturer uses in its
vehicle platforms and how those technologies are shared across
platforms. In addition, the Vehicles worksheet includes baseline
economic and safety inputs linked to each vehicle that allow the CAFE
Model to estimate economic and safety impacts resulting from any
simulated compliance pathway. The remaining two worksheets, the
Manufacturers worksheet and Credits and Adjustments worksheet, include
baseline compliance positions for each manufacturer, including each
manufacturer's starting CAFE credit banks and whether the manufacturer
is willing to pay civil penalties for noncompliance with CAFE
standards, among other inputs.
New inputs have been added for this analysis in the Vehicles
worksheet and Manufacturers worksheet. The new inputs indicate which
vehicles a manufacturer may reasonably be expected to convert to a zero
emissions vehicle (ZEV) at first redesign opportunity, to comply with
several states' ZEV program provisions. The new inputs also indicate if
a manufacturer has entered into an agreement with California to achieve
more stringent GHG emissions reductions targets than those promulgated
in the 2020 final rule.
The following sections discuss how we built the Market Data file,
including characterizing vehicles sold in MY 2020 and their technology
content, and baseline safety, economic, and manufacturer compliance
positions. A detailed discussion of the Market Data file development
process is in TSD Chapter 2.2.
(a) Characterizing Vehicles and Their Technology Content
The Market Data file integrates information from many sources,
including manufacturer compliance submissions, publicly available
information, and confidential business information. At times, DOT must
populate inputs using analyst judgment, either because information is
still incomplete or confidential, or because the information does not
yet exist.\93\ For this analysis DOT uses mid-MY 2020 compliance data
as the basis of the analysis fleet. The compliance data are
supplemented for each vehicle nameplate with manufacturer specification
sheets, usually from the manufacturer media website, or from online
marketing brochures.\94\ For additional information about how
specification sheets inform MY 2020 vehicle technology assignments, see
the technology specific assignments sections in Section III.D.
---------------------------------------------------------------------------
\93\ Forward looking refresh/redesign cycles are one example of
when analyst judgement is necessary.
\94\ The catalogue of reference specification sheets (broken
down by manufacturer, by nameplate) used to populate information in
the Market Data file is available in the docket.
---------------------------------------------------------------------------
DOT uses the mid-MY 2020 compliance data to create a row on the
Vehicles worksheet in the Market Data file for each vehicle (or vehicle
variant \95\) that lists a certification fuel economy, sales volume,
regulatory class, and footprint. DOT identifies which combination of
modeled technologies reasonably represents the fuel saving technologies
already on each vehicle, and assigns those technologies to each
vehicle, either on the Vehicles worksheet, the Engines worksheet, or
the Transmissions worksheet. The fuel saving technologies considered in
this analysis are listed in Table III-1.
---------------------------------------------------------------------------
\95\ The Market Data file often includes a few rows for vehicles
that may have identical certification fuel economies, regulatory
classes, and footprints (with compliance sales volumes divided out
among rows), because other pieces of information used in the CAFE
Model may be dissimilar. For instance, in the reference materials
used to create the Market Data file, for a nameplate curb weight may
vary by trim level (with premium trim levels often weighing more on
account of additional equipment on the vehicle), or a manufacturer
may provide consumers the option to purchase a larger fuel tank size
for their vehicle. These pieces of information may not impact the
observed compliance position directly, but curb weight (in relation
to other vehicle attributes) is important to assess mass reduction
technology already used on the vehicle, and fuel tank size is
directly relevant to saving time at the gas pump, which the CAFE
Model uses when calculating the value of avoided time spent
refueling.
---------------------------------------------------------------------------
BILLING CODE 4910-59-P
[[Page 25757]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.053
[[Page 25758]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.054
[[Page 25759]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.055
BILLING CODE 4910-59-P
[[Page 25760]]
For additional information on the characterization of these
technologies (including the cost, prevalence in the 2020 fleet,
effectiveness estimates, and considerations for their adoption) see the
appropriate technology section in Section III.D or TSD Chapter 3.
DOT also assigns each vehicle a technology class. The CAFE Model
uses the technology class (and engine class, discussed below) in the
Market Data file to reference the most relevant technology costs for
each vehicle, and fuel saving technology combinations. We assign each
vehicle in the fleet a technology class using a two-step algorithm that
takes into account key characteristics of vehicles in the fleet
compared to the baseline characteristics of each technology class.\96\
As discussed further in Section III.C.4.b), there are ten technology
classes used in the CAFE analysis that span five vehicle types and two
performance variants. The technology class algorithm and assignment
process is discussed in more detail in TSD Chapter 2.4.2.
---------------------------------------------------------------------------
\96\ Baseline 0 to 60 mph accelerations times are assumed for
each technology class as part of the Autonomie full vehicle
simulations. DOT calculates class baseline curb weights and
footprints by averaging the curb weights and footprints of vehicles
within each technology class as assigned in previous analyses.
---------------------------------------------------------------------------
We also assign each vehicle an engine technology class so that the
CAFE Model can reference the powertrain costs in the Technologies file
that most reasonably align with the observed vehicle. DOT assigns
engine technology classes for all vehicles, including electric
vehicles. If an electric powertrain replaces an internal combustion
engine, the electric motor specifications may be different (and hence
costs may be different) depending on the capabilities of the internal
combustion engine it is replacing, and the costs in the technologies
file (on the engine tab) account for the power output and capability of
the gasoline or electric drivetrain.
Parts sharing helps manufacturers achieve economies of scale,
deploy capital efficiently, and make the most of shared research and
development expenses, while still presenting a wide array of consumer
choices to the market. The CAFE Model simulates part sharing by
implementing shared engines, shared transmissions, and shared mass
reduction platforms. Vehicles sharing a part (as recognized in the CAFE
Model), will adopt fuel saving technologies affecting that part
together. To account for parts sharing across products, vehicle model/
configurations that share engines are assigned the same engine
code,\97\ vehicle model/configurations that share transmissions have
the same transmission code, and vehicles that adopt mass reduction
technologies together share the same platform. For more information
about engine codes, transmission codes, and mass reduction platforms
see TSD Chapter 3.
---------------------------------------------------------------------------
\97\ Engines (or transmissions) may not be exactly identical, as
specifications or vehicle integration features may be different.
However, the architectures are similar enough that it is likely the
powertrain systems share R&D, tooling, and production resources in a
meaningful way.
---------------------------------------------------------------------------
Manufacturers often introduce fuel saving technologies at a major
redesign of their product or adopt technologies at minor refreshes in
between major product redesigns. To support the CAFE Model accounting
for new fuel saving technology introduction as it relates to product
lifecycle, the Market Data file includes a projection of redesign and
refresh years for each vehicle. DOT projects future redesign years and
refresh years based on the historical cadence of that vehicle's product
lifecycle. For new nameplates, DOT considers the manufacturer's
treatment of product lifecycles for past products in similar market
segments. When considering year-by-year analysis of standards, the
sizing of redesign and refresh intervals will affect projected
compliance pathways and how quickly manufacturers can respond to
standards. TSD Chapter 2.2.1.7 includes additional information about
the product design cycles assumed for this action based on historical
manufacturer product design cycles.
The Market Data file also includes information about air
conditioning (AC) and off-cycle technologies, but the information is
not currently broken out at a row level, vehicle by vehicle.\98\
Instead, historical data (and forecast projections, which are used for
analysis regardless of regulatory scenario) are listed by manufacturer,
by fleet on the Credits and Adjustments worksheet of the Market Data
file. Section III.D.8 shows model inputs specifying estimated
adjustments (all in grams/mile) for improvements to air conditioner
efficiency and other off-cycle energy consumption, and for reduced
leakage of air conditioner refrigerants with high global warming
potential (GWP). DOT estimated future values based on an expectation
that manufacturers already relying heavily on these adjustments would
continue do so, and that other manufacturers would, over time, also
approach the limits on adjustments allowed for such improvements.
---------------------------------------------------------------------------
\98\ Regulatory provisions regarding off-cycle technologies are
new, and manufacturers have only recently begun including related
detailed information in compliance reporting data. For this
analysis, though, such information was not sufficiently complete to
support a detailed representation of the application of off-cycle
technology to specific vehicle model/configurations in the MY 2020
fleet.
---------------------------------------------------------------------------
(b) Characterizing Baseline Safety, Economic, and Compliance Positions
In addition to characterizing vehicles and their technology
content, the Market Data file contains a range of other inputs that,
though not specific to individual vehicle models, may be specific to
individual manufacturers, or that characterize baseline safety or
economic information.
First, the CAFE Model considers the potential safety effect of mass
reduction technologies and crash compatibility of different vehicle
types. Mass reduction technologies lower the vehicle's curb weight,
which may improve crash compatibility and safety, or not, depending on
the type of vehicle. DOT assigns each vehicle in the Market Data file a
safety class that best aligns with the mass-size-safety analysis. This
analysis is discussed in more detail in Section III.H of this action
and TSD Chapter 7.
The CAFE Model also includes procedures to consider the direct
labor impacts of manufacturer's response to CAFE regulations,
considering the assembly location of vehicles, engines, and
transmissions, the percent U.S. content (that reflects percent U.S. and
Canada content),\99\ and the dealership employment associated with new
vehicle sales. The Market Data file therefore includes baseline labor
information, by vehicle. Sales volumes also influence total estimated
direct labor projections in the analysis.
---------------------------------------------------------------------------
\99\ Percent U.S. content was informed by the 2020 Part 583
American Automobile Labeling Act Reports, appearing on NHTSA's
website.
---------------------------------------------------------------------------
We hold the percent U.S. content constant for each vehicle row for
the duration of the analysis. In practice, this may not be the case.
Changes to trade policy and tariff policy may affect percent U.S.
content in the future. Also, some technologies may be more or less
likely to be produced in the U.S., and if that is the case, their
adoption could affect future U.S. content. NHTSA does not have data at
this time to support varying the percent U.S. content.
We also hold the labor hours projected in the Market Data file per
unit transacted at dealerships, per unit produced for final assembly,
per unit produced for engine assembly, and per unit produced for
transmission assembly constant for the duration of the analysis, and
project that the origin
[[Page 25761]]
of these activities to remain unchanged. In practice, it is reasonable
to expect that plants could move locations, or engine and transmission
technologies are replaced by another fuel saving technology (like
electric motors and fixed gear boxes) that could require a meaningfully
different amount of assembly labor hours. NHTSA does not have data at
this time to support varying labor hours projected in the Market Data
file, but we will continue to explore methods to estimate the direct
labor impacts of manufacturer's responses to CAFE standards in future
analyses.
As observed from Table III-2, manufacturers employ U.S. labor with
varying intensity. In many cases, vehicles certifying in the light
truck (LT) regulatory class have a larger percent U.S. content than
vehicles certifying in the passenger car (PC) regulatory class.
BILLING CODE 4910-59-P
[GRAPHIC] [TIFF OMITTED] TR02MY22.056
BILLING CODE 4910-59-C
Next, manufacturers may over-comply with CAFE standards and bank
so-called over compliance credits. As discussed further in Section
III.C.7, manufacturers may use these credits later, sell them to other
manufacturers, or let them expire. The CAFE Model does not explicitly
trade credits between and among manufacturers, but staff have adjusted
starting credit banks in the Market Data file to reflect trades that
are likely to happen when the simulation begins (in MY 2020).
Considering information manufacturers have reported regarding
compliance credits, and considering recent manufacturers' compliance
positions, DOT estimates manufacturers' potential use of compliance
credits in earlier model years. This aligns to an extent that
represents how manufacturers could deplete their credit banks rather
than producing high volume vehicles with fuel saving technologies in
earlier model years. This also avoids the unrealistic application of
technologies for manufacturers in early analysis years that typically
rely on credits. For a complete discussion about how these data are
collected and assigned in the Market Data file, see TSD Chapter
2.2.2.3.
---------------------------------------------------------------------------
\100\ Tesla does not have internal combustion engines, or multi-
speed transmissions, even thought they are identified as producing
engine and transmission systems in the United States in the Market
Data file.
---------------------------------------------------------------------------
The Market Data file also includes assumptions about a vehicle
manufacturer's preferences towards civil penalty payments. EPCA
requires that if a manufacturer does not achieve
[[Page 25762]]
compliance with a CAFE standard in a given model year and cannot apply
credits sufficient to cover the compliance shortfall, the manufacturer
must pay civil penalties (i.e., fines) to the Federal Government. If
inputs indicate that a manufacturer treats civil penalty payment as an
economic choice (i.e., one to be taken if doing so would be
economically preferable to applying further technology toward
compliance), the CAFE Model, when evaluating the manufacturer's
response to CAFE standards in a given model year, will apply fuel-
saving technology only up to the point beyond which doing so would be
more expensive (after subtracting the value of avoided fuel outlays)
than paying civil penalties.
For this analysis, DOT exercises the CAFE Model with inputs
treating all manufacturers as treating civil penalty payment as an
economic choice through MY 2023. While DOT expects that only
manufacturers with some history of paying civil penalties would
actually treat civil penalty payment as an acceptable option, the CAFE
Model does not currently simulate compliance credit trading between
manufacturers, and DOT expects that this treatment of civil penalty
payment will serve as a reasonable proxy for compliance credit
purchases some manufacturers might actually make through MY 2023. These
input assumptions for model years through 2023 reduce the potential
that the model will overestimate technology application in the model
years leading up to those for which the agency is finalizing new
standards. As in past CAFE rulemaking analyses (except that supporting
the 2020 final rule), DOT has treated manufacturers with some history
of civil penalty payment (i.e., BMW, Daimler, FCA, Jaguar-Land Rover,
Volvo, and Volkswagen) as continuing to treat civil penalty payment as
an acceptable option beyond MY 2023, but has treated all other
manufacturers as unwilling to do so beyond MY 2023. DOT believes it is
more accurate, as in past analyses besides the 2020 final rule, to
reflect the possibility that these historical payers of civil penalties
may continue to do so in the future.
Next, the CAFE Model uses an ``effective cost'' metric to evaluate
options to apply specific technologies to specific engines,
transmissions, and vehicle model configurations. Expressed on a $/
gallon basis, the analysis computes this metric by subtracting the
estimated values of avoided fuel outlays and civil penalties from the
corresponding technology costs, and then dividing the result by the
quantity of avoided fuel consumption. The analysis computes the value
of fuel outlays over a ``payback period'' representing the
manufacturer's expectation that the market will be willing to pay for
some portion of fuel savings achieved through higher fuel economy. Once
the model has applied enough technology to a manufacturer's fleet to
achieve compliance with CAFE standards (and CO2 standards
and ZEV mandates) in a given model year, the model will apply any
further fuel economy improvements estimated to produce a negative
effective cost (i.e., any technology applications for which avoided
fuel outlays during the payback period are larger than the
corresponding technology costs). As discussed above in Section III.A
and below in Section III.C, DOT anticipates that manufacturers are
likely to act as if the market is willing to pay for avoided fuel
outlays expected during the first 30 months of vehicle operation.
In addition, the Market Data file includes two new sets of inputs
for this analysis. In 2020, five vehicle manufacturers reached a
voluntary commitment with the state of California to improve the
emissions levels of their future nationwide fleets above levels
required by the 2020 final rule. For this analysis, compliance with
this agreement is in the baseline case for designated manufacturers.
The Market Data file contains inputs indicating whether each
manufacturer has committed to exceed Federal requirements per this
agreement.
Finally, when considering other standards that may affect fuel
economy compliance pathways, DOT includes projected zero emissions
vehicles (ZEV) that would be required for manufacturers to meet
standards in California and Section 177 states, per the waiver granted
under the Clean Air Act. To support the inclusion of the ZEV program in
the analysis, DOT identifies specific vehicle model/configurations that
could adopt BEV technology in response to the ZEV program, independent
of CAFE standards, at the first redesign opportunity. These ZEVs are
identified in the Market Data file as future BEV200s, BEV300s, or
BEV400s. Not all announced BEV nameplates appear in the MY 2020 Market
Data file; in these cases, in consultation with CARB, DOT used the
volume from a comparable vehicle in the manufacturer's Market Data file
portfolio as a proxy. The Market Data file also includes information
about the portion of each manufacturer's sales that occur in California
and Section 177 states, which is helpful for determining how many ZEV
credits each manufacturer will need to generate in the future to comply
with the ZEV program with their own portfolio in the rulemaking
timeframe. These new procedures are described in detail below and in
TSD Chapter 2.3.
3. Simulating the Zero Emissions Vehicle Program
California's Zero Emissions Vehicle (ZEV) program is one part of a
program of coordinated standards that the California Air Resources
Board (CARB) has enacted to control emissions of criteria pollutants
and greenhouse gas emissions from vehicles. The program began in 1990
with the low-emission vehicle (LEV) regulation,\101\ and has since
expanded to include eleven other states.102 103 These states
may be referred to as Section 177 states, in reference to Section 177
of the Clean Air Act's grant of authority to allow these states to
adopt California's air quality standards,\104\ but it is important to
note that not all Section 177 states have adopted the ZEV program
component.\105\ In the following discussion of the incorporation of the
ZEV program into the CAFE Model, any reference to the Section 177
states refers to those states that have adopted California's ZEV
program requirements.
---------------------------------------------------------------------------
\101\ California Air Resource Board (CARB), Zero-Emission
Vehicle Program. California Air Resources Board. https://ww2.arb.ca.gov/our-work/programs/zero-emission-vehicle-program/about. (Accessed: February 16, 2022)
\102\ Through 2020, the Section 177 states that had adopted the
ZEV program included Colorado, Connecticut, Maine, Maryland,
Massachusetts, New Jersey, New York, Oregon, Rhode Island, Vermont,
and Washington. See Vermont Department of Environmental
Conservation, Zero Emission Vehicles. https://dec.vermont.gov/air-quality/mobile-sources/zev. (Accessed: February 16, 2022)
\103\ The states of Minnesota, Nevada, and Virginia have
recently adopted ZEV standards, which will go into effect for MY
2025. As discussed in this section, reflecting these three states'
adoption of ZEV mandates would have only negligibly impacted the
agency's national-scale modeling. See Green Car Reports, Minnesota
adopts California EV mandate, https://www.greencarreports.com/news/1133027_minnesota-adopts-california-ev-mandate-makes-it-tougher-for-plug-in-compliance-cars (accessed: February 16, 2022); State of
Nevada Climate Initiative, Adopt Low-and Zero-Emissions Passenger
Vehicle Standards, https://climateaction.nv.gov/policies/lev-zev
(accessed: February 16, 2022); Green Car Reports, Virginia becomes
15th Clean Cars State, https://www.greencarcongress.com/2021/03/20210330-virginia.html (accessed: February 16, 2022).
\104\ Section 177 of the Clean Air Act allows other states to
adopt California's new motor vehicle emission standards, if
specified criteria are met.
\105\ At the time of writing, Delaware and Pennsylvania are the
two states that have adopted the LEV standards, but not the ZEV
portion.
---------------------------------------------------------------------------
In their comments on the NPRM, Rivian stated that our ZEV program
modeling should include Minnesota, Virginia, and Nevada as ZEV states,
as those states have recently adopted the
[[Page 25763]]
regulation.\106\ We have not included those states as part of the ZEV
program in the modeling, but have ascertained that reflecting these
three states' adoption of ZEV mandates would have only negligibly
impacted the agency's national-scale modeling. Furthermore, the ZEV
standards for these states go into effect only beginning in MY 2025,
which created an inconsistency with our current modeling approach.
---------------------------------------------------------------------------
\106\ Rivian, Docket ID No. NHTSA-2021-0053-1562, at p. 2.
---------------------------------------------------------------------------
To account for the ZEV program, and particularly as other states
have recently adopted California's ZEV standards, DOT includes the main
provisions of the ZEV program in the CAFE Model's analysis of
compliance pathways. As explained below, incorporating the ZEV program
into the model includes converting vehicles that have been identified
as potential ZEV candidates into battery-electric vehicles (BEVs) at
the first redesign opportunity, so that a manufacturer's fleet meets
calculated ZEV credit requirements. Since ZEV program compliance
pathways happen independently from the adoption of fuel saving
technology in response to increasing CAFE standards, the ZEV program is
considered in the baseline of the analysis, and in all other regulatory
alternatives.
Through its ZEV program, California requires that all manufacturers
that sell cars within the state meet ZEV credit standards. The current
credit requirements are calculated based on manufacturers' California
sales volumes. Manufacturers primarily earn ZEV credits through the
production of BEVs, fuel cell vehicles (FCVs), and transitional zero-
emissions vehicles (TZEVs), which are vehicles with partial
electrification, namely plug-in hybrids (PHEVs). Total credits are
calculated by multiplying the credit value each ZEV receives by the
vehicle's volume.
The ZEV and PHEV/TZEV credit value per vehicle is calculated based
on the vehicle's range; ZEVs may earn up to four credits each and PHEVs
with a US06 all-electric range capability of 10 mi or higher receive an
additional 0.2 credits on top of the credits received based on all-
electric range.\107\ The maximum PHEV credit amount available per
vehicle is 1.10.\108\ Note however that CARB only allows intermediate-
volume manufacturers to meet their ZEV credit requirements through PHEV
production.\109\
---------------------------------------------------------------------------
\107\ US06 is one of the drive cycles used to test fuel economy
and all-electric range, specifically for the simulation of
aggressive driving. See https://www.epa.gov/vehicle-and-fuel-emissions-testing/dynamometer-drive-schedules for more information,
as well as Section III.C.4 and Section III.D.3.d). (Accessed: March
6, 2022)
\108\ 13 California Code of Regulations (CCR) 1962.2(c)(3).
\109\ 13 CCR 1962.2(c)(3).
---------------------------------------------------------------------------
DOT's method for simulating the ZEV program involves several steps;
first, DOT calculates an approximate ZEV credit target for each
manufacturer based on the manufacturer's national sales volumes, share
of sales in Section 177 states, and the CARB credit requirements. Next,
DOT identifies a general pathway to compliance that involves accounting
for manufacturers' potential use of ZEV overcompliance credits or other
credit mechanisms, and the likelihood that manufacturers would choose
to comply with the requirements with BEVs rather than PHEVs or other
types of compliant vehicles, in addition to other factors. For this
analysis, as discussed further below, DOT consulted with CARB to
determine reasonable assumptions for this compliance pathway. Finally,
DOT identifies vehicles in the MY 2020 analysis fleet that
manufacturers could reasonably adapt to comply with the ZEV standards
at the first opportunity for vehicle redesign, based on publicly
announced product plans and other information. Each of these steps is
discussed in turn, below, and a more detailed description of DOT's
simulation of the ZEV program is included in TSD Chapter 2.3.
The CAFE Model is designed to present outcomes at a national scale,
so the ZEV analysis considers the Section 177 states as a group as
opposed to estimating each state's ZEV credit requirements
individually. To capture the appropriate volumes subject to the ZEV
requirement, DOT calculates each manufacturer's total market share in
Section 177 states. DOT also calculates the overall market share of
ZEVs in Section 177 states, in order to estimate as closely as
possible, the number of predicted ZEVs we expect all manufacturers to
sell in those states. These shares are then used to scale down
national-level information in the CAFE Model to ensure that we
represent only Section 177 states in the final calculation of ZEV
credits that we project each manufacturer to earn in future years.
DOT uses MY 2019 National Vehicle Population Profile (NVPP) from
IHS Markit--Polk to calculate these percentages.\110\ These data
include vehicle characteristics such as powertrain, fuel type,
manufacturer, nameplate, and trim level, as well as the state in which
each vehicle is sold, which allows staff to identify the different
types of ZEVs manufacturers sell in the Section 177 state group.
---------------------------------------------------------------------------
\110\ National Vehicle Population Profile (NVPP) 2020, IHS
Markit--Polk. At the time of the analysis, MY 2019 data from the
NVPP contained the most current estimate of market shares by
manufacturer, and best represented the registered vehicle population
on January 1, 2020.
---------------------------------------------------------------------------
We calculate sales volumes for the ZEV credit requirement based on
each manufacturer's future assumed market share in Section 177 states.
DOT decided to carry each manufacturer's ZEV market shares forward to
future years, after examination of past market share data from MY 2016,
from the 2017 version of the NVPP.\111\ Comparison of these data to the
2020 version showed that manufacturers' market shares remain fairly
constant in terms of geographic distribution. Therefore, we determined
that it was reasonable to carry forward the recently calculated market
shares to future years.
---------------------------------------------------------------------------
\111\ National Vehicle Population Profile (NVPP) 2017, IHS
Markit--Polk.
---------------------------------------------------------------------------
We calculate total credits required for ZEV compliance by
multiplying the percentages from CARB's ZEV requirement schedule by the
Section 177 state volumes. CARB's credit percentage requirement
schedule for the years covered in this analysis begins at 9.5 percent
in 2020 and ramps up in increments to 22 percent by 2025.\112\ Note
that the requirements do not currently change after 2025.\113\
---------------------------------------------------------------------------
\112\ See 13 CCR 1962.2(b). The percentage credit requirements
are as follows: 9.5 percent in 2020, 12 percent in 2021, 14.5
percent in 2022, 17 percent in 2023, 19.5 percent in 2024, and 22
percent in 2025 and onward.
\113\ 13 CCR 1962.2(b).
---------------------------------------------------------------------------
We generate national sales volume predictions for future years
using the Compliance Report, a CAFE Model output file that includes
simulated sales by manufacturer, fleet, and model year. We use a
Compliance Report that corresponds to the baseline scenario of 1.5
percent per year increases in standards for both passenger car and
light truck fleets. The resulting national sales volume predictions by
manufacturer are then multiplied by each manufacturer's total market
share in the Section 177 states to capture the appropriate volumes in
the ZEV credits calculation. Required credits by manufacturer, per
year, are determined by multiplying the Section 177 state volumes by
CARB's ZEV credit percentage requirement. These required credits are
subsequently added to the CAFE Model inputs as targets for manufacturer
compliance with ZEV standards in the CAFE baseline.
The estimated ZEV credit requirements serve as a target for
simulating ZEV compliance in the baseline. To achieve this, DOT
determines a modeling philosophy for ZEV pathways, reviews various
sources
[[Page 25764]]
for information regarding upcoming ZEV programs, and inserts those
programs into the analysis fleet inputs. As manufacturers can meet ZEV
standards in a variety of different ways, using various technology
combinations, the analysis must include certain simplifying assumptions
in choosing ZEV pathways. We made these assumptions in conjunction with
guidance from CARB staff. The following sections discuss the approach
used to simulate a pathway to ZEV program compliance in this analysis.
First, DOT targeted 2025 compliance, as opposed to assuming
manufacturers would perfectly comply with their credit requirements in
each year prior to 2025. This simplifying assumption was made upon
review of past history of ZEV credit transfers, existing ZEV credit
banks, and redesign schedules. DOT focused on integrating ZEV
technology throughout that timeline with the target of meeting 2025
obligations; thus, some manufacturers are estimated to over-comply or
under-comply, depending on their individual situations, in the years
2021-2024.
Second, DOT determined that the most reasonable way to model ZEV
compliance would be to allow under-compliance in certain cases and
assume that some manufacturers would not meet their ZEV obligation on
their own in 2025. Instead, these manufacturers were assumed to prefer
to purchase credits from another manufacturer with a credit surplus.
Reviews of past ZEV credit transfers between manufacturers informed the
decision to make this simplifying assumption.\114\ CARB advised that
for these manufacturers, the CAFE Model should still project that each
manufacturer meet approximately 80 percent of their ZEV requirements
with technology included in their own portfolio. Manufacturers that
were observed to have generated many ZEV credits in the past or had
announced major upcoming BEV initiatives were projected to meet 100
percent of their ZEV requirements on their own, without purchasing ZEV
credits from other manufacturers.\115\
---------------------------------------------------------------------------
\114\ See https://ww2.arb.ca.gov/our-work/programs/advanced-clean-cars-program/zev-program/zero-emission-vehicle-credit-balances
for past credit balances and transfer information. (Accessed:
February 16, 2022)
\115\ The following manufacturers were assumed to meet 100-
percent ZEV compliance: Ford, General Motors, Hyundai, Kia, Jaguar
Land Rover, and Volkswagen Automotive. Tesla was also assumed to
meet 100 percent of its required standards, but the analyst team did
not need to add additional ZEV substitutes to the baseline for this
manufacturer.
---------------------------------------------------------------------------
Third, DOT agreed that manufacturers would meet their ZEV credit
requirements in 2025 though the production of BEVs. As discussed above,
manufacturers may choose to build PHEVs or FCVs to earn some portion of
their required ZEV credits. However, DOT projected that manufacturers
would rely on BEVs to meet their credit requirements, based on reviews
of press releases and industry news, as well as discussion with CARB.
Since nearly all manufacturers have announced some plans to produce
BEVs at a scale meaningful to future ZEV requirements, DOT agreed that
this was a reasonable assumption.\116\ Furthermore, as CARB only allows
intermediate-volume manufacturers to meet their ZEV credit requirements
through the production of PHEVs, and the volume status of these few
manufacturers could change over the years, assuming BEV production for
ZEV compliance is the most straightforward path.
---------------------------------------------------------------------------
\116\ See TSD Chapter 2.3 for a list of potential BEV programs
recently announced by manufacturers.
---------------------------------------------------------------------------
Fourth, to account for the new BEV programs announced by some
manufacturers, DOT identified vehicles in the 2020 fleet that closely
matched the upcoming BEVs, by regulatory class, market segment, and
redesign schedule. DOT made an effort to distribute ZEV candidate
vehicles by CAFE regulatory class (light truck, passenger car), by
manufacturer, in a manner consistent with the 2020 manufacturer fleet
mix. Since passenger car and light truck mixes by manufacturer could
change in response to the CAFE policy alternative under consideration,
this effort was deemed necessary in order to avoid redistributing the
fleet mix in an unrealistic manner. However, there were some exceptions
to this assumption, as some manufacturers are already closer to meeting
their ZEV obligation through 2025 with BEVs currently produced, and
some manufacturers underperform their compliance targets more so in one
fleet than another. In these cases, DOT deviated from keeping the LT/PC
mix of BEVs evenly distributed across the manufacturer's
portfolio.\117\
---------------------------------------------------------------------------
\117\ The GM light truck and passenger car distribution is one
such example.
---------------------------------------------------------------------------
DOT then identified future ZEV programs that could plausibly
contribute towards the ZEV requirements for each manufacturer by 2025.
To obtain this information, DOT examined various sources, including
trade press releases, industry announcements, and investor reports. In
many cases, these BEV programs are in addition to programs already in
production.\118\ Some manufacturers have not yet released details of
future electric vehicle programs at the time of writing, but have
indicated goals of reaching certain percentages of electric vehicles in
their portfolios by a specified year. In these cases, DOT reviewed the
manufacturer's current fleet characteristics as well as the
aspirational information in press releases and other news in order to
make reasonable assumptions about the vehicle segment and range of
those future BEVs. No changes in BEV program assumptions were made
between the NPRM and this document.
---------------------------------------------------------------------------
\118\ Examples of BEV programs already in production include the
Nissan Leaf and the Chevrolet Bolt.
---------------------------------------------------------------------------
Overall, analysts assumed that manufacturers would lean towards
producing BEV300s rather than BEV200s, based on the information
reviewed and an initial conversation with CARB.\119\ Phase-in caps were
also considered, especially for BEV200, with the understanding that the
CAFE Model will always pick BEV200 before BEV300 or BEV400, until the
quantity of BEV200s is exhausted. See Section III.D.3.c) for details
regarding BEV phase-in caps.
---------------------------------------------------------------------------
\119\ BEV300s are 300-mile range battery-electric vehicles. See
Section III.D.3.b) for further information regarding electrification
fleet assignments.
---------------------------------------------------------------------------
BEVs with smaller battery packs and less range are less likely to
meet all the performance needs of traditional pickup truck owners
today, such as long-range towing. However, longer-range BEV pickups are
being introduced, and may be joined by new markets in the form of
electric delivery trucks and some light-duty electric truck
applications in state and local government. The extent to which BEVs
will be used in these and other new markets is difficult to project.
DOT did identify certain trucks as upcoming BEVs for ZEV compliance,
and these BEVs were expected to have higher ranges, due to the specific
performance needs associated with these vehicles. Outside of the ZEV
inputs described here, the CAFE Model does not handle the application
of BEV technology with any special considerations as to whether the
vehicle is a pickup truck or not.
Finally, in order to simulate manufacturers' compliance with their
particular ZEV credits target, 142 rows in the analysis fleet were
identified as substitutes for future ZEV programs. As discussed above,
the analysis fleet summarizes the roughly 13.6 million light-duty
vehicles produced and sold in the United States in MY 2020 with more
than 3,500 rows, each reflecting
[[Page 25765]]
information for one vehicle type observed. Each row includes the
vehicle's nameplate and trim level, the sales volume, engine,
transmission, drive configuration, regulatory class, projected redesign
schedule, and fuel saving technologies, among other attributes.
As the goal of the ZEV analysis is to simulate compliance with the
ZEV program in the baseline, and the analysis fleet only contains
vehicles produced during MY 2020, DOT identified existing models in the
analysis fleet that shared certain characteristics with upcoming BEVs.
DOT also focused on identifying substitute vehicles with redesign years
similar to the future BEV's introduction year. The sales volumes of
those existing models, as predicted for 2025, were then used to
simulate production of the upcoming BEVs. DOT identified a combination
of rows that would meet the ZEV target, could contribute productively
towards CAFE program obligations (by manufacturer and by fleet), and
would introduce BEVs in each manufacturer's portfolio in a way that
reasonably aligned with projections and announcements. DOT tagged each
of these rows with information in the Market Data file, instructing the
CAFE Model to apply the specified BEV technology to the row at the
first redesign year, regardless of the scenario or type of CAFE or GHG
simulation.
The CAFE Model does not optimize compliance with the ZEV mandate;
it relies upon the inputs described in this section in order to
estimate each manufacturer's resulting ZEV credits. The resulting
amount of ZEV credits earned by manufacturer for each model year can be
found in the CAFE Model's Compliance file.
Not all ZEV-qualifying vehicles in the U.S. earn ZEV credits, as
they are not all sold in states that have adopted ZEV regulations. In
order to reflect this in the CAFE Model, which only estimates sales
volumes at the national level, the percentages calculated for each
manufacturer are used to scale down the national-level volumes.
Multiplying national-level ZEV sales volumes by these percentages
ensures that only the ZEVs sold in Section 177 states count towards the
ZEV credit targets of each manufacturer.\120\ See Section 5.8 of the
CAFE Model Documentation for a detailed description of how the model
applied these ZEV technologies and any changes made to the model's
programming for the incorporation of the ZEV program into the baseline.
---------------------------------------------------------------------------
\120\ The single exception to this assumption is Mazda, as Mazda
has not yet produced any ZEV-qualifying vehicles at the time of
writing. Thus, the percentage of ZEVs sold in Section 177 states
cannot be calculated from existing data. However, Mazda has
indicated its intention to produce ZEV-qualifying vehicles in the
future, so DOT assumed that 100 percent of future ZEVs would be sold
in Section 177 states for the purposes of estimating ZEV credits in
the CAFE Model.
---------------------------------------------------------------------------
As discussed above, DOT made an effort to distribute the newly
identified ZEV candidates between CAFE regulatory classes (light truck
and passenger car) in a manner consistent with the proportions seen in
the 2020 analysis fleet, by manufacturer. As mentioned previously,
there were a few exceptions to this assumption in cases where
manufacturers' regulatory class distribution of current or planned ZEV
programs clearly differed from their regulatory class distribution as a
whole.
In some instances, the regulatory distribution of flagged ZEV
candidates leaned towards a higher portion of PCs. The reasoning behind
this differs in each case, but there is an observed pattern in the 2020
analysis fleet of fewer BEVs being light trucks, especially pickups.
The 2020 analysis fleet contains no BEV pickups in the light truck
segment. The slow emergence of electric pickups could be linked to the
specific performance needs associated with pickup trucks. However, the
market for BEVs may emerge in unexpected ways that are difficult to
project. Examples of this include anticipated electric delivery trucks
and light-duty electric trucks used by state and local governments. Due
to these considerations, DOT tagged some trucks as BEVs for ZEV, and
expected that these would generally be of higher ranges.
TSD Chapter 2.3 includes more information about the process we use
to simulate ZEV program compliance in this analysis.
4. Technology Effectiveness Values
The next input we use to simulate manufacturers' decision-making
processes for the year-by-year application of technologies to specific
vehicles are estimates of how effective each technology would be at
reducing fuel consumption. For this analysis, we use full-vehicle
modeling and simulation to estimate the fuel economy improvements
manufacturers could make to a fleet of vehicles, considering the
vehicles' technical specifications and how combinations of technologies
interact. Full-vehicle modeling and simulation uses physics-based
models to predict how combinations of technologies perform as a full
system under defined conditions. We use full vehicle simulations
performed in Autonomie, a physics-based full-vehicle modeling and
simulation software developed and maintained by the U.S. Department of
Energy's Argonne National Laboratory.\121\
---------------------------------------------------------------------------
\121\ Islam, E.S., A. Moawad, N. Kim, R. Vijayagopal, and A.
Rousseau. A Detailed Vehicle Simulation Process to Support CAFE
Standards for the MY 2024-2026 Analysis. ANL/ESD-21/9 (hereinafter,
Autonomie model documentation).
---------------------------------------------------------------------------
A model is a mathematical representation of a system, and
simulation is the behavior of that mathematical representation over
time. In this analysis, the model is a mathematical representation of
an entire vehicle,\122\ including its individual components such as the
engine and transmission, overall vehicle characteristics such as mass
and aerodynamic drag, and the environmental conditions, such as ambient
temperature and barometric pressure. We simulate the model's behavior
over test cycles, including the 2-cycle laboratory compliance tests (or
2-cycle tests),\123\ to determine how the individual components
interact.
---------------------------------------------------------------------------
\122\ Each full vehicle model in this analysis is composed of
sub-models, which is why the full vehicle model could also be
referred to as a full system model, composed of sub-system models.
\123\ EPA's compliance test cycles are used to measure the fuel
economy of a vehicle. For readers unfamiliar with this process, it
is like running a car on a treadmill following a program--or more
specifically, two programs. The ``programs'' are the ``urban
cycle,'' or Federal Test Procedure (abbreviated as ``FTP''), and the
``highway cycle,'' or Highway Fuel Economy Test (abbreviated as
``HFET''), and they have not changed substantively since 1975. Each
cycle is a designated speed trace (of vehicle speed versus time)
that all certified vehicles must follow during testing. The FTP is
meant roughly to simulate stop and go city driving, and the HFET is
meant roughly to simulate steady flowing highway driving at about 50
mph.
---------------------------------------------------------------------------
Using full-vehicle modeling and simulation to estimate technology
efficiency improvements has two primary advantages over using single or
limited point estimates. An analysis using single or limited point
estimates may assume that, for example, one fuel economy-improving
technology with an effectiveness value of 5 percent by itself and
another technology with an effectiveness value of 10 percent by itself,
when applied together achieve an additive improvement of 15 percent.
Single point estimates generally do not provide accurate effectiveness
values because they do not capture complex relationships among
technologies. Technology effectiveness often differs significantly
depending on the vehicle type (e.g., sedan versus pickup truck) and the
way in which the technology interacts with other technologies on the
vehicle, as different technologies may provide different incremental
levels of fuel economy improvement if implemented alone or in
combination with other technologies. Any
[[Page 25766]]
oversimplification of these complex interactions leads to less accurate
and often overestimated effectiveness estimates.
In addition, because manufacturers often implement several fuel-
saving technologies simultaneously when redesigning a vehicle, it is
difficult to isolate the effect of individual technologies using
laboratory measurement of production vehicles alone. Modeling and
simulation offer the opportunity to isolate the effects of individual
technologies by using a single or small number of baseline vehicle
configurations and incrementally adding technologies to those baseline
configurations. This provides a consistent reference point for the
incremental effectiveness estimates for each technology and for
combinations of technologies for each vehicle type. Vehicle modeling
also reduces the potential for overcounting or undercounting technology
effectiveness.
An important feature of this analysis is that the incremental
effectiveness of each technology and combinations of technologies
should be accurate and relative to a consistent baseline vehicle. For
this analysis, the baseline absolute fuel economy value for each
vehicle in the analysis fleet is based on CAFE compliance data for each
make and model.\124\ The absolute fuel economy values of the full
vehicle simulations are used only to determine incremental
effectiveness and are never used directly to assign an absolute fuel
economy value to any vehicle model or configuration. For subsequent
technology changes, we apply the incremental effectiveness values of
one or more technologies to the baseline fuel economy value to
determine the absolute fuel economy achieved for applying the
technology change.
---------------------------------------------------------------------------
\124\ See Section III.C.2 for further discussion of CAFE
compliance data in the Market Data file.
---------------------------------------------------------------------------
As an example, if a Ford F-150 2-wheel drive crew cab and short bed
in the analysis fleet has a fuel economy value of 30 mpg for CAFE
compliance, 30 mpg will be considered the reference absolute fuel
economy value. A similar full vehicle model node in the Autonomie
simulation may begin with an average fuel economy value of 32 mpg, and
with incremental addition of a specific technology X its fuel economy
improves to 35 mpg, a 9.3 percent improvement. In this example, the
incremental fuel economy improvement (9.3 percent) from technology X
would be applied to the F-150's 30 mpg absolute value.
We determine the incremental effectiveness of technologies as
applied to the thousands of unique vehicle and technology combinations
in the analysis fleet. Although, as mentioned above, full-vehicle
modeling and simulation reduces the work and time required to assess
the impact of moving a vehicle from one technology state to another, it
would be impractical--if not impossible--to build a unique vehicle
model for every individual vehicle in the analysis fleet. Therefore, as
discussed in the following sections, the Autonomie analysis relies on
ten vehicle technology class models that are representative of large
portions of the analysis fleet vehicles. The vehicle technology classes
ensure that key vehicle characteristics are reasonably represented in
the full vehicle models.
We sought comment on the full vehicle modeling and simulation
assumptions used for this analysis and received some comments specific
to individual technologies, which are discussed further in the
individual technology subsections in final rule Section III.D. However,
we did not receive any comments on our use of Autonomie itself. The
next sections discuss the details of the technology effectiveness
analysis input specifications and assumptions that we continued to use
for this final rule analysis.
(a) Full Vehicle Modeling and Simulation
As discussed above, for this analysis we use Argonne's full vehicle
modeling tool, Autonomie, to build vehicle models with different
technology combinations and simulate the performance of those models
over regulatory test cycles. The difference in the simulated
performance between full vehicle models, with differing technology
combination, is used to determine effectiveness values. We consider
over 50 individual technologies as inputs to the Autonomie
modeling.\125\ These inputs consist of engine technologies,
transmission technologies, powertrain electrification, light-weighting,
aerodynamic improvements, and tire rolling resistance improvements.
Section III.D broadly discusses each of the technology groupings
definitions, inputs, and assumptions. A deeper discussion of the
Autonomie modeled subsystems, and how inputs feed the sub models
resulting in outputs, is contained in the Autonomie model documentation
that accompanies this analysis. The 50 individual technologies, when
considered with the ten vehicle technology classes, result in over 1
million individual vehicle technology combination models. For
additional discussion on the full vehicle modeling used in this
analysis see TSD Chapter 2.
---------------------------------------------------------------------------
\125\ See Autonomie model documentation; ANL--All
Assumptions_Summary_NPRM_022021.xlsx; ANL--Data Dictionary January
2021.xlsx.
---------------------------------------------------------------------------
While Argonne built full-vehicle models and ran simulations for
many combinations of technologies, it did not simulate literally every
single vehicle model/configuration in the analysis fleet. Not only
would it be impractical to assemble the requisite detailed information
specific to each vehicle/model configuration, much of which would
likely only be provided on a confidential basis, doing so would
increase the scale of the simulation effort by orders of magnitude.
Instead, Argonne simulated ten different vehicle types, corresponding
to the five ``technology classes'' generally used in CAFE analysis over
the past several rulemakings, each with two performance levels and
corresponding vehicle technical specifications (e.g., small car, small
performance car, pickup truck, performance pickup truck, etc.).
Technology classes are a means of specifying common technology
input assumptions for vehicles that share similar characteristics.
Because each vehicle technology class has unique characteristics, the
effectiveness of technologies and combinations of technologies is
different for each technology class. Conducting Autonomie simulations
uniquely for each technology class provides a specific set of
simulations and effectiveness data for each technology class. In this
analysis the technology classes are compact cars, midsize cars, small
SUVs, large SUVs, and pickup trucks. In addition, for each vehicle
class there are two levels of performance attributes (for a total of 10
technology classes). The high performance and low performance vehicles
classifications allow for better diversity in estimating technology
effectiveness across the fleet.
For additional discussion on the development of the vehicle
technology classes used in this analysis and the attributes used to
characterize each vehicle technology class, see TSD Chapter 2.4 and the
Autonomie model documentation.
Before any simulation is initiated in Autonomie, Argonne must
``build'' a vehicle by assigning reference technologies and initial
attributes to the components of the vehicle model representing each
technology class. The reference technologies are baseline
[[Page 25767]]
technologies that represent the first step on each technology pathway
used in the analysis. For example, a compact car is built by assigning
it a baseline engine (DOHC, VVT, PFI), a baseline transmission (AT5), a
baseline level of aerodynamic improvement (AERO0), a baseline level of
rolling resistance improvement (ROLL0), a baseline level of mass
reduction technology (MR0), and corresponding attributes from the
Argonne vehicle assumptions database like individual component weights.
A baseline vehicle will have a unique starting point for the simulation
and a unique set of assigned inputs and attributes, based on its
technology class. Argonne collected over a hundred baseline vehicle
attributes to build the baseline vehicle for each technology class. In
addition, to account for the weight of different engine sizes, like 4-
cylinder versus 8-cylinder or turbocharged versus naturally aspirated
engines, Argonne developed a relationship curve between peak power and
engine weight based on the A2Mac1 benchmarking data. Argonne uses the
developed relationship to estimate mass for all engines. For additional
discussion on the development and optimization of the baseline vehicle
models and the baseline attributes used in this analysis see TSD
Chapter 2.4 and the Autonomie model documentation.
The next step in the process is to run a powertrain sizing
algorithm that ensures the built vehicle meets or exceeds defined
performance metrics, including low-speed acceleration (time required to
accelerate from 0-60 mph), high-speed passing acceleration (time
required to accelerate from 50-80 mph), gradeability (the ability of
the vehicle to maintain constant 65 miles per hour speed on a six
percent upgrade), and towing capacity. Together, these performance
criteria are widely used by the automotive industry as metrics to
quantify vehicle performance attributes that consumers observe and that
are important for vehicle utility and customer satisfaction.
As with conventional vehicle models, electrified vehicle models
were also built from the ground up. For MY 2020, the U.S. market has an
expanded number of available hybrid and electric vehicle models. To
capture improvements for electrified vehicles for this analysis, DOT
applied a mass regression analysis process that considers electric
motor weight versus electric motor power (similar to the regression
analysis for internal combustion engine weights) for vehicle models
that have adopted electric motors. Benchmarking data for hybrid and
electric vehicles from the A2Mac1 database were analyzed to develop a
regression curve of electric motor peak power versus electric motor
weight.\126\
---------------------------------------------------------------------------
\126\ See Autonomie model documentation, Chapter 5.2.10,
Electric Machines System Weight.
---------------------------------------------------------------------------
We maintain performance neutrality in the full vehicle simulations
by resizing engines, electric machines, and hybrid electric vehicle
battery packs at specific incremental technology steps. To address
product complexity and economies of scale, engine resizing is limited
to specific incremental technology changes that would typically be
associated with a major vehicle or engine redesign. This is intended to
reflect manufacturers' comments to DOT on how they consider engine
resizing and product complexity, and DOT's observations on industry
product complexity. A detailed discussion on powertrain sizing can be
found in TSD Chapter 2.4 and in the Autonomie model documentation.
After all vehicle class and technology combination models have been
built, Autonomie simulates the vehicles' performance on test cycles to
calculate the effectiveness improvement of adding fuel-economy-
improving technologies to the vehicle. Simulating vehicles' performance
using tests and procedures specified by Federal law and regulations
minimizes the potential variation in determining technology
effectiveness.
For vehicles with conventional powertrains and micro hybrids,
Autonomie simulates the vehicles per EPA 2-cycle test procedures and
guidelines.\127\ For mild and full hybrid electric vehicles and FCVs,
Autonomie simulates the vehicles using the same EPA 2-cycle test
procedure and guidelines, and the drive cycles are repeated until the
initial and final state of charge are within a SAE J1711 tolerance. For
PHEVs, Autonomie simulates vehicles per similar procedures and
guidelines as prescribed in SAE J1711.\128\ For BEVs Autonomie
simulates vehicles per similar procedures and guidelines as prescribed
in SAE J1634.\129\
---------------------------------------------------------------------------
\127\ 40 CFR part 600.
\128\ PHEV testing is broken into several phases based on SAE
J1711: Charge-sustaining on the city cycle and HWFET cycle, and
charge-depleting on the city and HWFET cycles.
\129\ SAE J1634. ``Battery Electric Vehicle Energy Consumption
and Range Test Procedure.'' July 12, 2017.
---------------------------------------------------------------------------
We received comments from The International Council on Clean
Transportation (ICCT) regarding the application of the engine sizing
algorithm, and when it is applied in relation to vehicle road load
improvement technologies. ICCT stated that, ``[d]ue to the large
uncertainties in when and how to downsize engines for the variety of
vehicles, the only acceptable solution is to always model the
appropriate amount of engine downsizing to maintain performance.''
\130\
---------------------------------------------------------------------------
\130\ ICCT, Docket No. NHTSA-2021-0053-1581-A1, at p. 5.
---------------------------------------------------------------------------
We disagree with the comment implying that engine resizing is
required for every technology change on a vehicle platform. We believe
that this would artificially inflate effectiveness relative to cost.
Manufacturers have repeatedly and consistently conveyed that the costs
for redesign and the increased manufacturing complexity resulting from
continual resizing engine displacement for small technology changes
preclude them from doing so. NHTSA believes that it would not be
reasonable or cost-effective to expect resizing powertrains for every
unique combination of technologies, and even less reasonable and cost-
effective for every unique combination of technologies across every
vehicle model due to the extreme manufacturing complexity that would be
required to do so.\131\ In addition, a 2011 NAS report stated that
``[f]or small (under 5 percent [of curb weight]) changes in mass,
resizing the engine may not be justified, but as the reduction in mass
increases (greater than 10 percent [of curb weight]), it becomes more
important for certain vehicles to resize the engine and seek secondary
mass reduction opportunities.'' \132\
---------------------------------------------------------------------------
\131\ For more details, see comments and discussion in the 2020
Rulemaking Preamble Section VI.B.3.a)(6) Performance Neutrality.
\132\ National Research Council 2011. Assessment of Fuel Economy
Technologies for Light-Duty Vehicles. Washington, DC: The National
Academies Press. https://doi.org/10.17226/12924 (hereinafter, 2011
NAS Report), at 107.
---------------------------------------------------------------------------
We also believe that ICCT's comment regarding Autonomie's engine
resizing process is further addressed by the Autonomie's powertrain
calibration process. We do agree that the powertrain should be re-
calibrated for every unique technology combination and this calibration
is performed as part of the transmission shift initializer
routine.\133\ Autonomie runs the shift initializer routine for every
unique Autonomie full vehicle model configuration and generates
customized transmission shift maps. The algorithms' optimization is
designed to balance minimization of energy consumption and vehicle
performance.
---------------------------------------------------------------------------
\133\ See FRM ANL Model Documentation at Paragraph 4.4.5.2.
---------------------------------------------------------------------------
(b) Performance Neutrality
The purpose of the CAFE analysis is to examine the impact of
technology
[[Page 25768]]
application that can improve fuel economy. When the fuel economy-
improving technology is applied, frequently the manufacturer must
choose how the technology will affect the vehicle. The advantages of
the new technology can either be completely applied to improving fuel
economy or be used to increase vehicle performance while maintaining
the existing fuel economy, or some mix of the two effects.
Historically, vehicle performance, historically equated with
horsepower, has improved over the years as more technology is applied
to the fleet. The average horsepower is the highest that it has ever
been; all vehicle types have improved horsepower by at least 43 percent
compared to the 1978 model year, and pickup trucks have improved by 49
percent.\134\ Fuel economy has also improved, but the horsepower and
acceleration trends show that not 100 percent of technological
improvements have been applied to fuel savings. While future trends are
uncertain, the past trends suggest that vehicle performance is unlikely
to decrease, as it seems reasonable to assume that customers will, at a
minimum, demand vehicles that offer the same utility as today's fleet.
---------------------------------------------------------------------------
\134\ ``The 2021 EPA Automotive Trends Report, Greenhouse Gas
Emissions, Fuel Economy, and Technology since 1975,'' EPA-420-R-21-
023, November 2021, at pp. 20-7 (hereinafter, 2021 EPA Automotive
Trends Report).
---------------------------------------------------------------------------
For this rulemaking analysis, we analyzed technology pathways
manufacturers could use for compliance that attempt to maintain vehicle
attributes, utility, and performance. Using this approach allows us to
assess the costs and benefits of potential standards under a scenario
where consumers continue to get the similar vehicle attributes and
features, other than changes in fuel economy. The purpose of
constraining vehicle attributes is to simplify the analysis and reduce
variance in other attributes that consumers may value across the
analyzed regulatory alternatives. This allows for a streamlined
accounting of costs and benefits by not requiring the values of other
vehicle attributes.
To confirm minimal differences in performance metrics across
regulatory alternatives, we analyzed the sales-weighted average 0-60
mph acceleration performance of the entire simulated vehicle fleet for
MYs 2020 and 2029. The analysis compared performance under the baseline
standards and Preferred Alternative. For the NPRM, this analysis
identified that the analysis fleet under the No-Action Alternative in
MY 2029 had a 0.77 percent worse 0-60 mph acceleration time than under
the Preferred Alternative; in other words, the alternative with the
higher fuel economy standards also showed greater acceleration and
performance. For the final rule analysis, using the similar approach
yielded a 0.0615 percent better (as compared to the baseline) 0-60 mph
acceleration time, indicating there is minimal difference in
performance between the alternatives. This assessment shows that for
this analysis, the performance difference is minimal across regulatory
alternatives and across the simulated model years, which allows for
fair, direct comparison among the alternatives. Further details about
this assessment can be found in TSD Chapter 2.4.5.
Overall, commenters were supportive of our approach to maintaining
performance neutrality and the metrics we use to accomplish this.
Commenters said we should continue to improve our methodologies for
maintaining performance neutrality.\135\ Auto Innovators stated that
``[t]he [a]gencies have historically sought to maintain the performance
characteristics of vehicles modeled with fuel economy-improving
technologies.'' They added that they ``appreciate that the [a]gencies
continue to consider high- speed acceleration, gradeability, towing,
range, traction, and interior room (including headroom) in the analysis
when sizing powertrains and evaluating pathways for road-load
reductions.'' Finally, they stated that ``[a]ll of these parameters
should be considered separately, not just in combination. (For example,
we do not support an approach where various acceleration times are
added together to create a single `performance' statistic.
Manufacturers must provide all types of performance, not just one or
two to the detriment of others.).''
---------------------------------------------------------------------------
\135\ RV Industry Association, Docket No. NHTSA-2021-0053-0053,
at 4; Auto Innovators, Docket No. NHTSA-2021-0053-1492, at p. 62.
---------------------------------------------------------------------------
The RV Industry Association commented that the agency should
include towing capacity considerations for large SUVs because of the
public's reliance on large SUVs for RV towing.\136\ Currently, our
analysis assumes that SUVs are primarily used for carrying passengers
and cargo and towing is not their primary function, in contrast to how
full-size pickups are characterized in the analysis. Other aspects of
the analysis capture potential performance limitations for SUVs such as
limiting the adoption of technologies that could be considered less
practical for SUVs. For example, for some larger SUVs with higher power
density requirements, we limit HCR engine technologies and power-split
strong hybrid powertrains. For more details on these limitations, see
Section III.D.1.c) of this preamble for each technology pathway.
---------------------------------------------------------------------------
\136\ RV Industry Association, at p. 4.
---------------------------------------------------------------------------
For this final rule analysis, we continued to use the same
methodology for modeling full vehicles and maintaining performance
neutrality. As such, the estimated compliance costs reflect the
assumption that manufacturers will resize powertrains or make other
adjustments to maintain performance while increasing fuel economy. We
will continue to monitor performance neutrality metrics and their
incorporation as part of future analyses.
(c) Implementation in the CAFE Model
The CAFE Model uses two elements of information from the large
amount of data generated by the Autonomie simulation runs: Battery
costs, and fuel consumption on the city and highway cycles. We combine
the fuel economy information from the two cycles to produce a composite
fuel economy for each vehicle, and for each fuel used in dual fuel
vehicles. The fuel economy information for each simulation run is
converted into a single value for use in the CAFE Model.
In addition to the technologies in the Autonomie simulation, the
CAFE Model also incorporated a handful of technologies not explicitly
simulated in Autonomie. These technologies' performance either could
not be captured on the 2-cycle test, or there were no robust data
usable as an input for full-vehicle modeling and simulation. The
specific technologies are discussed in the individual technology
sections below and in TSD Chapter 3. To calculate fuel economy
improvements attributable to these additional technologies, estimates
of fuel consumption improvement factors were developed and scale
multiplicatively when applied together. See TSD Chapter 3 for a
complete discussion on how these factors were developed. The Autonomie-
simulated results and additional technologies are combined, forming a
single dataset used by the CAFE Model.
Each line in the CAFE Model dataset represents a unique combination
of technologies. We organize the records using a unique technology
state vector, or technology key (tech key), that describes the
technology content associated with each unique record. The modeled 2-
cycle fuel economy (miles per gallon) of each combination is converted
into fuel consumption (gallons per mile) and then normalized relative
to a baseline tech key. The improvement factors used by the model
[[Page 25769]]
are a given combination's fuel consumption improvement relative to the
baseline tech key in its technology class.
The tech key format was developed by recognizing that most of the
technology pathways are unrelated and are only logically linked to
designate the direction in which technologies are allowed to progress.
As a result, it is possible to condense the paths into groups based on
the specific technology. These groups are used to define the technology
vector, or tech key. The following technology groups defined the tech
key: Engine cam configuration (CONFIG), VVT engine technology (VVT),
VVL engine technology (VVL), SGDI engine technology (SGDI), DEAC engine
technology (DEAC), non-basic engine technologies (ADVENG), transmission
technologies (TRANS), electrification and hybridization (ELEC), low
rolling resistance tires (ROLL), aerodynamic improvements (AERO), mass
reduction levels (MR), EFR engine technology (EFR), electric accessory
improvement technologies (ELECACC), LDB technology (LDB), and SAX
technology (SAX). This summarizes to a tech key with the following
fields: CONFIG; VVT; VVL; SGDI; DEAC; ADVENG; TRANS; ELEC; ROLL; AERO;
MR; EFR; ELECACC; LDB; SAX. It should be noted that some of the fields
may be blank for some tech key combinations. These fields will be left
visible for the examples below, but blank fields may be omitted from
tech keys shown elsewhere in the documentation.
As an example, a technology state vector describing a vehicle with
a SOHC engine, variable valve timing (only), a 6-speed automatic
transmission, a belt-integrated starter generator, rolling resistance
(level 1), aerodynamic improvements (level 2), mass reduction (level
1), electric power steering, and low drag brakes, would be specified as
``SOHC; VVT; ; ; ; ; AT6; BISG; ROLL10; AERO20; MR1; ; EPS; LDB ; .''
\137\
---------------------------------------------------------------------------
\137\ In the example tech key, the series of semicolons between
VVT and AT6 correspond to the engine technologies which are not
included as part of the combination, while the gap between MR1 and
EPS corresponds to EFR and the omitted technology after LDB is SAX.
The extra semicolons for omitted technologies are preserved in this
example for clarity and emphasis and will not be included in future
examples.
---------------------------------------------------------------------------
Once a vehicle is assigned (or mapped) to an appropriate tech key,
adding a new technology to the vehicle simply represents progress from
a previous tech key to a new tech key. The previous tech key refers to
the technologies that are currently in use on a vehicle. The new tech
key is determined, in the simulation, by adding a new technology to the
combination represented by the previous state vector while
simultaneously removing any technologies that are superseded by the
newly added one.
For example, start with a vehicle with the tech key: SOHC; VVT;
AT6; BISG; ROLL10; AERO20; MR1; EPS; LDB. Assume the simulation is
evaluating PHEV20 as a candidate technology for application on this
vehicle. The new tech key for this vehicle is computed by removing
SOHC, VVT, AT6, and BISG technologies from the previous state
vector,\138\ and adding PHEV20, resulting a tech key that looks like
this: PHEV20; ROLL10; AERO20; MR1; EPS; LDB.
---------------------------------------------------------------------------
\138\ For more discussion of how the CAFE Model handles
technology supersession, see S4.5 of the CAFE Model Documentation.
---------------------------------------------------------------------------
From here, the simulation obtains a fuel economy improvement factor
for the new combination of technologies and applies that factor to the
fuel economy of a vehicle in the analysis fleet. The resulting
improvement is applied to the original compliance fuel economy value
for a discrete vehicle in the analysis fleet.
5. Defining Technology Adoption in the Rulemaking Timeframe
As discussed in Section III.C.2, starting with a fixed analysis
fleet (for this analysis, the MY 2020 fleet indicated in manufacturers'
early CAFE compliance data), the CAFE Model estimates ways each
manufacturer could potentially apply specific fuel-saving technologies
to specific vehicle model/configurations in response to, among other
things (such as fuel prices), CAFE standards, CO2 standards,
commitments some manufacturers have made to CARB's ``Framework
Agreements,'' and ZEV mandates imposed by California and several other
states. The CAFE Model follows a year-by-year approach to simulating
manufacturers' potential decisions to apply technology, accounting for
multiyear planning within the context of estimated schedules for future
vehicle redesigns and refreshes during which significant technology
changes may most practicably be implemented.
The modeled technology adoption for each manufacturer under each
regulatory alternative depends on this representation of multiyear
planning, and on a range of other factors represented by other model
characteristics and inputs, such as the logical progression of
technologies defined by the model's technology pathways; the
technologies already present in the analysis fleet; inputs directing
the model to ``skip'' specific technologies for specific vehicle model/
configurations in the analysis fleet (e.g., because secondary axle
disconnect cannot be applied to 2-wheel-drive vehicles, and because
manufacturers already heavily invested in engine turbocharging and
downsizing are unlikely to abandon this approach in favor of using high
compression ratios); inputs defining the sharing of engines,
transmissions, and vehicle platforms in the analysis fleet; the model's
logical approach to preserving this sharing; inputs defining each
regulatory alternative's specific requirements; inputs defining
expected future fuel prices, annual mileage accumulation, and valuation
of avoided fuel consumption; inputs defining the estimated efficacy and
future cost (accounting for projected future ``learning'' effects) of
included technologies; inputs controlling the maximum pace the
simulation is to ``phase in'' each technology; and inputs further
defining the availability of each technology to specific technology
classes.
Two of these inputs--the ``phase-in cap'' and the ``phase-in start
year''--apply to the manufacturer's entire estimated production and,
for each technology, define a share of production in each model year
that, once exceeded, will stop the model from further applying that
technology to that manufacturer's fleet in that model year. The
influence of these inputs varies with regulatory stringency and other
model inputs. For example, setting the inputs to allow immediate 100
percent penetration of a technology will not guarantee any application
of the technology if stringency increases are low and the technology is
not at all cost effective. Also, even if these are set to allow only
very slow adoption of a technology, other model aspects and inputs may
nevertheless force more rapid application than these inputs, alone,
would suggest (e.g., because an engine technology propagates quickly
due to sharing across multiple vehicles, or because BEV application
must increase quickly in response to ZEV requirements). For this
analysis, nearly all of these inputs are set at levels that do not
limit the simulation at all.
As discussed below, for the most advanced engines (advanced
cylinder deactivation, variable compression ratio, variable
turbocharger geometry, and turbocharging with cylinder deactivation),
we have specified phase-in caps and phase-in start years that limit the
pace at which the analysis shows the technology being adopted in
[[Page 25770]]
the rulemaking timeframe. For example, this analysis applies a 34-
percent phase-in cap and MY 2019 phase-in start year for advanced
cylinder deactivation (ADEAC), meaning that in MY 2021 (using a MY 2020
fleet, the analysis begins simulating further technology application in
MY 2021), the model will stop adding ADEAC to a manufacturer's MY 2021
fleet once ADEAC reaches more than 68-percent penetration, because 34%
x (2021-2019) = 34% x 2 = 68%.
We apply phase-in caps and corresponding start years to prevent the
simulation from showing unlikely rates of applying battery-electric
vehicles (BEVs), such as showing that a manufacturer producing very few
BEVs in MY 2020 could plausibly replace every product with a 300- or
400-mile BEV by MY 2025. Also, as discussed in Section III.D.4, we
apply phase-in caps and corresponding start years intended to ensure
that the simulation's plausible application of the highest included
levels of mass reduction (20 and 28.2 percent reductions of vehicle
``glider'' weight) do not, for example, outpace plausible supply of raw
materials and development of entirely new manufacturing facilities.
These model logical structures and inputs act together to produce
estimates of ways each manufacturer could potentially shift to new
fuel-saving technologies over time, reflecting some measure of
protection against rates of change not reflected in, for example,
technology cost inputs. This does not mean that every modeled solution
would necessarily be economically practicable. Using technology
adoption features like phase-in caps and phase-in start years is one
mechanism that can be used so that the analysis better represents the
potential costs and benefits of technology application in the
rulemaking timeframe.
6. Technology Costs
DOT estimates present and future costs for fuel-saving technologies
taking into consideration the type of vehicle, or type of engine if
technology costs vary by application. These cost estimates are based on
three main inputs. First, we estimate direct manufacturing costs
(DMCs), or the component and labor costs of producing and assembling
the physical parts and systems, assuming high volume production. DMCs
generally do not include the indirect costs of tools, capital
equipment, financing costs, engineering, sales, administrative support
or return on investment. DOT accounts for these indirect costs via a
scalar markup of direct manufacturing costs (the retail price
equivalent, or RPE). Finally, costs for technologies may change over
time as industry streamlines design and manufacturing processes. To
reflect this, DOT estimates potential cost improvements with learning
effects (LE). The retail cost of equipment in any future year is
estimated to be equal to the product of the DMC, RPE, and LE.
Considering the retail cost of equipment, instead of merely direct
manufacturing costs, is important to account for the real-world price
effects of a technology, as well as market realities.
(a) Direct Manufacturing Costs
Direct manufacturing costs (DMCs) are the component and assembly
costs of the physical parts and systems that make up a complete
vehicle. The analysis uses agency-sponsored tear-down studies of
vehicles and parts to estimate the DMCs of individual technologies, in
addition to independent tear-down studies, other publications, and
confidential business information. In the simplest cases, the agency-
sponsored studies produce results that confirm third-party industry
estimates and align with confidential information provided by
manufacturers and suppliers. In cases with a large difference between
the tear-down study results and credible independent sources, DOT
scrutinized the study assumptions, and sometimes revised or updated the
analysis accordingly.
Due to the variety of technologies and their applications, and the
cost and time required to conduct detailed tear-down analyses, the
agency did not sponsor teardown studies for every technology. In
addition, we consider some fuel-saving technologies that are pre-
production or are sold in very small pilot volumes. For those
technologies, DOT could not conduct a tear-down study to assess costs
because the product is not yet in the marketplace for evaluation. In
these cases, DOT relied upon third-party estimates and confidential
information from suppliers and manufacturers; however, there are some
common pitfalls with relying on confidential business information to
estimate costs. The agency and the source may have had incongruent or
incompatible definitions of ``baseline.'' The source may have provided
DMCs at a date many years in the future, and assumed very high
production volumes, important caveats to consider for agency analysis.
In addition, a source, under no contractual obligation to DOT, may
provide incomplete and/or misleading information. In other cases,
intellectual property considerations and strategic business
partnerships may have contributed to a manufacturer's cost information
and could be difficult to account for in the CAFE Model as not all
manufacturers may have access to proprietary technologies at stated
costs. The agency carefully evaluates new information in light of these
common pitfalls, especially regarding emerging technologies.
While costs for fuel-saving technologies reflect the best estimates
available today, technology cost estimates will likely change in the
future as technologies are deployed and as production is expanded. For
emerging technologies, DOT uses the best information available at the
time of the analysis and will continue to update cost assumptions for
any future analysis. The discussion of each category of technologies in
Section III.D (e.g., engines, transmissions, electrification) and
corresponding TSD Chapter 3 summarizes the specific cost estimates DOT
applied for this analysis.
(b) Indirect Costs (Retail Price Equivalent)
As discussed above, direct costs represent the cost associated with
acquiring raw materials, fabricating parts, and assembling vehicles
with the various technologies manufacturers are expected to use to meet
future CAFE standards. They include materials, labor, and variable
energy costs required to produce and assemble the vehicle. However,
they do not include overhead costs required to develop and produce the
vehicle, costs incurred by manufacturers or dealers to sell vehicles,
or the profit manufacturers and dealers make from their investments.
All of these items contribute to the price consumers ultimately pay for
the vehicle. These components of retail prices are illustrated in Table
III-3 below.
[[Page 25771]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.057
To estimate the impact of higher vehicle prices on consumers, both
direct and indirect costs must be considered. To estimate total
consumer costs, DOT multiplies direct manufacturing costs by an
indirect cost factor to represent the average price for fuel-saving
technologies at retail.
Historically, the method most commonly used to estimate indirect
costs of producing a motor vehicle has been the retail price equivalent
(RPE). The RPE markup factor is based on an examination of historical
financial data contained in 10-K reports filed by manufacturers with
the Securities and Exchange Commission (SEC). It represents the ratio
between the retail price of motor vehicles and the direct costs of all
activities that manufacturers engage in.
Figure III-4 indicates that for more than three decades, the retail
price of motor vehicles has been, on average, roughly 50 percent above
the direct cost expenditures of manufacturers. This ratio has been
remarkably consistent, averaging roughly 1.5 with minor variations from
year to year over this period. At no point has the RPE markup exceeded
1.6 or fallen below 1.4.\139\ During this time frame, the average
annual increase in real direct costs was 2.5 percent, and the average
annual increase in real indirect costs was also 2.5 percent. Figure
III-4 illustrates the historical relationship between retail prices and
direct manufacturing costs.\140\
---------------------------------------------------------------------------
\139\ Based on data from 1972-1997 and 2007. Data were not
available for intervening years, but results for 2007 seem to
indicate no significant change in the historical trend.
\140\ Rogozhin, A., Gallaher, M., & McManus, W., 2009,
Automobile Industry Retail Price Equivalent and Indirect Cost
Multipliers. Report by RTI International to Office of Transportation
Air Quality. U.S. Environmental Protection Agency, RTI Project
Number 0211577.002.004, February, Research Triangle Park, N.C.
Spinney, B.C., Faigin, B., Bowie, N., & S. Kratzke, 1999, Advanced
Air Bag Systems Cost, Weight, and Lead Time analysis Summary Report,
Contract NO. DTNH22-96-0-12003, Task Orders--001, 003, and 005.
Washington, DC, U.S. Department of Transportation.
---------------------------------------------------------------------------
An RPE of 1.5 does not imply that manufacturers automatically mark
up each vehicle by exactly 50 percent. Rather, it means that, over
time, the competitive marketplace has resulted in pricing structures
that average out to this relationship across the entire industry.
Prices for any individual model may be marked up at a higher or lower
rate depending on market demand. The consumer who buys a popular
vehicle may, in effect, subsidize the installation of a new technology
in a less marketable vehicle. But, on average, over time and across the
vehicle fleet, the retail price paid by consumers has risen by about
$1.50 for each dollar of direct costs incurred by manufacturers.
[[Page 25772]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.058
It is also important to note that direct costs associated with any
specific technology will change over time as some combination of
learning and resource price changes occurs. Resource costs, such as the
price of steel, can fluctuate over time and can experience real long-
term trends in either direction, depending on supply and demand.
However, the normal learning process generally reduces direct
production costs as manufacturers refine production techniques and seek
out less costly parts and materials for increasing production volumes.
By contrast, this learning process does not generally influence
indirect costs. The implied RPE for any given technology would thus be
expected to grow over time as direct costs decline relative to indirect
costs. The RPE for any given year is based on direct costs of
technologies at different stages in their learning cycles, and that may
have different implied RPEs than they did in previous years. The RPE
averages 1.5 across the lifetime of technologies of all ages, with a
lower average in earlier years of a technology's life, and, because of
learning effects on direct costs, a higher average in later years.
The RPE has been used in all NHTSA safety and most previous CAFE
rulemakings to estimate costs. In 2011, the National Academy of
Sciences (NAS) recommended RPEs of 1.5 for suppliers and 2.0 for in-
house production be used to estimate total costs.\141\ Auto Innovators,
formerly known as the Alliance of Automobile Manufacturers, also
advocated these values as appropriate markup factors for estimating
costs of technology changes.\142\ In their 2015 report, NAS recommended
1.5 as an overall RPE markup.\143\ An RPE of 2.0 has also been adopted
by a coalition of environmental and research groups (NESCCAF, ICCT,
Southwest Research Institute, and TIAX-LLC) in a report on reducing
heavy truck emissions, and 2.0 is recommended by the U.S. Department of
Energy for estimating the cost of hybrid-electric and automotive fuel
cell costs (see Vyas et al. (2000) in Table III-4 below). Table III-4
below also lists other estimates of the RPE. Note that all RPE
estimates vary between 1.4 and 2.0, with most in the 1.4 to 1.7 range.
---------------------------------------------------------------------------
\141\ Effectiveness and Impact of Corporate Average Fuel Economy
Standards, Washington, DC--The National Academies Press; NRC, 2011.
\142\ Communication from Chris Nevers (Auto Innovators) to
Christopher Lieske (EPA) and James Tamm (NHTSA), http://www.regulations.gov Docket ID Nos. NHTSA-2018-0067; EPA-HQ-OAR-2018-
0283, p .143.
\143\ National Research Council 2015. Cost, Effectiveness, and
Deployment of Fuel Economy Technologies for Light Duty Vehicles.
Washington, DC: The National Academies Press. https://doi.org/10.17226/21744 (hereafter, ``2015 NAS Report''). (Accessed: February
16, 2022)
---------------------------------------------------------------------------
[[Page 25773]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.059
The RPE has thus enjoyed widespread use and acceptance by a variety
of governmental, academic, and industry organizations.
---------------------------------------------------------------------------
\144\ Duleep, K.G. ``2008 Analysis of Technology Cost and Retail
Price.'' Presentation to Committee on Assessment of Technologies for
Improving Light Duty Vehicle Fuel Economy, January 25, Detroit, MI.;
Jack Faucett Associates, September 4, 1985. Update of EPA's Motor
Vehicle Emission Control Equipment Retail Price Equivalent (RPE)
Calculation Formula. Chevy Chase, MD--Jack Faucett Associates;
McKinsey & Company, October 2003. Preface to the Auto Sector Cases.
New Horizons--Multinational Company Investment in Developing
Economies, San Francisco, CA.; NRC (National Research Council),
2002. Effectiveness and Impact of Corporate Average Fuel Economy
Standards, Washington, DC--The National Academies Press; NRC, 2011.
Assessment of Fuel Economy Technologies for Light Duty Vehicles.
Washington, DC--The National Academies Press; Cost, Effectiveness,
and Deployment of Fuel Economy Technologies in Light Duty Vehicles.
Washington, DC--The National Academies Press, 2015; Sierra Research,
Inc., November 21, 2007, Study of Industry-Average Mark-Up Factors
used to Estimate Changes in Retail Price Equivalent (RPE) for
Automotive Fuel Economy and Emissions Control Systems, Sacramento,
CA--Sierra Research, Inc.; Vyas, A. Santini, D., & Cuenca, R. 2000.
Comparison of Indirect Cost Multipliers for Vehicle Manufacturing.
Center for Transportation Research, Argonne National Laboratory,
April. Argonne, Ill.
---------------------------------------------------------------------------
In past rulemakings, a second type of indirect cost multiplier has
also been examined. Known as the ``Indirect Cost Multiplier'' (ICM)
approach, ICMs were first examined alongside the RPE approach in the
2010 rulemaking regarding standards for MYs 2012-2016. Both methods
have been examined in subsequent rulemakings.
Consistent with the 2020 final rule, we continue to employ the RPE
approach to account for indirect manufacturing costs. The RPE accounts
for indirect costs like engineering, sales, and administrative support,
as well as other overhead costs, business expenses, warranty costs, and
return on capital considerations. A detailed discussion of indirect
cost methods and the basis for our use of the RPE to reflect these
costs is available in the FRIA for the 2020 final rule.\145\
---------------------------------------------------------------------------
\145\ FRIA, The Safer Affordable Fuel-Efficient (SAFE) Vehicles
Rule for Model Year 2021-2026 Passenger Cars and Light Trucks,
USDOT, EPA, March 2020, at pp. 354-76.
---------------------------------------------------------------------------
The Consumer Federation of America (CFA) noted that the inputs we
use for indirect costs produce less optimistic results than those used
by EPA. They cite these differing results as evidence that our analysis
should use the EPA values. CFA states that, ``EPA's benefit cost ratios
are much higher affirming that their analysis is more appropriate.''
\146\ CFA provided no new data or discussion to justify a conclusion
that their preferred values are justified empirically, and NHTSA
continues to believe that an RPE of 1.5 is the most justified by
empirical evidence and research, without regard to the outcomes that a
different RPE would produce. We have provided a full description of the
basis for choosing the indirect cost values that we use in Chapter
2.6.2 of the TSD accompanying this final rule, as well as in the FRIA
accompanying the 2020 final rule. In addition, we note that the RPE
value of 1.5 was also used by EPA in its regulatory impact analysis to
calculate RPE-inclusive vehicle manufacturer costs.\147\
---------------------------------------------------------------------------
\146\ CFA, Docket No. NHTSA-2021-0053-1535, at p. 5.
\147\ FRIA, Revised 2023 and Later Model Year Light-Duty Vehicle
GHG Emissions Standards: Regulatory Impact Analysis, US EPA,
December 2021, at pp. 4-8.
---------------------------------------------------------------------------
(c) Stranded Capital Costs
The idea behind stranded capital is that manufacturers amortize
research, development, and tooling expenses over many years, especially
for engines and transmissions. The traditional production life-cycles
for transmissions and engines have been a decade or longer. If a
manufacturer launches or updates a product with fuel-saving technology,
and then later replaces that technology with an unrelated or different
fuel-saving technology before the equipment and research and
development investments have been fully paid off, there will be
unrecouped, or stranded, capital costs. Quantifying stranded capital
costs accounts for such lost investments.
As DOT has observed previously, manufacturers may be shifting their
investment strategies in ways that may alter how stranded capital could
be considered. For example, some suppliers sell similar transmissions
to multiple manufacturers. Such arrangements allow manufacturers to
share in capital expenditures or amortize expenses more quickly.
Manufacturers share parts on vehicles around the globe, achieving
greater scale and greatly affecting tooling strategies and costs.
As a proxy for stranded capital in recent CAFE analyses, the CAFE
Model has accounted for platform and engine sharing and includes
redesign and refresh cycles for significant and less significant
vehicle updates. This analysis continues to rely on the CAFE Model's
explicit year-by-year accounting for estimated refresh and redesign
cycles, and shared vehicle platforms and engines, to moderate the
cadence of technology adoption and thereby limit the implied occurrence
of stranded capital and the need to account for it explicitly. In
addition, confining some manufacturers to specific advanced technology
pathways through technology adoption features acts as a proxy to
indirectly account for stranded capital. Adoption features specific to
each technology, if applied on a manufacturer-by-manufacturer basis,
are
[[Page 25774]]
discussed in each technology section. The agency will monitor these
trends to assess the role of stranded capital moving forward.
(d) Cost Learning
Manufacturers make improvements to production processes over time,
which often result in lower costs. ``Cost learning'' reflects the
effect of experience and volume on the cost of production, which
generally results in better utilization of resources, leading to higher
and more efficient production. As manufacturers gain experience through
production, they refine production techniques, raw material and
component sources, and assembly methods to maximize efficiency and
reduce production costs. Typically, a representation of this cost
learning, or learning curves, reflects initial learning rates that are
relatively high, followed by slower learning as additional improvements
are made and production efficiency peaks. This eventually produces an
asymptotic shape to the learning curve, as small percent decreases are
applied to gradually declining cost levels. These learning curve
estimates are applied to various technologies that are used to meet
CAFE standards.
We estimate cost learning by considering methods established by
T.P. Wright and later expanded upon by J.R. Crawford.148 149
Wright, examining aircraft production, found that every doubling of
cumulative production of airplanes resulted in decreasing labor hours
at a fixed percentage. This fixed percentage is commonly referred to as
the progress rate or progress ratio, where a lower rate implies faster
learning as cumulative production increases. J.R. Crawford expanded
upon Wright's learning curve theory to develop a single unit cost
model, which estimates the cost of the nth unit produced
given the following information is known: (1) Cost to produce the first
unit; (2) cumulative production of n units; and (3) the progress ratio.
---------------------------------------------------------------------------
\148\ Wright, T. P., Factors Affecting the Cost of Airplanes.
Journal of Aeronautical Sciences, Vol. 3 (1936), at pp. 124-25.
Available at https://www.uvm.edu/pdodds/research/papers/others/1936/wright1936a.pdf. (Accessed: February 16, 2022)
\149\ Crawford, J.R., Learning Curve, Ship Curve, Ratios,
Related Data, Burbank, California-Lockheed Aircraft Corporation
(1944).
---------------------------------------------------------------------------
As pictured in Figure III-5, Wright's learning curve shows the
first unit is produced at a cost of $1,000. Initially cost per unit
falls rapidly for each successive unit produced. However, as production
continues, cost falls more gradually at a decreasing rate. For each
doubling of cumulative production at any level, cost per unit declines
20 percent, so that 80 percent of cost is retained. The CAFE Model uses
the basic approach by Wright, where cost reduction is estimated by
applying a fixed percentage to the projected cumulative production of a
given fuel economy technology.
[GRAPHIC] [TIFF OMITTED] TR02MY22.060
The analysis accounts for learning effects with model year-based
cost learning forecasts for each technology that reduces direct
manufacturing costs over time. We evaluate the historical use of
technologies, and reviews industry forecasts to estimate future volumes
to develop the model year-based technology cost learning curves.
The following section discusses the development of model year-based
cost learning forecasts for this analysis, including how the approach
has evolved from the 2012 rulemaking for MY 2017-2025 vehicles, and how
the progress ratios were developed for different technologies
considered in the analysis. Finally, we discuss how these learning
effects are applied in the CAFE Model.
(l) Time Versus Volume-Based Learning
For the 2012 joint CAFE and GHG rulemaking, DOT developed learning
curves as a function of vehicle model year.\150\ Although the concept
of this methodology is derived from Wright's cumulative production
volume-based learning curve, its application for CAFE technologies was
more of a function of time. More than a dozen learning curve schedules
were developed, varying
[[Page 25775]]
between fast and slow learning, and assigned to each technology
corresponding to its level of complexity and maturity. The schedules
were applied to the base year of direct manufacturing cost and
incorporate a percentage of cost reduction by model year, declining at
a decreasing rate through the technology's production life. Some newer
technologies experience 20 percent cost reductions for introductory
model years, while mature or less complex technologies experience 0-3
percent cost reductions over a few years.
---------------------------------------------------------------------------
\150\ 77 FR 62624 (Oct. 15, 2012).
---------------------------------------------------------------------------
In their 2015 report to Congress, NAS recommended NHTSA should
``continue to conduct and review empirical evidence for the cost
reductions that occur in the automobile industry with volume,
especially for large-volume technologies that will be relied on to meet
the CAFE/GHG standards.'' \151\
---------------------------------------------------------------------------
\151\ 2015 NAS Report.
---------------------------------------------------------------------------
In response, we incorporated statically projected cumulative volume
production data of fuel economy-improving technologies, representing an
improvement over the previously used time-based method. Dynamic
projections of cumulative production are not feasible with current CAFE
Model capabilities, so one set of projected cumulative production data
for most vehicle technologies was developed for the purpose of
determining cost impact. We obtained historical cumulative production
data for many technologies produced and/or sold in the U.S. to
establish a starting point for learning schedules. Groups of similar
technologies or technologies of similar complexity may share identical
learning schedules.
The slope of the learning curve, which determines the rate at which
cost reductions occur, has been estimated using research from an
extensive literature review and automotive cost tear-down reports (see
below). The slope of the learning curve is derived from the progress
ratio of manufacturing automotive and other mobile source technologies.
(2) Deriving the Progress Ratio Used in This Analysis
Learning curves vary among different types of manufactured
products. Progress ratios can range from 70 to 100 percent, where 100
percent indicates no learning can be achieved.\152\ Learning effects
tend to be greatest in operations where workers often touch the
product, while effects are less substantial in operations consisting of
more automated processes. As automotive manufacturing plant processes
become increasingly automated, a progress ratio towards the higher end
would seem more suitable. We incorporated findings from automotive
cost-teardown studies with EPA's 2015 literature review of learning-
related studies to estimate a progress ratio used to determine learning
schedules of fuel economy-improving technologies.
---------------------------------------------------------------------------
\152\ Martin, J., ``What is a Learning Curve?'' Management and
Accounting Web, University of South Florida, available at: https://www.maaw.info/LearningCurveSummary.htm. (Accessed: February 16,
2022)
---------------------------------------------------------------------------
EPA's literature review examined and summarized 20 studies related
to learning in manufacturing industries and mobile source
manufacturing.\153\ The studies focused on many industries, including
motor vehicles, ships, aviation, semiconductors, and environmental
energy. Based on several criteria, EPA selected five studies providing
quantitative analysis from the mobile source sector (progress ratio
estimates from each study are summarized in Table III-5, below).
Further, those studies expand on Wright's learning curve function by
using cumulative output as a predictor variable, and unit cost as the
response variable. As a result, EPA determined a best estimate of 84
percent as the progress ratio in mobile source industries. However, of
those five studies, EPA at the time placed less weight on the Epple et
al. (1991) study, because of a disruption in learning due to incomplete
knowledge transfer from the first shift to introduction of a second
shift at a North American truck plant. While learning may have
decelerated immediately after adding a second shift, we note that unit
costs continued to fall as the organization gained experience operating
with both shifts. We recognize that disruptions are an essential part
of the learning process and should not, in and of themselves, be
discredited. For this reason, the analysis uses a re-estimated average
progress ratio of 85 percent from those five studies (equally
weighted).
---------------------------------------------------------------------------
\153\ Cost Reduction through Learning in Manufacturing
Industries and in the Manufacture of Mobile Sources, U.S.
Environmental Protection Agency (2015). Prepared by ICF
International and available at https://19january2017snapshot.epa.gov/sites/production/files/2016-11/documents/420r16018.pdf. (Accessed: February 16, 2022)
\154\ Argote, L., Epple, D., Rao, R. D., & Murphy, K., The
acquisition and depreciation of knowledge in a manufacturing
organization--Turnover and plant productivity, Working paper,
Graduate School of Industrial Administration, Carnegie Mellon
University (1997).
\155\ Benkard, C. L., Learning and Forgetting--The Dynamics of
Aircraft Production, The American Economic Review, Vol. 90(4), at
1034-54 (2000).
\156\ Epple, D., Argote, L., & Devadas, R., Organizational
Learning Curves--A Method for Investigating Intra-Plant Transfer of
Knowledge Acquired through Learning by Doing, Organization Science,
Vol. 2(1), at 58-70 (1991).
\157\ Epple, D., Argote, L., & Murphy, K., An Empirical
Investigation of the Microstructure of Knowledge Acquisition and
Transfer through Learning by Doing, Operations Research, Vol. 44(1),
at 77-86 (1996).
\158\ Levitt, S. D., List, J. A., & Syverson, C., Toward an
Understanding of Learning by Doing--Evidence from an Automobile
Assembly Plant, Journal of Political Economy, Vol. 121 (4), at 643-
81 (2013).
[GRAPHIC] [TIFF OMITTED] TR02MY22.061
[[Page 25776]]
In addition to EPA's literature review, this progress ratio
estimate was informed based on findings from automotive cost-teardown
studies. NHTSA routinely performs evaluations of costs of previously
issued Federal Motor Vehicle Safety Standards (FMVSS) for new motor
vehicles and equipment. NHTSA engages contractors to perform detailed
engineering ``tear-down'' analyses for representative samples of
vehicles, to estimate how much specific FMVSS add to the weight and
retail price of a vehicle. As part of the effort, the agency examines
cost and production volume for automotive safety technologies. In
particular, we estimated costs from multiple cost tear-down studies for
technologies with actual production data from the Cost and weight added
by the Federal Motor Vehicle Safety Standards for MY 1968-2012
passenger cars and LTVs (2017).\159\
---------------------------------------------------------------------------
\159\ Simons, J. F., Cost and weight added by the Federal Motor
Vehicle Safety Standards for MY 1968-2012 Passenger Cars and LTVs
(Report No. DOT HS 812 354). Washington, DC--National Highway
Traffic Safety Administration (November 2017), at pp. 30-33.
---------------------------------------------------------------------------
We chose five vehicle safety technologies with sufficient data to
estimate progress ratios of each, because these technologies are large-
volume technologies and are used by almost all vehicle manufacturers.
Table III-6 includes these five technologies and yields an average
progress rate of 92 percent.
[GRAPHIC] [TIFF OMITTED] TR02MY22.062
For the final progress ratio used in the CAFE Model, the five
progress rates from EPA's literature review and five progress rates
from NHTSA's evaluation of automotive safety technologies results were
averaged. This resulted in an average progress rate of approximately 89
percent. We placed equal weight on progress ratios from all 10 sources.
More specifically, we placed equal weight on the Epple et al. (1991)
study, because disruptions have more recently been recognized as an
essential part in the learning process, especially in an effort to
increase the rate of output.
(3) Obtaining Appropriate Baseline Years for Direct Manufacturing Costs
DOT obtained direct manufacturing costs for each fuel economy-
improving technology from various sources, as discussed above. To
establish a consistent basis for direct manufacturing costs in the
rulemaking analysis, we adjusted each technology cost to MY 2018
dollars. For each technology, the DMC is associated with a specific
model year, and sometimes a specific production volume, or cumulative
production volume. The base model year is established as the model year
in which direct manufacturing costs were assessed (with learning factor
of 1.00). With the aforementioned data on cumulative production volume
for each technology and the assumption of a 0.89 progress ratio for all
automotive technologies, we can solve for an implied cost for the first
unit produced. For some technologies, we used modestly different
progress ratios to match detailed cost projections if available from
another source (for instance, batteries for plug-in hybrids and battery
electric vehicles).
This approach produces reasonable estimates for technologies
already in production, and some additional steps are required to set
appropriate learning rates for technologies not yet in production.
Specifically, for technologies not yet in production in MY 2017, the
cumulative production volume in MY 2017 is zero, because manufacturers
have not yet produced the technologies. For pre-production cost
estimates in previous CAFE rulemakings, we often relied on confidential
business information sources to predict future costs. Many sources for
pre-production cost estimates include significant learning effects,
often providing cost estimates assuming high volume production, and
often for a timeframe late in the first production generation or early
in the second generation of the technology. Rapid doubling and re-
doubling of a low cumulative volume base with Wright's learning curves
can provide unrealistic cost estimates. In addition, direct
manufacturing cost projections can vary depending on the initial
production volume assumed. Accordingly, we carefully examined direct
costs with learning, and made adjustments to the starting point for
those technologies on the learning curve to better align with the
assumptions used for the initial direct cost estimate.
(4) Cost Learning Applied in the CAFE Model
For this analysis, we apply learning effects to the incremental
cost over the null technology state on the applicable technology tree.
After this step, we calculate year-by-year incremental costs over
preceding technologies on the tech tree to create the CAFE Model
inputs.\160\ The shift from incremental cost accounting to absolute
cost accounting in recent CAFE analyses made cost inputs more
transparently relatable to detailed model output, and relevant to this
discussion, made it easier to apply learning curves in the course of
developing inputs to the CAFE Model.
---------------------------------------------------------------------------
\160\ These costs are located in the CAFE Model Technologies
file.
---------------------------------------------------------------------------
We group certain technologies, such as advanced engines, advanced
transmissions, and non-battery electric components and assign them to
the same learning schedule. While these grouped technologies differ in
operating characteristics and design, we chose to group them based on
their complexity, technology integration, and economies of scale across
manufacturers. The low volume of certain advanced technologies, such as
hybrid and electric technologies, poses a significant issue for
suppliers and prevents them
[[Page 25777]]
from producing components needed for advanced transmissions and other
technologies at more efficient high scale production. The technology
groupings consider market availability, complexity of technology
integration, and production volume of the technologies that can be
implemented by manufacturers and suppliers. The details of these
technologies are discussed in Section III.D.
In addition, we expanded model inputs to extend the explicit
simulation of technology application through MY 2050. Accordingly, we
updated the learning curves for each technology group to cover MYs
through 2050. For MYs 2017-2032, we expect incremental improvements in
all technologies, particularly in electrification technologies because
of increased production volumes, labor efficiency, improved
manufacturing methods, specialization, network building, and other
factors. While these and other factors contribute to continual cost
learning, we believe that many fuel economy-improving technologies
considered in this rule will approach a flat learning level by the
early 2030s. Specifically, older, and less complex internal combustion
engine technologies and transmissions will reach a flat learning curve
sooner when compared to electrification technologies, which have more
opportunity for improvement. For batteries and non-battery
electrification components, we estimated a steeper learning curve that
will gradually flatten after MY 2040. For a more detailed discussion of
the electrification learning curves, see Section III.D.3.
Each technology in the CAFE Model is assigned a learning schedule
developed from the methodology explained previously. For example, the
following chart shows learning rates for several technologies
applicable to midsize sedans, demonstrating that while we estimate that
such learning effects have already been almost entirely realized for
engine turbocharging (a technology that has been in production for many
years), we estimate that significant opportunities to reduce the cost
of the greatest levels of mass reduction (e.g., MR5) remain, and even
greater opportunities remain to reduce the cost of batteries for HEVs,
PHEVs, BEVs. In fact, for certain advanced technologies, we determined
that the results predicted by the standard learning curves progress
ratio was not realistic, based on unusual market price and production
relationships. For these technologies, we developed specific learning
estimates that may diverge from the 0.89 progress rate. As shown in
Figure III-6, these technologies include: Turbocharging and downsizing
level 1 (TURBO1), variable turbo geometry electric (VTGE), aerodynamic
drag reduction by 15 percent (AERO15), mass reduction level 5 (MR5), 20
percent improvement in low-rolling resistance tire technology (ROLL20)
over the baseline, and belt integrated starter/generator (BISG).
BILLING CODE 4910-59-P
[GRAPHIC] [TIFF OMITTED] TR02MY22.063
[[Page 25778]]
BILLING CODE 4910-59-C
CFA noted that the inputs we use for learning rates produce less
optimistic results than those used by EPA. They cite these differing
results as evidence that NHTSA should use the EPA values. CFA states
that, ``EPA's benefit cost ratios are much higher affirming that their
analysis is more appropriate.'' \161\ CFA provided no new data or
discussion to justify a conclusion that their preferred values are
justified empirically, and NHTSA continues to believe that the
appropriate values to use in estimating the impacts of CAFE standards
are those most justified by empirical evidence and research, consistent
with E.O. 12866, without reference to the outcomes they produce. We
have provided a full description of the basis for choosing the learning
values that we use in Chapter 2.6.4 of the TSD accompanying this final
rule, as well as in the FRIA accompanying the 2020 final rule.
---------------------------------------------------------------------------
\161\ CFA, at p. 5.
---------------------------------------------------------------------------
(e) Cost Accounting
To facilitate specification of detailed model inputs and review of
detailed model outputs, the CAFE Model continues to use absolute cost
inputs relative to a known base component cost, such that the estimated
cost of each technology is specified relative to a common reference
point for the relevant technology pathway. For example, the cost of a
7-speed transmission is specified relative to a 5-speed transmission,
as is the cost of every other transmission technology. Conversely, in
some earlier versions of the CAFE Model, incremental cost inputs were
estimated relative to the technology immediately preceding on the
relevant technology pathway. For our 7-speed transmission example, the
incremental cost would be relative to a 6-speed transmission. This
change in the structure of cost inputs does not, by itself, change
model results, but it does make the connection between these inputs and
corresponding outputs more transparent. The CAFE Model Documentation
accompanying our analysis presents details of the structure for model
cost inputs.\162\ The individual technology sections in Section III.D
provide a detailed discussion of cost accounting for each technology.
---------------------------------------------------------------------------
\162\ CAFE Model Documentation, S4.7.
---------------------------------------------------------------------------
7. Manufacturer's Credit Compliance Positions
This rule involves a variety of provisions regarding ``credits''
and other compliance flexibilities. Some regulatory provisions allow a
manufacturer to earn ``credits'' that will be counted toward a
vehicle's rated CO2 emissions level, or toward a fleet's
rated average CO2 or CAFE level, without reference to
required levels for these average levels of performance. Such
flexibilities effectively modify emissions and fuel economy test
procedures or methods for calculating fleets' CAFE and average
CO2 levels. Other provisions (for CAFE, statutory
provisions) allow manufacturers to earn credits by achieving CAFE or
average CO2 levels beyond required levels; these provisions
may hence more appropriately be termed ``compliance credits.'' We
described in the 2020 final rule how the CAFE Model simulates these
compliance credit provisions for both the CAFE program and for EPA's
CO2 standards.\163\ For this analysis, we modeled the No-
Action and Action Alternatives as a set of CAFE standards in place
simultaneously with EPA's 2020 final rule CO2
standards,\164\ related CARB agreements with five manufacturers, and
ZEV mandates in place in California and some other states. The modeling
of CO2 standards and standard-like contractual obligations
includes our representation of applicable credit provisions.
---------------------------------------------------------------------------
\163\ See 85 FR 24174, 24303 (April 30, 2020).
\164\ The baseline for this analysis is the set of standards in
place when NHTSA initiated this rulemaking.
---------------------------------------------------------------------------
EPCA has long provided that, by exceeding the CAFE standard
applicable to a given fleet in a given model year, a manufacturer may
earn corresponding ``credits'' that the same manufacturer may, within
the same regulatory class, apply toward compliance in a different model
year. EISA amended these provisions by providing that manufacturers
may, subject to specific statutory limitations, transfer compliance
credits between regulatory classes and trade compliance credits with
other manufacturers. Under the CAA, EPA has broad standard-setting
authority and has long provided for averaging, banking, and trading
programs in certain circumstances, and in particular for GHGs.
EPCA also specifies that NHTSA may not consider the availability of
CAFE credits (for transfer, trade, or direct application) toward
compliance with new standards when establishing the standards
themselves.\165\ Therefore, this analysis excludes MYs 2024-2026 from
those in which carried-forward or transferred credits can be applied
for the CAFE program.
---------------------------------------------------------------------------
\165\ 49 U.S.C. 32902(h)(3).
---------------------------------------------------------------------------
The ``unconstrained'' perspective acknowledges that these
flexibilities exist as part of the program and, while not considered by
NHTSA in setting standards, are nevertheless important to consider when
attempting to estimate the real impact of any alternative. Under the
``unconstrained'' perspective, credits may be earned, transferred, and
applied to deficits in the CAFE program throughout the full range of
model years in the analysis. The Final SEIS accompanying this rule
presents ``unconstrained'' modeling results. Also, consistent with the
program EPA established under the CAA, this analysis includes
simulation of carried-forward and transferred CO2 credits in
all model years.
The CAFE Model, therefore, does provide means to simulate
manufacturers' potential application of some compliance credits, and
both the analysis of CO2 standards and the NEPA analysis of
CAFE standards do make use of this aspect of the model. On the other
hand, 49 U.S.C. 32902(h) prevents NHTSA from, in its standard setting
analysis, considering the potential that manufacturers could use
compliance credits in model years for which the agency is establishing
maximum feasible CAFE standards. Further, as discussed below, we also
continue to find it appropriate for the analysis largely to refrain
from simulating two of the mechanisms allowing the use of compliance
credits.
The CAFE Model's approach to simulating compliance decisions
accounts for the potential to earn and use CAFE credits as provided by
EPCA/EISA. The model similarly accumulates and applies CO2
credits when simulating compliance with EPA's standards. Like past
versions, the current CAFE Model can simulate credit carry-forward
(i.e., banking) between model years and transfers between the passenger
car and light truck fleets but not credit carry-back (i.e., borrowing)
from future model years or trading between manufacturers.
While NHTSA's ``unconstrained'' evaluation can consider the
potential to carry back compliance credits from later to earlier model
years, past examples of failed attempts to carry back CAFE credits
(e.g., a MY 2014 carry back default leading to a civil penalty payment)
underscore the riskiness of such ``borrowing.'' Recent evidence
indicates manufacturers are disinclined to take such risks, and we find
it reasonable and prudent to refrain from attempting to simulate such
``borrowing'' in rulemaking analysis.
Like the previous version, the current CAFE Model provides a basis
to specify (in model inputs) CAFE credits available from model years
earlier than
[[Page 25779]]
those being explicitly simulated. For example, with this analysis
representing MYs 2020-2050 explicitly, credits earned in the MY 2015
are made available for use through the MY 2020 (given the current five-
year limit on carry-forward of credits). The banked credits are
specific to both the model year and fleet in which they were earned.
To increase the realism with which the model transitions between
the early model years (MYs 2020-2023) and the later years that are the
subject of this action, we have accounted for the potential that some
manufacturers might trade credits earned prior to 2020 to other
manufacturers. However, the analysis refrains from simulating the
potential that manufacturers might continue to trade credits during and
beyond the model years covered by this action. In 2018 and 2020, the
analysis included idealized cases simulating ``perfect'' (i.e., wholly
unrestricted) trading of CO2 compliance credits by treating
all vehicles as being produced by a single manufacturer. Even for
CO2 compliance credit trading, these scenarios were not
plausible, because it is exceedingly unlikely that some pairs of
manufacturers would trade compliance credits. NHTSA did not include
such cases for CAFE compliance credits, because EPCA provisions (such
as the minimum domestic passenger car standard requirement) make such
scenarios impossible. At this time, we remain concerned that any
realistic simulation of such trading would require assumptions
regarding which specific pairs of manufacturers might trade compliance
credits, and the evidence to date makes it clear that the credit market
is far from fully ``open.'' \166\
---------------------------------------------------------------------------
\166\ See, Automotive Innovators, NHTSA-2021-0053-1492, at p.
73.
---------------------------------------------------------------------------
We also remain concerned that to set standards based on an analysis
that presumes the use of program flexibilities risks making the
corresponding actions mandatory. Some flexibilities--credit carry-
forward (banking) and transfers between fleets in particular--involve
little risk because they are internal to a manufacturer and known in
advance. As discussed above, credit carry-back involves significant
risk because it amounts to borrowing against future improvements,
standards, and production volume and mix. Similarly, credit trading may
also involve significant risk, because the ability of manufacturer A to
acquire credits from manufacturer B depends not just on manufacturer B
actually earning the expected amount of credit, but also on
manufacturer B being willing to trade with manufacturer A, and on
potential interest by other manufacturers. Manufacturers' compliance
plans have already evidenced cases of compliance credit trades that
were planned and subsequently aborted, reinforcing our judgment that,
like credit borrowing, credit trading involves too much risk to be
included in an analysis that informs decisions about the stringency of
future standards. NHTSA will continue to carefully monitor
manufacturers' practices regarding use of credit trading and other
flexibilities to ensure that future analyses appropriately account for
realistic market conditions and statutory requirements as applicable.
As discussed in the CAFE Model Documentation, the model's default
logic attempts to maximize credit carry-forward--that is, to ``hold
on'' to credits for as long as possible. If a manufacturer needs to
cover a shortfall that occurs when insufficient opportunities exist to
add technology to achieve compliance with a standard, the model will
apply credits. Otherwise, the manufacturer carries forward credits
until they are about to expire, at which point it will use them before
adding technology that is not considered cost-effective. The model
attempts to use credits that will expire within the next three years as
a means to smooth out technology applications over time to avoid both
compliance shortfalls and high levels of over-compliance that can
result in a surplus of credits. Although it remains impossible
precisely to predict the manufacturer's actual earning and use of
compliance credits, and this aspect of the model may benefit from
future refinement as manufacturers and regulators continue to gain
experience with these provisions, this approach is generally consistent
with manufacturers' observed practices.
NHTSA introduced the CAFE Public Information Center (PIC) to
provide public access to a range of information regarding the CAFE
program,\167\ including manufacturers' credit balances. However, there
is a data lag in the information presented on the CAFE PIC that may not
capture credit actions across the industry for as much as several
months. Furthermore, CAFE credits that are traded between manufacturers
are adjusted to preserve the gallons saved that each credit
represents.\168\ The adjustment occurs at the time of application
rather than at the time the credits are traded. This means that a
manufacturer who has acquired credits through trade, but has not yet
applied them, may show a credit balance that is either considerably
higher or lower than the real value of the credits when they are
applied. For example, a manufacturer that buys 40 million credits from
Tesla may show a credit balance in excess of 40 million. However, when
those credits are applied, they may be worth only \1/10\ as much--
making that manufacturer's true credit balance closer to 4 million than
40 million (e.g., when another manufacturer uses credits acquired from
Tesla, the manufacturer may only be able to offset a 1 mpg compliance
shortfall, even though the credits' ``face value'' suggests the
manufacturer could offset a 10-mpg compliance shortfall).
---------------------------------------------------------------------------
\167\ CAFE Public Information Center, https://one.nhtsa.gov/cafe_pic/home (accessed: March 6, 2022).
\168\ CO2 credits for EPA's program are denominated
in metric tons of CO2 rather than gram/mile compliance
credits and require no adjustment when traded between manufacturers
or fleets.
---------------------------------------------------------------------------
Specific inputs accounting for manufacturers' accumulated
compliance credits are discussed in TSD Chapter 2.
In addition to the inclusion of these existing credit banks, the
CAFE Model also updated its treatment of credits in the rulemaking
analysis. EPCA requires that NHTSA set CAFE standards at maximum
feasible levels for each model year without consideration of the
program's credit mechanisms. However, as recent CAFE rulemakings have
evaluated the effects of standards over longer time periods, the early
actions taken by manufacturers required more nuanced representation.
Accordingly, the CAFE Model now provides means to exclude the simulated
application of CAFE compliance credits only from specific model years
for which standards are being set (for this analysis, 2024-2026), while
allowing CAFE credits to be applied in other model years.
In addition to more rigorous accounting of CAFE and CO2
compliance credits, the model also accounts for air conditioning
efficiency and off-cycle adjustments. NHTSA's program considers those
adjustments in a manufacturer's compliance calculation starting in MY
2017, and specific estimates of each manufacturer's reliance on these
adjustments are discussed above in Section III.C.2.a). Because air
conditioning efficiency and off-cycle adjustments are not credits in
NHTSA's program, but rather adjustments to compliance fuel economy,
they may be included under either a ``standard setting'' or
``unconstrained'' analysis perspective.
[[Page 25780]]
The manner in which the CAFE Model treats the EPA and CAFE AC
efficiency and off-cycle credit programs is similar, but the model also
accounts for AC leakage (which is not part of NHTSA's program). When
determining the compliance status of a manufacturer's fleet (in the
case of EPA's program, PC and LT are the only fleet distinctions), the
CAFE Model weighs future compliance actions against the presence of
existing (and expiring) CO2 credits resulting from over-
compliance with earlier years' standards, AC efficiency credits, AC
leakage credits, and off-cycle credits.
The model currently accounts for any off-cycle adjustments
associated with technologies that are included in the set of fuel-
saving technologies simulated explicitly (for example, start-stop
systems that reduce fuel consumption during idle or active grille
shutters that improve aerodynamic drag at highway speeds) and
accumulates these adjustments up to levels defined in the Market Data
file. As discussed further in Section III.D.8, this analysis considers
that some manufacturers may apply up to 15.0 g/mi of off-cycle credit
by MY 2032. We considered the potential to model the application of
off-cycle technologies explicitly. However, doing so would require data
regarding which vehicle models already possess these improvements as
well as the cost and expected value of applying them to other models in
the future. Such data are currently too limited to support explicit
modeling of these technologies and adjustments.
When establishing maximum feasible fuel economy standards, NHTSA is
prohibited from considering the availability of alternatively fueled
vehicles,\169\ and credit provisions related to AFVs that significantly
increase their fuel economy for CAFE compliance purposes. Under the
``standard setting'' perspective, these technologies (pure battery
electric vehicles and fuel cell vehicles \170\) are not available in
the compliance simulation to improve fuel economy. Under the
``unconstrained'' perspective, such as is documented in the Final SEIS,
the CAFE Model considers these technologies in the same manner as other
available technologies and may apply them if they represent cost-
effective compliance pathways. However, under both perspectives, the
analysis continues to include dedicated AFVs that could be produced in
response to CAFE standards outside the model years for which standards
are being set, or for other reasons (e.g., ZEV mandates, as accounted
for in this analysis).
---------------------------------------------------------------------------
\169\ 49 U.S.C. 32902(h).
\170\ Dedicated compressed natural gas (CNG) vehicles should
also be excluded in this perspective but are not considered as a
compliance strategy under any perspective in this analysis.
---------------------------------------------------------------------------
EPCA also provides that CAFE levels may, subject to limitations, be
adjusted upward to reflect the sale of flexible fuel vehicles (FFVs).
Because these adjustments ended in MY 2020, this analysis assumes no
manufacturer will earn FFV credits within the modeling horizon.
In contrast, the CAA allows consideration of alternative fuels, and
EPA has provided that manufacturers selling PHEVs, BEVs, and FCVs may,
when calculating fleet average CO2 levels, ``count'' each
unit of production as more than a single unit. The CAFE Model accounts
for these ``multipliers.''
There were no natural gas vehicles in the baseline fleet, and the
analysis did not apply natural gas technology due to cost
effectiveness. The application of production multipliers for natural
gas vehicles for MY 2022 would have no impact on the analysis because
given the state of natural gas vehicle refueling infrastructure, the
cost to equip vehicles with natural gas tanks, the outlook for
petroleum prices, and the outlook for battery prices, we have little
basis to project more than an inconsequential response to this
incentive in the foreseeable future.
D. Technology Pathways, Effectiveness, and Cost
Vehicle manufacturers meet increasingly stringent fuel economy
standards by applying additional fuel-economy-improving technologies to
their vehicles. To assess what increases in fuel economy standards
could be achievable at what cost, we first need accurate
characterizations of fuel-economy-improving technologies. We collected
data on over 50 fuel-economy-improving technologies that manufacturers
could apply to their vehicles to meet future stringency levels. This
includes determining technology effectiveness values, technology costs,
and how we realistically expect manufacturers could apply the
technologies in the rulemaking timeframe. The characterizations of
these fuel-economy-improving technologies are built on work performed
by DOT, EPA, NAS, and other Federal and state government agencies
including the Department of Energy's Argonne National Laboratory and
the California Air Resources Board.
In the NPRM we described spending approximately a decade refining
the technology pathways, effectiveness, and cost assumptions used in
successive CAFE Model analyses. We discussed developing guiding
principles to ensure the CAFE Model reasonably simulates manufacturers'
possible real-world compliance behavior. These guiding principles are
as follows:
The fuel economy improvement from any individual technology must be
considered in conjunction with any other fuel-economy-improving
technologies applied to the vehicle. Certain technologies will have
complementary or non-complementary interactions with the full vehicle
technology system. For example, there is an obvious fuel economy
benefit that results from converting a vehicle with a traditional
internal combustion engine to a battery electric vehicle; however, the
benefit of the electrification technology depends on the other road
load reducing technologies (i.e., mass reduction, aerodynamic, and
rolling resistance) on the vehicle.
Technologies added in combination to a vehicle will not result in a
simply additive fuel economy improvement from each individual
technology. As discussed in Section III.C.4, full vehicle modeling and
simulation provides the required degree of accuracy to project how
different technologies will interact in the vehicle system. For
example, as discussed further in Sections III.D.1 and III.D.3, a
parallel hybrid architecture powertrain improves fuel economy, in part,
by allowing the internal combustion engine to spend more time operating
at efficient engine speed and load conditions. This reduces the
advantage of adding advanced internal combustion engine technologies,
which also improve fuel economy, by broadening the range of speed and
load conditions for the engine to operate at high efficiency. This
redundancy in fuel savings mechanism results in a reduced effectiveness
improvement when the technologies are added to each other.
The effectiveness of a technology depends on the type of vehicle
the technology is being applied to. For example, applying mass
reduction technology results in varying effectiveness as the absolute
mass reduced is a function of the starting vehicle mass, which varies
across vehicle technology classes. See Section III.D.4 for more
details.
The cost and effectiveness values for each technology should be
reasonably representative of what can be achieved across the entire
industry. Each technology model employed in the analysis is designed to
be representative of a wide range of specific technology applications
used in industry. Some vehicle manufacturer's systems may
[[Page 25781]]
perform better and cost less than our modeled systems and some may
perform worse and cost more. However, employing this approach will
ensure that, on balance, the analysis captures a reasonable level of
costs and benefits that would result from any manufacturer applying the
technology.
The baseline for cost and effectiveness values must be identified
before assuming that a cost or effectiveness value could be employed
for any individual technology. For example, as discussed further in
Section III.D.1.d) below, this analysis uses a set of engine map models
that were developed by starting with a small number of baseline engine
configurations, and then, in a very systematic and controlled process,
adding specific well-defined technologies to create a new map for each
unique technology combination.
Historically, we have received comments concerned with specific
technology assumptions, such as technology effectiveness or cost, or
how we applied adoption features. In response to this proposal,
however, commenters instead focused on broader portions of our modeling
approach. Specifically, we received comments about the range of
technologies considered on the advanced engine technology pathway and
hybrid/electric pathway, considering the potential future of light duty
vehicle fuel economy and greenhouse gas emissions regulations. We did
still receive some comments regarding specific technology values, but
fewer than previous rules.\171\
---------------------------------------------------------------------------
\171\ Comments regarding specific technology modeling values,
such as battery cost, strong hybrid electric vehicle costs, and high
compression ratio engine adoptions features are addressed under
their respective paragraphs below.
---------------------------------------------------------------------------
Vehicle manufacturers emphasized the diminishing returns to
investing in advanced internal combustion engine technologies, and a
current trend of shifting resources from ICE development into
electrification technologies. Ford Motor Company (Ford) commented that
``[t]he transformation of the light-duty fleet toward electrification
will require unprecedented levels of ingenuity and investment to
succeed. Over the last 10 years, rapid improvements in internal
combustion engine (ICE) fuel efficiency and criteria emissions
performance have been accomplished. Further improvements are possible,
but will be marginal, and will come at high cost.'' \172\ Similarly,
Volkswagen Group of America (Volkswagen) commented that they have
``publicly stated that investments into combustion technologies will
wane with a point in the next several years where there will be no new
combustion engine families developed for the Group. Volkswagen
recognizes that remaining combustion models will continue to be sold in
high volume for the next several years and that it is important to
preserve the fuel economy of remaining ICEs as electrification volumes
increase. As noted earlier, Volkswagen's remaining ICE engines will
[sic]primary focus on evolutions of existing downsized, charged engines
to incorporate incremental hardware and software improvements.'' \173\
Toyota Motor North America, Inc. (Toyota) also commented that ``data
has consistently documented that even advanced ICE-only powertrains
will fall short of the proposed standards and that while future
advancements are possible, a point of diminishing returns is in part
driving the transition to electrified powertrains, including
conventional hybrids.'' \174\
---------------------------------------------------------------------------
\172\ Ford, Docket No. NHTSA-2021-0053-1545-A1, at p. 1.
\173\ Volkswagen, Docket No. NHTSA-2021-0053-1548-A1, at pp. 21-
22.
\174\ Toyota, Docket No. NHTSA-2021-0053-1568, at p. 2.
---------------------------------------------------------------------------
In contrast, Union of Concerned Scientists (UCS) acknowledged that
``given automaker investments and future product plans, it is likely
that manufacturers' compliance strategies will include increased
electrification. However, there are significant opportunities for
improvements to internal combustion engine vehicles as well.'' \175\
Similarly, ICCT provided examples of vehicle technologies that can
``boost ICE efficiency well beyond even HCR2 efficiency levels,''
including technologies that are not modeled in the analysis like
negative valve overlap (NVO) fuel reforming, passive prechamber
engines, and high energy ignition systems.\176\ Borg Warner also
provided hydrogen combustion as ``an advanced technology that has been
under development for some time and could be more rapidly deployed in
high volumes to make an impact.'' \177\
---------------------------------------------------------------------------
\175\ UCS, Docket No. NHTSA-2021-0053-1567-A1, at p. 6.
\176\ ICCT, Docket No. NHTSA-2021-0053-1581-A1, at p. 2.
\177\ BorgWarner Inc. (BorgWarner), Docket No. NHTSA-2021-0053-
1473, at p. 2.
---------------------------------------------------------------------------
First and foremost, we want to emphasize that the purpose of this
regulation is to set maximum feasible CAFE standards for passenger cars
and light trucks that improve energy conservation, and not to advocate
for specific technology solutions. We acknowledge that the industry is
not going to quickly abandon ICE technologies and we anticipate
improvements in those vehicles for years to come; however, we also
acknowledge that many manufacturers have announced significant shifts
in product line-up, moving toward electrification technologies and
likely slowing the rate of new ICE technology introduction.\178\ That
said, we agree with comments urging us to staying abreast of the
feasibility of advanced engine and other powertrain technologies. For
this analysis we evaluated over 50 different technologies for
effectiveness and cost and continue to research the feasibility of
additional technology models. However, we also agree with comments
regarding constraining some advanced technology options as an
acknowledgment of the realities of limited investment resources.
Accordingly, we expect an actual pathway to compliance in the
rulemaking timeframe to fall somewhere between the extremes suggested
by the commenters above. This expectation is discussed further in the
results/legal justification section \179\ and in the engine technology
section.\180\
---------------------------------------------------------------------------
\178\ ``Mercedes-Benz Prepares to Go All-Electric,'' Mercedes-
Benz Media Newsroom USA (Jul. 22, 2021), https://media.mbusa.com/releases/release-ee5a810c1007117e79e1c871354679e4-mercedes-benz-prepares-to-go-all-electric (accessed: February 16, 2022).
``Investments into combustion engines and plug-in hybrid
technologies will drop by 80% between 2019 and 2026.''; Hannah Lutz,
``Shifting into E,'' Automotive News (Jul. 26, 2021). ``Some
existing vehicles, such as the Chevy Malibu and Camaro, won't stick
to the standard cadence of face-lifts and redesigns. Instead,
they'll ride out the current generation before making way for EVs.''
Jordyn Grzelewski, ``Ford Slated to Spend More On EVs Than On
Internal Combustion Engine Vehicles in 2023,'' The Detroit News
(Aug. 2, 2021).; Lindsay Chappell, ``All-In On EVs,'' Automotive
News (May 17, 2021). ``Mini will become an all-electric brand by
early 2030, and the British marque will roll out its last new
combustion engine variant in 2025.'' (Emphasis added); Bibhu
Pattnaik, ``Audi Will Not Introduce ICE Vehicles After 2026, No
Hybrid Vehicles Either,'' Benzinga (Jun. 19, 2021), https://finance.yahoo.com/news/audi-not-introduce-ice-vehicles-160320055.html (accessed February 16, 2022); Mike Colias, ``Gas
Engines, and the People Behind Them, Are Cast Aside for Electric
Vehicles,'' The Wall Street Journal (Jul. 23, 2021). ``Auto
executives have concluded, to varying degrees, that they can't meet
tougher tailpipe-emission rules globally by continuing to improve
gas or diesel engines . . . Over the past several decades, auto
makers in most years rolled out between 20 and 70 new engines
globally, according to research firm IHS Markit. That number will
fall below 10 this year, and then essentially go to zero, the
research firm said.''
\179\ See Section VI.
\180\ See Section III.D.1.
---------------------------------------------------------------------------
As a result, we believe the range of technologies modeled on the
advanced engine technologies and hybrid/electric pathways appropriately
represent the range of technologies that will be available in the
rulemaking time frame. The technologies in our analysis are
[[Page 25782]]
based on guidance from NAS \181\ and align with technologies considered
by the EPA as part of their final rulemaking for MYs 2023-2026.\182\
---------------------------------------------------------------------------
\181\ 2021 NAS Report.
\182\ For detailed discussions on all the technologies used in
this analysis see TSD Chapter 3, For more detailed discussion of the
comments discussed here see Section III.D.1.
---------------------------------------------------------------------------
However, the CAFE Model is a tool that offers many ways to evaluate
a cost-effective technology pathway for vehicle manufacturers to reach
given levels of CAFE standards, based on user-provided inputs and
constraints. As a result of the concerns expressed in the comments
above, we included a sensitivity analysis with inputs assuming that
vehicle manufacturers would no longer deploy advanced engine
technologies.\183\ The sensitivity analysis demonstrates a technology
path where manufacturers choose to stop applying additional ICE
improvements and only invest in partial or full electrification
technologies going forward.\184\ Our ``no advanced engines''
sensitivity analysis shows a modest increase in strong hybrid (SHEV)
and plug-in hybrid (PHEV) technology adoption compared to the reference
analysis. This modest increase, about 5-6 percent increased technology
penetration of SHEVs and PHEVs, enables the manufacturers to meet more
stringent standards without the adoption of additional advanced ICE
technology. The ``no advanced engine'' technology pathway increases the
estimated average vehicle costs by $25 over the reference analysis by
MY 2029.\185\
---------------------------------------------------------------------------
\183\ See TSD Chapter 3.1 for a definition of advanced engine
technologies.
\184\ See FRIA Chapter 7.1 for more details; the sensitivity
case ``conv-tech-imprlimited'' is referred to as ``no advanced
engine'' in this discussion.
\185\ Effects of standards on the fleet out to MY 2029 are
considered to account for years the regulation covers, and years of
potential carry back credit use.
---------------------------------------------------------------------------
In consideration of comments received on the NPRM analysis and the
results of additional sensitivity analysis, we believe that the
technologies included in the CAFE Model's technology tree are currently
appropriate, and we have made no changes in the technology tree for the
analysis supporting this final rule. We believe the selected
technologies provide a realistic representation of options that
manufacturers have to comply with standards in the rulemaking
timeframe.
We made changes to just three technology inputs from the NPRM to
this final rule. The changes are discussed in detail in the respective
technology sections, and include:
Decreased eCVT and cable costs associated with strong
hybrid electric vehicle technologies;
Decreased start/stop micro hybrid battery costs; and
Correction of the high compression ratio with cylinder
deactivations setting in the Technologies input file.
The following sections discuss the engine, transmission,
electrification, mass reduction, aerodynamic, tire rolling resistance,
and other vehicle technologies considered in this analysis. Each
section discusses how we define the technology in the CAFE Model,\186\
how we assign the technology to vehicles in the MY 2020 analysis fleet
used as a starting point for this analysis, any adoption features that
we apply to the technology so the analysis better represents
manufacturers' real-world decisions, the technology effectiveness
values, and technology cost. In addition, each section discusses the
comments received for that technology pathway, and the changes made to
input values because of comments.
---------------------------------------------------------------------------
\186\ Note, due to the diversity of definitions industry uses
for technology terms, or in describing the specific application of
technology, the terms defined here may differ from how the
technology is defined in the industry.
---------------------------------------------------------------------------
Please note that the following technology effectiveness sections
provide examples of the range of effectiveness values that a technology
could achieve when applied to the entire vehicle system, in conjunction
with the other fuel-economy-improving technologies already in use on
the vehicle.\187\ To see the incremental effectiveness values for any
particular vehicle moving from one technology key to a more advanced
technology key, see the FE_1 and FE_2 Adjustments files that are
integrated in the CAFE Model executable file. Similarly, the technology
costs provided in each section are examples of absolute costs seen in
specific model years (MYs 2020, 2025, and 2030 for most technologies),
for specific vehicle classes.\188\ Please refer to the Technologies
file to see all absolute technology costs used in the analysis across
all model years.
---------------------------------------------------------------------------
\187\ This serves as a visual example of the conditional
effectiveness of adding `one technology at a time' discussed in the
guiding principles above.
\188\ The values shown serve as examples of cost origins and how
cost values were treated to account for changes due to learning or
time value of money.
---------------------------------------------------------------------------
1. Engine Paths
We classified the extensive variety of light duty vehicle internal
combustion (IC) engine technologies into discrete engine technology
paths for this analysis. These engine technology paths model the most
representative characteristics, costs, and performance of the fuel-
economy improving technologies likely available during the rulemaking
time frame. It is our intent that the technology paths be
representative of the range of potential performance levels for each of
the technologies. We also acknowledge that some new and pre-production
technologies are not part of this analysis because of uncertainties in
the cost and capabilities of these emerging technologies. As a result,
we did not include technologies unlikely to be feasible in the
rulemaking timeframe, technologies unlikely to be compatible with U.S.
fuels, or technologies where there were not appropriate data available
to allow the simulation of effectiveness across all vehicle technology
classes in this analysis.
We briefly discuss IC engine technologies considered in this
analysis, the CAFE Model's general engine technology categories, and
how we assign engine technologies in the analysis fleet in the
following sections. We also touch on engine technologies' adoption
features, costs, and effectiveness when used as part of a full vehicle
model. For a complete discussion on all of these topics please see the
TSD.\189\
---------------------------------------------------------------------------
\189\ See TSD Chapter 3.1.
---------------------------------------------------------------------------
(a) Engine Modeling in the CAFE Model
Engine modeling in the CAFE Model involves the application of
internal combustion engine technologies that manufacturers use to
improve fuel economy. Of the engine technologies we model, some can be
incorporated into existing engines with minor or moderate changes, but
many require an entirely new engine architecture. As a result, we
divide engine technologies into two categories, ``basic engine
technologies'' and ``advanced engine technologies.'' ``Basic engine
technologies'' refer to technologies adaptable to an existing engine
with minor or moderate changes to the engine. ``Advanced engine
technologies'' refer to technologies that generally require significant
changes or an entirely new engine architecture.
We do not intend for the words ``basic'' and ``advanced'' to confer
any information about the level of sophistication of the technology or
to indicate relative cost. Many advanced engine technology definitions
include some basic engine technologies in their design, and these basic
technologies are accounted for in the costs and effectiveness values of
the advance engine. Figure III-7 shows how we organize the engine
technologies pathways evaluated in the compliance simulation. We
briefly describe each
[[Page 25783]]
engine technology below. It is important to note the ``Basic Engine
Path'' shows that every engine starts with VVT and can add one, some,
or all of the technologies in the dotted box, as discussed in Section
III.D.1.a)(1).
[GRAPHIC] [TIFF OMITTED] TR02MY22.064
In response to our proposal, some commenters, particularly in the
automotive industry, commented in support of the number of advanced
engine technologies in the engine tree especially in light of
forthcoming electrification investments. Other commenters, in
particular some environmental groups, commented with examples of
advanced engine technologies that they believed we should consider in
the analysis.
More specifically, the automotive industry believes that the future
of ICE technology is very limited, as manufacturers turn their focus to
the electrification of the fleet. The new focus would result in
limitation or even removal of resources dedicated to further ICE
development. Major manufacturers provided information indicating that
they will not develop advanced engine technologies beyond the current
generation. Commenters who provided information suggesting engine
technology may stagnate as manufacturers dedicate resources to
electrification technology included Ford, Toyota, Volkswagen, and the
Auto Innovators.
Ford stated:
Over the last 10 years, rapid improvements in internal
combustion engine (ICE) fuel efficiency and criteria emissions
performance have been accomplished. Further improvements are
possible, but will be marginal, and will come at high cost. Ford
requests that the agencies carefully weigh these considerations in
the current and future rulemakings to ensure that resources and
investment are not diverted from our primary objective: Fulfilling
President Biden's goal of achieving 40-50 [percent] ZEV sales by
2030.\190\
---------------------------------------------------------------------------
\190\ Ford, Docket No. NHTSA-2021-0053-1545-A1, at p. 1.
---------------------------------------------------------------------------
Toyota stated:
Toyota has provided extensive information, in public comments
and under CBI, on the effectiveness of [CO2] reduction
technologies including those for advanced gasoline engines.\191\ The
data has consistently documented that even advanced ICE-only
powertrains will fall short of the proposed standards and that while
future advancements are possible, a point of diminishing returns is
in part driving the transition to electrified powertrains, including
conventional hybrids. EPA notes manufacturer plans and announcements
of ``a rapidly growing shift in investment away from internal-
combustion technologies and toward high levels of electrification.''
192 193
---------------------------------------------------------------------------
\191\ Toyota comments on: Draft Technical Assessment Report on
2022-2025 Model Year Light-Duty Vehicle Greenhouse Gas Emission
Standards and Corporate Average Fuel Economy Standards, EPA-420-D-
16-900 pp. 2-5 and Appendix 1; Proposed Determination on the
Appropriateness of the Model Year 2022-2025 Light-Duty Vehicle
Greenhouse Gas Emissions Standards under the Midterm Evaluation,
EPA-420-R-16-020, pp. 3-8; Request for Comment on Reconsideration of
the Final Determination of the Mid-Term Evaluation of Greenhouse Gas
Emissions Standards for Model Year 2022-2025 Light-Duty Vehicles;
Request for Comment on Model Year 2021 Greenhouse Gas Emissions
Standards, EPA-HQ-OAR-2015-0827, pp. 3-9; Safer Affordable Fuel-
Efficient (SAFE) Vehicles Rule For Model Years 2020-2026 Model Year
Passenger Cars and Light Trucks, NHTSA-2018-0067; EPA-HQ-OAR-2018-
0283, pp. 2-9 and Appendices A-C.
\192\ U.S. EPA. Revised 2023 and Later Model Year Light-Duty
Vehicle GHG Emissions Standards, EPA-HQ-OAR-2021-0208, August 2021,
at p. 43766.
\193\ Toyota, Docket No. NHTSA-2021-0053-1568, at p. 2.
---------------------------------------------------------------------------
Volkswagen stated:
As noted earlier, Volkswagen has implemented a capital spending
plan and technology roadmap that primary focuses on electrification
as our main pathway for achieving deep decarbonization and petroleum
reduction goals. In parallel with increasing consumer demand for
electrification, the increase in States with ZEV mandates and the
emergence and recent passage of State legislation banning
combustion, it is unlikely that OEMs will invest significant
resources in researching new combustion technologies or developing
all new powertrains.
Engine development programs are long-lead time, often requiring
5 years to fully design and validate new engines. Powertrain
production is also capital intensive, and the
[[Page 25784]]
high upfront costs often consider 10 plus years of steady volume to
amortize the production and development costs. The effects have been
studied extensively by NHTSA and the National Academies and are
reflected in such factors as Retail Price Equivalency (RPE) values.
However, with the shift to legislative and regulatory programs that
are reducing and eliminating future market volumes for combustion
technologies, it is unlikely that OEMs will make significant
investments in this space.
Volkswagen has publicly stated that investments into combustion
technologies will wane with a point in the next several years where
there will be no new combustion engine families developed for the
Group. Volkswagen recognizes that remaining combustion models will
continue to be sold in high volume for the next several years and
that it is important to preserve the fuel economy of remaining ICEs
as electrification volumes increase. As noted earlier, Volkswagen's
remaining ICE engines will primarily focus on evolutions of existing
downsized, charged engines to incorporate incremental hardware and
software improvements.\194\
---------------------------------------------------------------------------
\194\ Volkswagen, Docket No. NHTSA-2021-0053-1548-A1, at pp. 21-
22.
---------------------------------------------------------------------------
Auto Innovators stated:
Manufacturers are also already announcing plans to reduce or
eliminate investments in ICEs. Some automotive executives are saying
that they no longer intend to develop new ICEs, are no longer
setting aside significant money for new ICEs, or that ICEs will only
get incremental work. Others, such as policymakers, may suggest that
little or no investment is needed in ICE technologies because they
are ``off-the-shelf'' or present in the fleet today. This view
ignores that technologies can't simply be ``bolted on'' to existing
engines. Instead, they must be carefully integrated into existing
designs, requiring engineering resources, and in many cases, new
engine designs. A new engine design can cost as much as $1
billion.\195\
---------------------------------------------------------------------------
\195\ Auto Innovators, Docket No. NHTSA-2021-0053-0021-A1, at 8
(citing ``Mercedes-Benz Prepares to Go All-Electric,'' Mercedes-Benz
Media Newsroom USA (Jul. 22, 2021), https://media.mbusa.com/releases/release-ee5a810c1007117e79e1c871354679e4-mercedes-benz-prepares-to-go-all-electric (accessed: February 16, 2022).
``Investments into combustion engines and plug-in hybrid
technologies will drop by 80% between 2019 and 2026.''; Hannah Lutz,
``Shifting into E,'' Automotive News (Jul. 26, 2021). ``Some
existing vehicles, such as the Chevy Malibu and Camaro, won't stick
to the standard cadence of face-lifts and redesigns. Instead,
they'll ride out the current generation before making way for
EVs.''; Jordyn Grzelewski, ``Ford Slated to Spend More On EVs Than
On Internal Combustion Engine Vehicles in 2023,'' The Detroit News
(Aug. 2, 2021).; Lindsay Chappell, ``All-In On EVs,'' Automotive
News (May 17, 2021). ``Mini will become an all-electric brand by
early 2030, and the British marque will roll out its last new
combustion engine variant in 2025.'' (Emphasis added.); Bibhu
Pattnaik, ``Audi Will Not Introduce ICE Vehicles After 2026, No
Hybrid Vehicles Either,'' Benzinga (Jun. 19, 2021), https://finance.yahoo.com/news/audi-not-introduce-ice-vehicles-160320055.html (accessed: February 16, 2022), Mike Colias, ``Gas
Engines, and the People Behind Them, Are Cast Aside for Electric
Vehicles,'' The Wall Street Journal (Jul. 23, 2021). ``Auto
executives have concluded, to varying degrees, that they can't meet
tougher tailpipe-emission rules globally by continuing to improve
gas or diesel engines . . . Over the past several decades, auto
makers in most years rolled out between 20 and 70 new engines
globally, according to research firm IHS Markit. That number will
fall below 10 this year, and then essentially go to zero, the
research firm said.'').
These comments reflect an increasing industry trend to divest from
internal combustion engine technology, to increase investments in
alternative powertrains such as electrification or fuel cells. The
provided comments also support NAS's finding: ICE technology
advancements are seeing diminishing returns, with future gains
requiring significant investment, driving manufacturers to alternative
technology development in place of further ICE development, such as
electrification.\196\
---------------------------------------------------------------------------
\196\ 2021 NAS Report, Finding 4.7, at p. 70.
---------------------------------------------------------------------------
On the other hand, some commenters were concerned that our modeled
technology paths do not adequately keep pace with potential significant
improvements in ICE technologies that manufacturers will continue to
make. ICCT and UCS suggested that additional advanced versions of
modeled technologies as well as additional technologies should be added
to the engine technology paths. Both commenters provided information on
emerging technologies currently in the research phase, and the
commenters stated these new technologies should be included in the
engine technology path options.
ICCT stated, ``two recent reports demonstrate that further
technology improvements are coming that can boost ICE efficiency well
beyond even HCR2 efficiency levels.'' \197\ ICCT further stated,
``Indeed, it appears that no technology improvements or cost reductions
from EPA's independent evaluations or from any comments submitted to
NHTSA or new studies over the last 5 years were included in the
proposed rule, beyond the additional of DEAC to HCR1. This basis for
NHTSA's analysis is an overly conservative assessment of the costs of
the standards.''
---------------------------------------------------------------------------
\197\ ICCT, Docket No. NHTSA-2021-0053-1581-A1, at 2 (citing AVL
Webinar on Passenger Car powertrain 4.x--Fuel Consumption,
Emissions, and Cost on June 2, 2020 https://www.avl.com/-/passenger-car-powertrain-4.x-fuel-consumption-emissions-and-cost plus slides
are attached to these comments (AVL 2020); Roush report on Gasoline
Engine Technologies for Improved Efficiency (Roush 2021 LDV) https://www.regulations.gov/comment/EPA-HQ-OAR-2021-0208-0210).
---------------------------------------------------------------------------
UCS also provided a comment suggesting the need for more advanced
engine technology models:
Given automaker investments and future product plans, it is
likely that manufacturers' compliance strategies will include
increased electrification. However, there are significant
opportunities for improvements to internal combustion engine
vehicles as well. The importance of both strategies is evident in
our own modeling. Internal combustion engine vehicles will continue
to improve in the timeframe considered under this rule and show no
sign of exhausting their potential. While our modeling suggests that
manufacturers will deploy a significant number of EVs due to the
improvement they can make in a fleet's performance, this is by no
means the only path available, as indicated by the relatively low
levels of vehicle technology modeled as being deployed in the
remaining gasoline-powered fleet, which leave many other options
open.\198\
---------------------------------------------------------------------------
\198\ UCS, Docket No. NHTSA-2021-0053-1567-A1, at 6 (citing
Murphy, John. 2021. ``US Automotive Product Pipeline: Car Wars 2022-
2025 (Electric Vehicles shock the product pipeline).'' Media
briefing, June 10, 2021, on behalf of Bank of America Securities.
https://s3-prod.autonews.com/2021-06/BofA%20Global%20Research%20Car%20Wars.pdf).
For this final rule analysis, the agency has made no changes to the
Engine technology pathway.\199\ While we agree with the potential of
the technologies as they are described in the provided comments,\200\
we do not believe that the application of the technologies is feasible
in the rulemaking timeframe. As stated in the NPRM and discussed above,
we did not include technologies unlikely to be feasible in the
rulemaking timeframe, technologies unlikely to be compatible with U.S.
fuels, or technologies for which there were not appropriate data
available to allow the simulation of effectiveness across all vehicle
technology classes used in the analysis. For example, ICCT recommended
the inclusion of passive prechamber combustion in our analysis.
Currently, the technology is under development by two vendors, but
neither vendor has indicated the system has progressed past the
technology demonstration phase, or the technology is currently only
used for specialty purposes.201 202
---------------------------------------------------------------------------
\199\ See TSD Chapter 3.1 for a detailed discussion of the
engine technology pathways used in the final rule analysis.
\200\ ICCT comments at pp. 8-10.
\201\ https://www.iav.com/en/what-moves-us/pre-chamber-ignition-small-spark-great-effect/--Accessed 10DEC2021.
\202\ https://www.mahle-powertrain.com/en/experience/mahle-jet-ignition/--Accessed 10DEC2021.
---------------------------------------------------------------------------
In light of the comments provided by manufacturers, such as
Volkswagen's comment above, it is very unlikely that major
manufacturers will introduce these technologies in the time frame of
the regulation.203 204 We also believe this
[[Page 25785]]
approach is in agreement with the assessments on ICE technologies
provided by NAS, discussed above.\205\
---------------------------------------------------------------------------
\203\ Volkswagen, at 21-22 (``Engine development programs are
long-lead time, often requiring 5 years to fully design and validate
new engines. Powertrain production is also capital intensive and the
high upfront costs often consider 10 plus years of steady volume to
amortize the production and development costs.'').
\204\ Auto Innovators, at 8 (``Others, such as policymakers, may
suggest that little or no investment is needed in ICE technologies
because they are ``off-the-shelf'' or present in the fleet today.
This view ignores that technologies can't simply be ``bolted on'' to
existing engines. Instead, they must be carefully integrated into
existing designs, requiring engineering resources, and in many
cases, new engine designs. A new engine design can cost as much as
$1 billion.'').
\205\ 2021 NAS Report, at 369 (``Internal combustion engines
(ICEs) will continue to play a significant role in the new vehicle
fleet in MY 2025-2035 in ICE-only vehicles, as well as in hybrid
electric vehicles (HEVs) from mild hybrids to plug-in hybrids, but
will decrease in number with increasing battery electric vehicle
(BEV) and fuel cell electric vehicle penetration. In this period,
manufacturers will continue to develop and deploy technologies to
further improve the efficiency of conventional powertrains, for ICE-
only vehicles and as implemented in HEVs. Developments in the ICE
for hybrids will advance toward engines optimized for a limited
range of engine operating conditions, with associated efficiency
benefits. Major automakers are on differing paths, with some
focusing their research and development and advanced technology
deployment more squarely on BEVs, and others more focused on
advanced HEVs to maximize ICE efficiency.'').
---------------------------------------------------------------------------
(1) Basic Engines
We applied basic engine technologies individually or in combination
with other basic engine technologies in the CAFE Model. The basic
engine technologies we used include variable valve timing (VVT),
variable valve lift (VVL), stoichiometric gasoline direct injection
(SGDI), and cylinder deactivation. The cylinder deactivation
technologies we used includes a basic level (DEAC) and an advanced
level (ADEAC). DOT applies the basic engine technologies across two
engine architectures: Dual over-head camshaft (DOHC) engine
architecture and single over-head camshaft (SOHC) engine architecture.
VVT: Variable valve timing is a family of valve-train designs that
dynamically adjusts the timing of the intake valves, exhaust valves, or
both, in relation to piston position. VVT can reduce pumping losses,
provide increased engine torque and horsepower over a broad engine
operating range, and allow unique operating modes, such as Atkinson
cycle operation, to further enhance efficiency.\206\ VVT is nearly
universally used in the MY 2020 fleet. VVT enables more control of in-
cylinder air flow for exhaust scavenging and combustion relative to
fixed valve timing engines. Engine parameters such as volumetric
efficiency, effective compression ratio, and internal exhaust gas
recirculation (iEGR) can all be enabled and controlled by a VVT system.
---------------------------------------------------------------------------
\206\ 2015 NAS Report, at p. 31.
---------------------------------------------------------------------------
VVL: Variable valve lift dynamically adjusts the distance a valve
travels from the valve seat. The dynamic adjustment can optimize
airflow over a broad range of engine operating conditions. The
technology can increase effectiveness by reducing pumping losses and by
affecting the fuel and air mixture motion and combustion in-
cylinder.\207\ VVL is less common in the MY 2020 fleet than VVT, but
still prevalent. Some manufacturers have implemented a limited,
discrete approach to VVL. The discrete approach allows only limited
(e.g., two) valve lift profiles versus allowing a continuous range of
lift profiles.
---------------------------------------------------------------------------
\207\ 2015 NAS Report, at p. 32.
---------------------------------------------------------------------------
SGDI: Stoichiometric gasoline direct injection sprays fuel at high
pressure directly into the combustion chamber, which provides cooling
of the in-cylinder charge via in-cylinder fuel vaporization to improve
spark knock tolerance and enable an increase in compression ratio and/
or more optimal spark timing for improved efficiency.\208\ SGDI is
common in the MY 2020 fleet, and the technology is used in many
advanced engines as well.
---------------------------------------------------------------------------
\208\ 2015 NAS Report, at p. 34.
---------------------------------------------------------------------------
DEAC: Basic cylinder deactivation disables intake and exhaust
valves and turns off fuel injection for the deactivated cylinders
during light load operation. DEAC is characterized by a small number of
discrete operating configurations.\209\ The engine runs temporarily as
though it were a smaller engine, reducing pumping losses and improving
efficiency. DEAC is present in the MY 2020 baseline fleet.
---------------------------------------------------------------------------
\209\ 2015 NAS Report, at p. 33.
---------------------------------------------------------------------------
ADEAC: Advanced cylinder deactivation systems, also known as
rolling or dynamic cylinder deactivation systems, allow a further
degree of cylinder deactivation than the base DEAC. ADEAC allows the
engine to vary the percentage of cylinders deactivated and the sequence
in which cylinders are deactivated, essentially providing
``displacement on demand'' for low load operations. A small number of
vehicles have ADEAC in the MY 2020 baseline fleet.
Section III.D.1.d) contains additional information about each basic
engine technology used in this analysis, including information about
the engine map models used in the full vehicle technology effectiveness
modeling.
(2) Advanced Engines
We define advanced engine technologies in the analysis as
technologies that require significant changes in engine structure, or
an entirely new engine architecture.\210\ Currently there are two types
of advanced engine technologies, the application of alternate
combustion cycles or application of forced induction to the engine.
Each advanced engine technology has a discrete pathway for progression
to improved versions of the technology, as seen above in Figure III-7.
The advanced engine technology pathways include a turbocharged pathway,
a high compression ratio (Atkinson) engine pathway, a variable turbo
geometry (Miller Cycle) engine pathway, a variable compression ratio
pathway, and a diesel engine pathway. Although the CAFE Model includes
a compressed natural gas (CNG) pathway, that technology is a baseline-
only technology and was not included in the analysis; there are no
dedicated CNG vehicles in the MY 2020 analysis fleet.
---------------------------------------------------------------------------
\210\ Examples of this include but are not limited to changes in
cylinder count, block geometry or combustion cycle changes.
---------------------------------------------------------------------------
TURBO: Forced induction engines, or turbocharged downsized engines,
are characterized by technology that can create greater-than-
atmospheric pressure in the engine intake manifold when higher output
is needed. The raised pressure results in an increased amount of
airflow into the cylinder supporting combustion, increasing the
specific power of the engine. Increased specific power means the engine
can generate more power per unit of cylinder volume. The higher power
per cylinder volume allows the overall engine volume to be reduced,
while maintaining performance. The overall engine volume decrease
results in an increase in fuel efficiency by reducing parasitic loads
associated with larger engine volumes.\211\
---------------------------------------------------------------------------
\211\ 2015 NAS Report, at p. 34.
---------------------------------------------------------------------------
Cooled exhaust gas recirculation is also part of the advanced
forced induction technology path. The basic recycling of exhaust gases
using VVT is called internal EGR (iEGR) and is included as part of the
performance improvements provided by the VVT basic engine technology.
Cooled EGR (cEGR) is a second method for diluting the incoming air that
takes exhaust gases, passes them through a heat exchanger to reduce
their temperature, and then mixes them with incoming air in the intake
manifold.\212\ As discussed
[[Page 25786]]
in Section III.D.1.d), many advanced engine maps include EGR.
---------------------------------------------------------------------------
\212\ 2015 NAS Report, at p. 35.
---------------------------------------------------------------------------
Five levels of turbocharged engine downsizing technologies are
considered in this analysis: A `basic' level of turbocharged downsized
technology (TURBO1), an advanced turbocharged downsized technology
(TURBO2), an advanced turbocharged downsized technology with cooled
exhaust gas recirculation applied (cEGR), a turbocharged downsized
technology with basic cylinder deactivation applied (TURBOD), and a
turbocharged downsized technology with advanced cylinder deactivation
applied (TURBOAD).
HCR: Atkinson engines, or high compression ratio engines, represent
a class of engines that achieve a higher level of fuel efficiency by
implementing an alternate combustion cycle.\213\ Historically, the Otto
combustion cycle has been used by most gasoline-based spark ignition
engines. Increased research into improving fuel economy has resulted in
the application of alternate combustion cycles that allow for greater
levels of thermal efficiency. One such alternative combustion cycle is
the Atkinson cycle. Atkinson cycle operation is achieved by allowing
the expansion stroke of the engine to overextend, allowing the
combustion products to achieve the lowest possible pressure before the
exhaust stroke.214 215 216
---------------------------------------------------------------------------
\213\ See the 2015 NAS Report, Appendix D, for a short
discussion on thermodynamic engine cycles.
\214\ Otto cycle is a four-stroke cycle that has four piston
movements over two engine revolutions for each cycle. First stroke:
Intake or induction; seconds stroke: Compression; third stroke:
Expansion or power stroke; and finally, fourth stroke: Exhaust.
\215\ Compression ratio is the ratio of the maximum to minimum
volume in the cylinder of an internal combustion engine.
\216\ Expansion ratio is the ratio of maximum to minimum volume
in the cylinder of an IC engine when the valves are closed (i.e.,
the piston is traveling from top to bottom to produce work).
---------------------------------------------------------------------------
Descriptions of Atkinson cycle engines and Atkinson mode or
Atkinson-enabled engine technologies have been used interchangeably in
association with high compression ratio (HCR) engines, for past
rulemaking analyses. Both technologies achieve a higher thermal
efficiency than traditional Otto cycle-only engines, however, the two
engine types operate differently. For purposes of this analysis,
Atkinson technologies can be categorized into two groups to reduce
confusion: (1) Atkinson-enabled engines and (2) Atkinson engines.
Atkinson-enabled engines, or high compression ratio (HCR) engines,
dynamically swing between an Otto cycle like behavior (very little
expansion over-stroke) to a more Atkinson cycle intensive behavior
(large expansion over-stroke) based on engine demand. During high loads
the engine will reduce the Atkinson level behavior by increasing the
dynamic compression ratio, reducing over-stroke, sacrificing efficiency
for increased power density. While at low loads the engine will
increase the Atkinson level behavior by reducing the dynamic
compression ratio, increasing the over-stroke, improve efficiency but
reduce power density. The hybrid combustion cycle can be used to
address, but not eliminate, the low power density issues that can
constrain the application of an Atkinson-only engine and allow for a
wider application of the technology.
The level of efficiency improvement experienced by a vehicle
employing an Atkinson-enabled engine is directly related to how much of
the engine's operation time is spent at high Atkinson levels. Vehicles
that must maintain a high level of torque reserve, that experience
operation at a high load for long portions of their operating cycle, or
that have high base road loads, will see little to no benefit from this
technology compared with other advanced engine technologies. This power
density constraint results in manufacturers typically limiting the
application of this technology to vehicles with a lower road load, and
lower relative need for torque reserves.
Three HCR or Atkinson-enabled engines are available in the
analysis: (1) The baseline Atkinson-enabled engine (HCR0), (2) the
enhanced Atkinson enabled engine (HCR1), and finally, (3) the enhanced
Atkinson enabled engine with cylinder deactivation (HCR1D).
Next, Atkinson engines (as opposed to Atkinson-enabled engines,
discussed above) in this analysis are defined as engines that operate
full-time in Atkinson cycle. The most common method of achieving
Atkinson operation is the use of late intake valve closing. This method
allows backflow from the combustion chamber into the intake manifold,
reducing the dynamic compression ratio, and providing a higher over-
expansion ratio during the expansion stroke. The higher expansion ratio
improves thermal efficiency but reduces power density. The low power
density relegates these engines to hybrid vehicle (SHEVPS) applications
only in this analysis. Coupling the engines to electric motors and
significantly reducing road loads compensates for the lower power
density and maintains desired performance levels for the vehicle.\217\
The Toyota Prius is an example of a vehicle that uses an Atkinson
engine. The 2017 Toyota Prius achieved a peak thermal efficiency of 40
percent.\218\
---------------------------------------------------------------------------
\217\ Toyota. ``Under the Hood of the All-new Toyota Prius.''
Oct. 13, 2015. Available at https://global.toyota/en/detail/9827044.
(Accessed: February 17, 2022)
\218\ Matsuo, S., Ikeda, E., Ito, Y., and Nishiura, H., ``The
New Toyota Inline 4 Cylinder 1.8L ESTEC 2ZR-FXE Gasoline Engine for
Hybrid Car,'' SAE Technical Paper 2016-01-0684, 2016, https://doi.org/10.4271/2016-01-0684.
---------------------------------------------------------------------------
VTG: The Miller cycle is another type of overexpansion combustion
cycle, similar to the Atkinson cycle. The Miller cycle, however,
operates in combination with a forced induction system that helps
address the impacts of reduced power density during high load operating
conditions. Miller cycle-enabled engines use a similar technology
approach as seen in Atkinson-enabled engines to effectively create an
expanded expansion stroke of the combustion cycle.
In the analysis, the baseline Miller cycle-enabled engine includes
the application of a variable turbo geometry technology (VTG). The
advanced Miller cycle enabled system includes the application of a 48V-
based electronic boost system (VTGE). VTG technology allows the system
to vary boost level based on engine operational needs. The use of a
variable geometry turbocharger also supports the use of cooled exhaust
gas recirculation.\219\ An electronic boost system has an electric
motor added to assist a turbocharger at low engine speeds. The motor
assist mitigates turbocharger lag and low boost pressure at low engine
speeds. The electronic assist system can provide extra boost needed to
overcome the torque deficits at low engine speeds.\220\
---------------------------------------------------------------------------
\219\ 2015 NAS Report, at p. 116.
\220\ 2015 NAS Report, at p. 62.
---------------------------------------------------------------------------
ICCT provided comments regarding Miller Cycle technology as part of
its comments about technologies that may not have been incorporated in
NHTSA's proposal, stating that, ``VW is already using Miller Cycle
engines as the base engine in the Passat, Arteon, Atlas, and Tiguan and
a hybrid-specific version of this engine with cEGR and VGT is under
development by VW that demonstrates a peak BTE of 41.5 percent. The
fact that Miller cycle is already included on the standard engine for
many of VW's most popular vehicles supports that Miller cycle is a
cost-effective addition to turbocharged engines. Yet there are no
Miller cycle applications in 2026 beyond the specific Mazda and Volvo
models that already had Miller cycle in 2017.'' \221\
---------------------------------------------------------------------------
\221\ ICCT, at p. 4.
---------------------------------------------------------------------------
[[Page 25787]]
NHTSA's NPRM used a MY 2020 fleet that appropriately characterized
Volkswagen, Volvo, and Mazda engines with VTG and VTGe technology.\222\
We believe our use of the MY 2020 baseline fleet addresses some of the
concerns expressed by ICCT. As far as additional application of the
technology in the MY 2026 fleet results, we did not place any adoption
restrictions on the use of VTG and VTGe technology and it can be
applied to any basic and turbocharged engine. This means that while VTG
and VTGe may be a cost-effective technology for some manufacturers in
the real world--particularly for Volkswagen, a manufacturer that
already has the technology refined for use on its vehicles--the CAFE
Model did not consider it to be a cost-effective pathway to compliance
for manufacturers in the analysis, that did not already use the
technology in MY 2020. NHTSA does not have any alternative relative
effectiveness \223\ data or cost estimates to consider that would
affect the CAFE Model's compliance pathway. Therefore, we have made no
changes to this engine technology's inputs in the final rule analysis
from what was used in the NPRM. We will continue to follow any updates
on the effectiveness and cost of VTG and VTGe technology for future
actions.
---------------------------------------------------------------------------
\222\ See Section III.C.2, The Market Data File.
\223\ As a reminder, our analysis considers the relative
technology effectiveness improvement from a previously applied
technology. Therefore, while VW may be developing a hybrid version
of its Miller engine technology with a peak BTE of 41.5 percent, the
relevant data point for our analysis would be the relative
effectiveness improvement from the previous version of the
technology.
---------------------------------------------------------------------------
VCR: Variable compression ratio (VCR) engines work by changing the
length of the piston stroke of the engine to optimize the compression
ratio and improve thermal efficiency over the full range of engine
operating conditions. Engines using VCR technology are currently in
production, but appear to be targeted primarily towards limited
production, high performance applications. Nissan is the only
manufacturer to use this technology in the MY 2020 baseline fleet. Few
manufacturers and suppliers provided information about VCR
technologies, and we reviewed several design concepts that could
achieve a similar functional outcome. In addition to design concept
differences, intellectual property ownership complicates the ability to
define a VCR hardware system that could be widely adopted across the
industry. Because of these issues, adoption of the VCR engine
technology is limited to specific OEMs only.
ADSL: Diesel engines have several characteristics that result in
superior fuel efficiency over traditional gasoline engines. These
advantages include reduced pumping losses due to lack of (or greatly
reduced) throttling, high pressure direct injection of fuel, a more
efficient combustion cycle,\224\ and a very lean air/fuel mixture
relative to an equivalent-performance gasoline engine.\225\ However,
diesel technologies require additional enablers, such as a
NOX adsorption catalyst system or a urea/ammonia selective
catalytic reduction system, for control of NOX emissions.
---------------------------------------------------------------------------
\224\ Diesel cycle is also a four-stroke cycle like the Otto
Cycle, except in the intake stroke no fuel is injected and fuel is
injected late in the compression stroke at higher pressure and
temperature.
\225\ See the 2015 NAS Report, Appendix D, for a short
discussion on thermodynamic engine cycles.
---------------------------------------------------------------------------
DOT considered three levels of diesel engine technology: The
baseline diesel engine technology (ADSL) is based on a standard 2.2L
turbocharged diesel engine; the more advanced diesel engine (DSLI)
starts with the ADSL system and incorporates a combination of low
pressure and high pressure EGR, reduced parasitic loss, friction
reduction, a highly integrated exhaust catalyst with low temp light off
temperatures, and closed loop combustion control; and finally the most
advanced diesel system (DSLIAD) is the DSLI system with advanced
cylinder deactivation technology added.
EFR: Engine friction reduction technology is a general engine
improvement meant to represent future technologies that reduce the
internal friction of an engine. EFR technology is not available for
application until MY 2023. The future technologies do not significantly
change the function or operation of the engine but reduce the energy
loss due to the rotational or rubbing friction experienced in the
bearings or cylinder during normal operation. These technologies can
include improved surface coatings, lower-tension piston rings, roller
cam followers, optimal thermal management and piston surface
treatments, improved bearing design, reduced inertial loads, improved
materials, or improved geometry.
(b) Engine Analysis Fleet Assignments
As a first step in assigning baseline levels of engine technologies
in the analysis fleet, DOT uses data for each manufacturer to determine
which platforms share engines. Within each manufacturer's fleet, DOT
assigns unique identification designations (engine codes) based on
configuration, technologies applied, displacement, compression ratio,
and power output. DOT uses power output to distinguish between engines
that might have the same displacement and configuration but
significantly different horsepower ratings.
The CAFE Model identifies leaders and followers for a
manufacturer's vehicles that use the same engine, indicated by sharing
the same engine code. The model automatically determines which engines
are leaders by using the highest sales volume row of the highest sales
volume nameplate that is assigned an engine code. This leader-follower
relationship allows the CAFE Model simulation to maintain engine
sharing as more technology is applied to engines.
DOT accurately represents each engine using engine technologies and
engine technology classes. The first step is to assign engine
technologies to each engine code. Technology assignment is based on the
identified characteristics of the engine being modeled, and based on
technologies assigned, the engine will be aligned with a technology key
that most closely corresponds.
The engine technology classes are a second identifier used to
accurately account for engine costs. The engine technology class is
formatted as number of cylinders followed by the letter C, number of
banks followed by the letter B, and an engine head configuration
designator, which is _SOHC for single overhead cam, _ohv for overhead
valve, or blank for dual overhead cam. As an example, one variant of
the GMC Acadia has a naturally aspirated DOHC inline 4-cylinder engine,
so DOT assigned the vehicle to the `4C1B' engine technology class and
assigned the technology VVT and SGDI. Table III-7 shows examples of
observed engines with their corresponding assigned engine technologies
as well as engine technology classes.
[[Page 25788]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.065
The cost tables for a given engine class include downsizing (to an
engine architecture with fewer cylinders) when turbocharging technology
is applied, and therefore, the turbocharged engines observed in the
2020 fleet (that have already been downsized) often map to an engine
class with more cylinders. For instance, an observed TURBO1 V6 engine
would map to an 8C2B (V8) engine class, because the turbo costs on the
8C2B engine class worksheet assume a V6 (6C2B) engine architecture.
Diesel engines map to engine technology classes that match the observed
cylinder count since naturally aspirated diesel engines are not found
in new light duty vehicles in the U.S. market. Similarly, as indicated
above, the TURBO1 I3 in the Ford Escape maps to the 4C1B_L (I4) engine
class, because the turbo costs on the 4C1B_L engine class worksheet
assume a I3 (3C1B) engine architecture. Some instances can be more
complex, including low horsepower variants for 4 cylinder engines, and
are shown in Table III-8.
For this analysis, we allow additional downsizing beyond what has
been previously modeled in prior rulemaking analyses. We allow enhanced
downsizing because manufacturers have downsized low output naturally
aspirated engines to turbo engines with smaller architectures than
traditionally observed.226 227 228 To capture this new level
of turbo downsizing we created a new category of low output naturally
aspirated engines, which is only applied to 4-cylinder engines in the
MY 2020 fleet. These engines use the costing tabs in the Technologies
file with the `L' designation and are assumed to downsize to
turbocharged 3-cylinder engines for costing purposes. We sought comment
regarding the expected further application of this technology to larger
cylinder count engines, such as 8-cylinder engines that may be turbo
downsized to 4-cylinder engines. We also sought comment on how to
define the characteristic of an engine that may be targeted for
enhanced downsizing. We received no additional comments regarding
enhanced downsizing.
---------------------------------------------------------------------------
\226\ Richard Truett, ``GM Bringing 3-Cylinder back to North
America.'' Automotive News, December 01, 2019. https://www.autonews.com/cars-concepts/gm-bringing-3-cylinder-back-na.
(Accessed: February 17, 2022)
\227\ Stoklosa, Alexander, ``2021 Mini Cooper Hardtop.'' Car and
Driver, December 2, 2014. https://www.caranddriver.com/reviews/a15109143/2014-mini-cooper-hardtop-manual-test-review/. (Accessed:
February 17, 2022)
\228\ Leanse, Alex, ``2020 For Escape Options: Hybrid vs. 3-
Cylinder EcoBoost vs. 4-Cylinder EcoBoost.'' MotorTrend, Sept 24,
2019. https://www.motortrend.com/news/2020-ford-escape-engine-options-pros-and-cons-comparison/. (Accessed: February 17, 2022)
---------------------------------------------------------------------------
[[Page 25789]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.066
TSD Chapter 3.1.2 includes more details about baseline engine
technology assignment logic, and details about the levels of engine
technology penetration in the MY 2020 fleet.
(c) Engine Adoption Features
We defined engine adoption features through a combination of (1)
refresh and redesign cycles, (2) technology path logic, (3) phase-in
capacity limits, and (4) SKIP logic. Figure III-7 above shows the
technology paths available for engines in the CAFE Model. Engine
technology development and application typically results in an engine
design moving from the basic engine tree to one of the advanced engine
trees. Once an engine design moves to the advanced engine tree it is
not allowed to move to alternate advanced engine trees. Specific path
logic, phase-in caps, and SKIP logic applied to each engine technology
are discussed by engine technology, in turn.
Refresh and redesign cycles dictate when we apply engine
technology. Technologies applicable only during a platform redesign can
be applied during a platform refresh if another vehicle platform that
shares engine codes (uses the same engine) has already applied the
technology during a redesign. For example, models of the GMC Acadia and
the Cadillac XT4 use the same engine (assigned engine code 112011 in
the Market Data file); if the XT4 adds a new engine technology during a
redesign, then the Acadia may also add the same engine technology
during the next refresh or redesign. This allows the model to maintain
engine sharing relationships while also maintaining refresh and
redesign schedules.\229\ For engine technologies, DOHC, OHV, VVT, and
CNG engine technologies are baseline only, while all other engine
technologies can only be applied at a vehicle redesign.
---------------------------------------------------------------------------
\229\ See Section III.C.2.a) for more discussion on platform
refresh and redesign cycles.
---------------------------------------------------------------------------
Basic engine technologies in the CAFE Model are represented by four
technologies: VVT, VVL, SGDI, and DEAC. DOT assumes that 100 percent of
basic engine platforms use VVT as a baseline, based on wide
proliferation of the technology in the U.S. fleet. The remaining three
technologies, VVL, SGDI, and DEAC, can all be applied individually or
in any combination of the three. An engine can jump from the basic
engines path to any other engine path except the Alternative Fuel
Engine Path.
Turbo downsizing allows manufacturers to maintain vehicle
performance characteristics while reducing engine displacement and
cylinder count. Any basic engine can adopt one of the turbo engine
technologies (TURBO1, TURBO2, and CEGR1). Vehicles that have
turbocharged engines in the baseline fleet will stay on the turbo
engine path to prevent unrealistic engine technology change in the
short timeframe considered in the rulemaking analysis. Turbo technology
is a mutually exclusive technology in that it cannot be adopted for
HCR, diesel, ADEAC, or CNG engines.
Non-HEV Atkinson enabled engines are a collection of engines in the
HCR engine pathway (HCR0, HCR1, HCR1D, and HCR2). Atkinson enabled
engines excel in lower power applications for lower load conditions,
such as driving around a city or steady state highway driving without
large payloads. As a result, their adoption is more limited than some
other technologies. We expanded the availability of HCR technology
compared to the 2020 final rule because of new observed applications in
the market.\230\ However, there are three categories of adoption
features specific to the HCR engine pathway: \231\
---------------------------------------------------------------------------
\230\ For example, the Hyundai Palisade and Kia Telluride have a
291 hp V6 HCR1 engine. The specification sheets for these vehicles
are located in the docket for this action.
\231\ See Section III.D.1.d)(1) (Engine Maps), for a discussion
of why HCR2 and P2HCR2 were not used in the central analysis.
``SKIP'' logic was used to remove this engine technology from
application, however as discussed below, we maintain HCR2 and P2HCR2
in the model architecture for sensitivity analysis and for future
engine map model updates.
---------------------------------------------------------------------------
We currently do not allow vehicles with 405 or more
horsepower to adopt HCR engines due to their prescribed duty cycle
being more demanding and likely not supported by the lower power
density found in HCR-based engines.\232\
---------------------------------------------------------------------------
\232\ Heywood, John B. Internal Combustion Engine Fundamentals.
McGraw-Hill Education, 2018. Chapter 5.
---------------------------------------------------------------------------
Pickup trucks and vehicles that share engines with pickup
trucks are currently excluded from receiving HCR engines; the duty
cycle for these heavy vehicles, particularly the need for large torque
reserves, results in an engine calibration that minimizes the advantage
of Atkinson cycle use.\233\
---------------------------------------------------------------------------
\233\ This is based on CBI conversation with manufacturers that
currently employ HCR-based technology but saw no benefit when the
technology was applied to truck platforms in their fleet.
---------------------------------------------------------------------------
HCR engine application is also currently restricted for
some manufacturers that are heavily
[[Page 25790]]
performance-focused and have demonstrated a significant commitment to
power dense technologies such as turbocharged downsizing.\234\
---------------------------------------------------------------------------
\234\ There are three manufacturers that met the criteria (near
100 percent turbo downsized fleet, and future hybrid systems are
based on turbo-downsized engines) described and were excluded: BMW,
Daimler, and Jaguar Land Rover.
---------------------------------------------------------------------------
Advanced cylinder deactivation technology (ADEAC), or dynamic
cylinder deactivation (e.g., Dynamic Skip Fire), can be applied to any
engine with basic technology. This technology represents a naturally
aspirated engine with ADEAC. Additional technology can be applied to
these engines by moving to the Advanced Turbo Engine Path.
Miller cycle (VTG and VTGe) engines can be applied to any basic and
turbocharged engine. VTGe technology is enabled by the use of a 48V
system that presents an improvement from traditional turbocharged
engines, and accordingly VTGe includes the application of a mild hybrid
(BISG) system.
VCR engines can be applied to basic and turbocharged engines, but
the technology is limited to specific OEMs.\235\ VCR technology
requires a complete redesign of the engine, and in the analysis fleet,
only two platforms had incorporated this technology. The agency does
not believe any other manufacturers will invest to develop and market
this technology in their fleet in the rulemaking time frame.
---------------------------------------------------------------------------
\235\ Nissan and Mitsubishi are strategic partners and members
of the Renault-Nissan-Mitsubishi Alliance.
---------------------------------------------------------------------------
Advanced turbo engines are becoming more prevalent as the
technologies mature. TURBOD combines TURBO1 and DEAC technologies and
represents the first advanced turbo. TURBOAD combines TURBO1 and ADEAC
technologies and is the second and last level of advanced turbos.
Engines from either the Turbo Engine Path or the ADEAC Engine Path can
adopt these technologies.
Any basic engine technologies (VVT, VVL, SGDI, and DEAC) can adopt
ADSL and DSLI engine technologies. Any basic engine and diesel engine
can adopt DSLIAD technology in this analysis; however, we applied a
phase in cap and year for this technology at 34 percent and MY 2023,
respectively. In our engineering judgement, this is a rather complex
and costly technology to adopt and it would take significant investment
for a manufacturer to develop. For more than a decade, diesel engine
technologies have been used in less than one percent of the total
light-duty fleet production and have been found mostly on medium and
heavy-duty vehicles.
Finally, we allow the CAFE Model to apply EFR to any engine
technology except for DSLI and DSLIAD. DSLI and DSLIAD inherently have
incorporated engine friction technologies from ADSL. In addition,
friction reduction technologies that apply to gasoline engines cannot
necessarily be applied to diesel engines due to the higher temperature
and pressure operation in diesel engines.
We sought comment on the appropriateness of engine adoption
features, specifically for the HCR engines, and received feedback. Some
commenters felt the constraints on application of HCR technology in the
CAFE Model were too strict. Specifically, comments on this issue were
received from ICCT, California Air Resources Board (CARB), a coalition
of States and Cities, and a joint group of non-governmental
organizations.236 237 238 239 240 ICCT described NHTSA's
characterization of HCR with respect to the duty cycle requirements of
high horsepower or high towing vehicles as ``backwards and wrong,''
stating that:
---------------------------------------------------------------------------
\236\ ICCT, at p. 11.
\237\ CARB, Docket No. NHTSA-2021-0053-1521-A2, at pp. 6-8.
\238\ States of California, Colorado, Connecticut, Delaware,
Hawaii, Illinois, Maine, Maryland, Michigan, Minnesota, Nevada, New
Jersey, New Mexico, New York, North Carolina, Oregon, Rhode Island,
Vermont, Washington, and Wisconsin; the Commonwealths of
Massachusetts and Pennsylvania; the District of Columbia; the Cities
and Counties of Denver and San Francisco; and the Cities of Los
Angeles, New York, Oakland, and San Jose (NHTSA-2021-0053-1499)
(California Attorney General et al.), Docket No. NHTSA-2021-0053-
1499-A1, at p. 33.
\239\ Natural Resources Defense Council (NRDC), Docket No.
NHTSA-2021-0053-1572-A1, at p. 7.
\240\ NRDC, A2, at pp. 46-47.
engines in pickup trucks and high-performance vehicles are sized and
powered to handle higher peak loads and, thus, operate at lower
loads relative to their maximum capacity. According to supplemental
tables for the 2020 EPA FE Trends report found online, pickups have
18 [percent] to 19 [percent] higher power to weight than both cars
and truck SUVs, which means that pickup trucks and high-performance
vehicles will spend more time in Atkinson Cycle operation than lower
performance vehicles on both the test cycles and in the real world,
not less. Any need for ``additional torque reserve'' is met by
switching to Otto cycle. The one exception is towing, which does
impose constant high loads on the engine. However, Strategic Vision
data finds that ``percent of [pickup] truck owners use their truck
for towing one time a year or less''. The large majority of pickup
trucks spend the vast majority of driving at low loads relative to
the engine's capability, where Atkinson Cycle engines are very
effective. Thus, all restrictions on HCR engines should be
removed.\241\
---------------------------------------------------------------------------
\241\ ICCT, at p. 11.
We disagree with ICCT's and other comments regarding the
appropriateness of the HCR technology constraints. Current HCR engines
achieve the effects of a longer expansion stroke, necessary for
Atkinson operation, using continuous variable valve timing. The timing
of the intake valve closure is based on the current load demand on the
engine. Under higher loads, the intake values will close sooner in the
cycle, increasing the dynamic compression ratio and decreasing the
over-stroke of the expansion cycle, decreasing thermal efficiency, and
increasing torque. This causes the engine to operate closer to an Otto
combustion cycle than an Atkinson cycle. However, under these
conditions, the engine is not able to completely achieve a traditional
Otto cycle due to knock limitations and maintains a minimum of over-
expansion behavior. While under lower loads the engine decreases the
dynamic compression ratio, closing the intake valve later, and
increasing the over-stroke of the expansion stroke reducing torque
while increasing efficiency. Having the ability to continuously adjust
the shape of the combustion cycle significantly improves the engine
efficiency but does not give the engine the functional flexibility
suggested by ICCT's interpretation of the technology description.
This is exemplified by Toyota's comment to the 2018 CAFE NPRM on
the application of the HCR-based engine to the Tacoma platform, where
Toyota stated that:
Tacoma has a greater coefficient of drag from a larger frontal
area, greater tire rolling resistance from larger tires with a more
aggressive tread, and higher driveline losses from 4WD. Similarly,
the towing, payload, and off-road capability of pick-up trucks
necessitate greater emphasis on engine torque and horsepower over
fuel economy. This translates into engine specifications such as a
larger displacement and a higher stroke-to-bore ratio. Tacoma's
higher road load and more severe utility requirements push engine
operation more frequently to the less efficient regions of the
engine map and limit the level of Atkinson operation.\242\
---------------------------------------------------------------------------
\242\ Toyota, Docket No. NHTSA-2018-0067-12376-A1, at pp. 8-9.
In addition to operating issues, comments such as those provided by
the Auto Innovators, also to the 2018 NPRM (83 FR 42986, Aug. 24,
2018), highlight packaging issues that make the application of HCR in
high horsepower/high torque applications less practical. Specifically,
the Alliance of Automobile
[[Page 25791]]
Manufacturer's \243\ comments to the 2018 NPRM stated that ``[t]he
Alliance agrees with the more restrained application of HCR1 in the
Proposed Rule,'' and agreed with the agencies' rationale for the
restrictions that included ``[p]ackaging and emission constraints
associated with intricate exhaust manifolds needed to mitigate high
load/low revolutions per minute knock'' and ``Inherent performance
limitations of Atkinson cycle engines.'' \244\ Ford echoed this
concern, stating that ``Ford supports the more restrained application
of HCR1 in the Proposed Rule, an approach that recognizes the
investment, packaging, performance and emissions factors that will
limit penetration of this technology.'' \245\
---------------------------------------------------------------------------
\243\ Now Alliance for Automotive Innovation, also referred to
as Auto Innovators.
\244\ Auto Innovators, Docket No. NHTSA-2018-0067-12073-A1, at
p. 139.
\245\ Ford, Docket No. NHTSA-2018-0067-11928-A1, at p. 8.
---------------------------------------------------------------------------
Based on this discussion, and previously provided data, we have
kept the HCR adoptions features used in the NPRM for the final rule,
except for a correction to the HCR1D application. Keeping the
constraints in place also aligns us with the most recent EPA rulemaking
analysis.\246\ We do intend to continue research into the
appropriateness of HCR technology applications in future analysis, as
we look at timeframes beyond the current rulemaking.
---------------------------------------------------------------------------
\246\ See U.S. EPA, ``Revised 2023 and Later Model Year Light-
Duty Vehicle GHG Emissions Standards: Regulatory Impact Analysis.''
December 2021. EPA-420-R-21-028. https://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=P1013ORN.pdf. (Accessed: March 9, 2022)
---------------------------------------------------------------------------
Regarding the application of the HCR1D technology, a joint group of
NGO comments, and others, pointed out an error in the CAFE Model input
files used in the NPRM. The HCR1D technology was not set to `true' for
the central analysis.\247\ We agree the setting was left blank in error
and is correctly assigned a `true' value in the technology input file
for the final rule analysis.
---------------------------------------------------------------------------
\247\ NRDC, at pp. 46-47.
---------------------------------------------------------------------------
(d) Engine Effectiveness Modeling
Engine effectiveness values used for engine technologies in two
ways. The values are either calculated based on the difference in full
vehicle simulation results created using the Autonomie modeling tool,
or determined by the effectiveness values using an alternate
calculation method, including analogous improvement or fuel economy
improvement factors.
(1) Engine Maps
Effectiveness values used as inputs for the CAFE Model are
determined by comparing results of full vehicle simulations using the
Autonomie simulation tool. For a full discussion about how Autonomie
was used, see Section III.C.4 and TSD Chapter 2.4, in addition to the
Autonomie model documentation. Engine map models are the primary inputs
used to simulate the effects of different engine technologies in the
Autonomie full vehicle simulations.
Engine maps provide a three-dimensional representation of engine
performance characteristics at each engine speed and load point across
the operating range of the engine. Engine maps have the appearance of
topographical maps, typically with engine speed on the horizontal axis
and engine torque, power, or brake mean effective pressure (BMEP) \248\
on the vertical axis. A third engine characteristic, such as brake-
specific fuel consumption (BSFC),\249\ is displayed using contours
overlaid across the speed and load map. The contours provide the values
for the third characteristic in the regions of operation covered on the
map. Other characteristics typically overlaid on an engine map include
engine emissions, engine efficiency, and engine power. The engine maps
developed to model the behavior of the engines used in this analysis
are referred to as engine map models.
---------------------------------------------------------------------------
\248\ Brake mean effective pressure is an engineering measure,
independent of engine displacement, which indicates the actual work
an engine performs.
\249\ Brake-specific fuel consumption is the rate of fuel
consumption divided by the power being produced.
---------------------------------------------------------------------------
The engine map models used in this analysis are representative of
technologies that are currently in production or are expected to be
available in the rulemaking timeframe. The engine map models are
developed to be representative of the performance achievable across
industry for a given technology and are not intended to represent the
performance of a single manufacturer's specific engine. The broadly
representative performance level was targeted because the same
combination of technologies produced by different manufacturers will
have differences in performance, due to manufacturer-specific designs
for engine hardware, control software, and emissions calibration.
Accordingly, we expect that the engine maps developed for this
analysis will differ from engine maps for manufacturers' specific
engines. However, we intend and expect that the incremental changes in
performance modeled for this analysis, due to changes in technologies
or technology combinations, will be similar to the incremental changes
in performance observed in manufacturers' engines for the same changes
in technologies or technology combinations.
The analysis never applies absolute BSFC levels from the engine
maps to any vehicle model or configuration for the rulemaking analysis.
The absolute fuel economy values from the full vehicle Autonomie
simulations are used only to determine incremental effectiveness for
switching from one technology to another technology. The incremental
effectiveness is applied to the absolute fuel economy of vehicles in
the analysis fleet, which are based on CAFE compliance data. For
subsequent technology changes, incremental effectiveness is applied to
the absolute fuel economy level of the previous technology
configuration. Therefore, for a technically sound analysis, it is most
important that the differences in BSFC among the engine maps be
accurate, and not the absolute values of the individual engine maps.
For this analysis, we use a small number of baseline engine
configurations with well-defined BSFC maps, and then, in a very
systematic and controlled process, add specific well-defined
technologies to create a BSFC map for each unique technology
combination. This can theoretically be done using engine or vehicle
testing, but testing would need to be conducted on a single engine, and
each configuration would require physical parts and associated engine
calibrations to assess the impact of each technology configuration,
which is impractical for the rulemaking analysis because of the
extensive design, prototype part fabrication, development, and
laboratory resources that are required to evaluate each unique
configuration. Modeling is an approach used by industry to assess an
array of technologies with more limited testing. Modeling offers the
opportunity to isolate the effects of individual technologies by using
a single or small number of baseline engine configurations and
incrementally adding technologies to those baseline configurations.
This provides a consistent reference point for the BSFC maps for each
technology and for combinations of technologies that enables the
differences in effectiveness among technologies to be carefully
identified and quantified.
[[Page 25792]]
The Autonomie model documentation provides a detailed discussion on
how the engine map models were used as inputs to the full vehicle
simulations performed using the Autonomie tool. The Autonomie model
documentation contains the engine map model topographic figures, and
additional engine map model data can be found in the Autonomie input
files.\250\
---------------------------------------------------------------------------
\250\ See additional Autonomie supporting materials in docket
number NHTSA-2021-0053 for this rule.
---------------------------------------------------------------------------
We received a comment from the High Octane Low Carbon Fuel Alliance
regarding the potential use of high octane fuels. The High Octane Low
Carbon Fuel Alliance stated, ``Higher octane enables greater engine
efficiency and improved vehicle performance through higher compression
ratios and/or more aggressive turbocharging and downsizing--also
facilitated by ethanol's cylinder ``charge cooling'' effect due to its
high heat of vaporization.\251\ Raising the engine's compression ratio
from 10:1 to 12:1 could increase vehicle efficiency by 5 to 7
percent.'' 252 253
---------------------------------------------------------------------------
\251\ J.E. Anderson et al., ``High octane number ethanol-
gasoline blends: Quantifying the potential benefits in the United
States,'' Fuel (2012): 97: pp. 585-594: https://www.sciencedirect.com/science/article/pii/S0016236112002268.
(Accessed: February 17, 2022)
\252\ David S. Hirshfeld et al., ``Refining Economics of U.S.
Gasoline: Octane Ratings and Ethanol Content,'' Environmental
Science & Technology (2014): 48(19): pp. 11064-11071: https://pubs.acs.org/doi/pdf/10.1021/es5021668. (Accessed: February 17,
2022)
\253\ Thomas G. Leone et al., ``The Effect of Compression Ratio,
Fuel Octane Rating, and Ethanol Content on Spark- Ignition Engine
Efficiency,'' Environmental Science & Technology (2015): 49(18): pp.
10778-10789: https://pubs.acs.org/doi/abs/10.1021/acs.est.5b01420.
(Accessed: February 17, 2022)
---------------------------------------------------------------------------
We agree with the data provided; however, we simulate the use of
Tier 3 fuel in our engine technology models to represent the fuel
available and most commonly used by consumers.\254\ If we assumed that
high octane fuel was used in the engine map models, we would be
assuming a greater fuel economy benefit than would actually be achieved
in the real world, which would overestimate the benefits of more
stringent standards. Moreover, to date, vehicle manufacturers do not
appear to be pursuing this technology path. As we have stated
previously, regulation of fuels is also outside of the scope of NHTSA's
authority. Accordingly, we made no updates to the fuel assumed used in
the engine map models.
---------------------------------------------------------------------------
\254\ See TSD Chapter 3.1 for a detailed discussion on engine
map model assumptions.
---------------------------------------------------------------------------
(a) IAV Engine Map Models
Most of the engine map models used in this analysis were developed
by IAV GmbH (IAV) Engineering. IAV is one of the world's leading
automotive industry engineering service partners with an over 35-year
history of performing research and development for powertrain
components, electronics, and vehicle design.\255\ The primary outputs
of IAV's work for this analysis are engine maps that model the
operating characteristics of engines equipped with specific
technologies.
---------------------------------------------------------------------------
\255\ IAV Automotive Engineering, https://www.iav.com/en/.
(Accessed: February 17, 2022)
---------------------------------------------------------------------------
The generated engine maps are validated against IAV's global
database of benchmarked data, engine test data, single cylinder test
data, prior modeling studies, technical studies, and information
presented at conferences.\256\ The effectiveness values from the
simulation results are also validated against detailed engine maps
produced from Argonne engine benchmarking programs, as well as
published information from industry and academia, ensuring reasonable
representation of simulated engine technologies.\257\ The engine map
models used in this analysis and their specifications are shown in
Table III-9.
---------------------------------------------------------------------------
\256\ Friedrich, I., Pucher, H., and Offer, T., ``Automatic
Model Calibration for Engine-Process Simulation with Heat-Release
Prediction,'' SAE Technical Paper 2006-01-0655, 2006, https://doi.org/10.4271/2006-01-0655. (Accessed: February 17, 2022) Rezaei,
R., Eckert, P., Seebode, J., and Behnk, K., ``Zero-Dimensional
Modeling of Combustion and Heat Release Rate in DI Diesel Engines,''
SAE Int. J. Engines 5(3):874-885, 2012, https://doi.org/10.4271/2012-01-1065. (Accessed: February 17, 2022) Multistage Supercharging
for Downsizing with Reduced Compression Ratio (2015). MTZ Rene
Berndt, Rene Pohlke, Christopher Severin and Matthias Diezemann IAV
GmbH. Symbiosis of Energy Recovery and Downsizing (2014). September
2014 MTZ Publication Heiko Neukirchner, Torsten Semper, Daniel
Luederitz and Oliver Dingel IAV GmbH.
\257\ Bottcher, L., Grigoriadis, P. ``ANL--BSFC map prediction
Engines 22-26.'' IAV (April 30, 2019). https://lindseyresearch.com/wp-content/uploads/2021/09/NHTSA-2021-0053-0002-20190430_ANL_Eng-22-26-20190430_ANL_Eng22-26Updated_Docket.pdf. (Accessed: February 17,
2022)
---------------------------------------------------------------------------
BILLING CODE 4910-59-P
[[Page 25793]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.067
BILLING CODE 4910-59-C
We received a comment from ICCT regarding the validity of the
continued use of the IAV engine map models. ICCT stated that ``[t]he
engine maps that are
[[Page 25794]]
included in the agency modeling are severely outdated. For example, all
base naturally aspirated engine maps are based on an unidentified 2013
or older vehicle, all turbo (non-Miller cycle) maps are based on a
vehicle whose specifications match that of the 2011 MINI R56 N18/BMW
N13 engine, the hybrid Atkinson cycle map (for PS and PHEV) is based on
the 2010 Toyota Prius, and the HCR1 map is based on the 2014 Mazda
SkyActiv 2.0L engine. Essentially, NHTSA is assuming there will be no
efficiency improvements in any of these technologies through at least
2026, or for 12 to 16 years from the model year of the vehicle used to
generate the maps.'' \258\
---------------------------------------------------------------------------
\258\ ICCT, at p. 3.
---------------------------------------------------------------------------
We disagree with statements that the IAV engine maps are outdated.
Many of the engine maps were developed specifically to support analysis
for the current rulemaking time frame. The engine map models encompass
engine technologies that are present in the analysis fleet and
technologies that could be applied in the rulemaking timeframe. In many
cases those engine technologies are mainstream today and will continue
to be during the rulemaking timeframe. For example, the engines on some
MY 2020 vehicles in the analysis fleet have technologies that were
initially introduced ten or more years ago. Having engine maps
representative of those technologies is important for the analysis. The
most basic engine technology levels also provide a useful baseline for
the incremental improvements for other engine technologies. The
timeframe for the testing or modeling is unimportant because time by
itself doesn't impact engine map data. A given engine or model will
produce the same BSFC map regardless of when testing or modeling is
conducted. Simplistic discounting of engine maps based on temporal
considerations alone could result in discarding useful technical
information.
If we did use a mix of engine maps from engine modeling and from
benchmarking data, no common reference for measuring impacts of adding
specific technological improvements would exist. Additionally,
manufacturers often implement multiple fuel-saving technologies
simultaneously when redesigning a vehicle and it is not possible to
isolate the effect of individual technologies by using laboratory
measurements of a single production engine or vehicle with a
combination of technologies.\259\ Because so many vehicle and engine
changes are involved, it is not possible to attribute effectiveness
improvements accurately for benchmarked engines to specific technology
changes. Further, while two or more different manufacturers may produce
engines with the same high level technologies (such as a DOHC engine
with VVT and SGDI), each manufacturer's engine will have unique
component designs that cause its version of the engine to have a unique
engine map. For example, engines with the same high level technologies
have unique intake manifold and exhaust manifold runners, cylinder head
ports and combustion chamber geometry that impact charge motion,
combustion and efficiency, as well as unique valve control, compression
ratios, engine friction, cooling systems, and fuel injector spray
characteristics, among other factors. All of these differences lead to
potential overcounting or undercounting technology effectiveness per
cost. As described above, our approach allows the analysis to isolate
the effects of individual technologies by incrementally adding
individual technologies to baseline engine configurations. We selected
this approach for the NPRM and final rule and discuss it in detail in
the TSD.\260\
---------------------------------------------------------------------------
\259\ See e.g., Toyota Supplemental Comments to the 2018 NPRM,
Docket No. NHTSA-2018-0067-12431 (``Atkinson-cycle operation is just
one of several measures responsible for the 2.5L Dynamic Force
engine achieving a world-best 40 percent thermal efficiency. The
Late Intake Valve Closing (LIVC) of the Atkinson cycle reduces low-
load pumping losses and supports the 13:1 CR by suppressing engine
knock. However, the engine's increased stroke-to-bore ratio (S/B
ratio) and improved cooling, engine warmup, friction reduction, and
exhaust system play an equally important role. For example, the 1.18
S/B ratio preserves stable combustion under high EGR flow rates
which improves thermal efficiency as much as the longer effective
expansion ratio from the Atkinson cycle. The increased S/B ratio
also compliments intake port, valve timing (VVT-iE) and piston
enhancements resulting in greater tumble intensity of the charge-air
intake, higher speed combustion, and increased thermal efficiency.
Greater detail on factors contributing to the thermal efficiency of
the 2018 Camry 2.5L engine can be found in Toyota SAE paper 2017-01-
1021 contained in Appendix 1 of this submission.'').
\260\ See TSD Chapter 3.1.
---------------------------------------------------------------------------
As a result, it should not be expected that any of our engine maps
would necessarily align with a specific manufacturer's engine, unless
of course the engine map was developed from that specific engine. We do
not agree that comparing an engine map used for the rulemaking analysis
to a single specific benchmarked engine has technical relevance, beyond
serving as a general corroboration for the engine map. When a vehicle
is benchmarked, the resulting data are dictated by the unique
combination of technologies and design constraints for the whole
vehicle system.
ICCT further stated: ``As just two examples of how absurd it is to
assume no improvements in any of these engine technologies for at least
12 years, the turbocharged engine introduced by Honda in 2016 was
significantly more efficient than the engine used to generate all the
turbocharged maps in the proposed rule and the 2018 Camry hybrid
improved fuel economy by 15 (XLE/SE) to 25 percent (LE) compared to the
2017 Camry hybrid. And these (unincorporated) improvements were already
in the market by 2016 and 2018--still 8 to 10 years before 2026. For
additional information see UCS Reconsideration Petition pages 68-72.''
\261\ ICCT also stated ``EPA added a 2nd generation turbocharged
downsized engine package based on EPA benchmark testing of the Honda
L15B7 1.5L turbocharged, direct-injection engine to its 2018 MTE, which
was not used in NHTSA's proposed rule.'' \262\
---------------------------------------------------------------------------
\261\ ICCT, at p. 4.
\262\ Id.
---------------------------------------------------------------------------
Our effectiveness data, including engine map models, is not used in
the rulemaking analysis in the manner described in ICCT's comments. Our
analysis does not apply absolute BSFC levels from the engine maps to
any vehicle model or configuration for the rulemaking analysis. The
absolute fuel economy values from the full vehicle Autonomie
simulations are used only to determine incremental effectiveness for
switching from one technology to another technology. The incremental
effectiveness is applied to the absolute fuel economy of vehicles in
the analysis fleet, which are based on CAFE compliance data. For
subsequent technology changes, incremental effectiveness is applied to
the absolute fuel economy level of the previous technology
configuration. Therefore, for a technically sound analysis, it is most
important that the differences in BSFC among the engine maps be
accurate, and not the absolute values of the individual engine maps.
This comment also mirrors a similar ICCT comment to the 2018
NPRM.\263\ In the 2020 final rule, we compared two IAV engine maps to
the EPA's benchmarked Toyota 2017 2.5L naturally aspirated engine and
Honda's 2016 1.5L turbocharged downsized engine for predicted
effectiveness improvements. The IAV engines were modeled and simulated
in a midsize non-performance vehicle with an automatic transmission and
the same
[[Page 25795]]
road load technologies, MR0, ROLL0 and AERO0, to isolate for the
benefits associated with the specific engine maps.\264\ Eng 12, a 1.6L,
4-cylinder, turbocharged, SGDI, DOHC, dual cam VVT, VVL engine was
selected as the closest engine configuration to the Honda 1.5L.\265\
Eng 22b, a 2.5L, 4 cylinder, VVT Atkinson cycle engine, was selected as
the closest engine configuration to the Toyota 2.5L.\266\ Both the
Toyota 2.5L naturally aspirated engine and Honda's 1.5L engine have
incorporated a number of fuel saving technologies, including improved
accessories and engine friction reduction. To assure an ``apples-to-
apples'' comparison, both IACC and EFR technologies were applied to the
IAV engine maps. IACC technology provides an additional 3.6 percent
incremental improvement and EFR provides an additional 1.4 percent
incremental improvement beyond the IAV engine maps for midsize non-
performance vehicles.
---------------------------------------------------------------------------
\263\ ICCT, Attachment 3, Docket No. NHTSA-2018-0067-11741, at
p. I-49.
\264\ See TSD Chapter 3.4, TSD Chapter 3.5, and TSD Chapter 3.6
for more information on road load modeling.
\265\ See TSD Chapter 3.1 for more discussion on modeled engine
technologies.
\266\ See TSD Chapter 3.1 for more discussion on modeled engine
technologies.
---------------------------------------------------------------------------
The comparison shows that the relative effectiveness of the IAV
engine maps are in line with the Honda 1.5L and the Toyota 2.5L
benchmarked engines. Figure III-8 below shows the effectiveness
improvements for the EPA benchmarked engines and the corresponding IAV
engine maps incremental to a baseline vehicle. Accordingly, we believe
that the methodology used in this analysis, and the engine maps and
incremental effectiveness values used, are in line with benchmarking
data and are reasonable for the rulemaking analysis. We believe the
approach used in this rulemaking analysis appropriately allows us to
account for a wide array of engine technologies that could be adopted
during the rulemaking timeframe. Declining to use manufacturer-specific
engines allows us to ensure that all effectiveness and cost
improvements due to the incremental addition of fuel economy improving
technologies are appropriately accounted for.
[GRAPHIC] [TIFF OMITTED] TR02MY22.068
(b) Other Engine Map Models
Two of the engine map models we show in Table III-9, Eng24 and
Eng25, were not developed as part of the IAV modeling effort and we
only used Eng24 in this analysis. The Eng24 and Eng25 engine maps are
equivalent to the ATK and ATK2 engine map models developed for the 2016
Draft TAR, EPA Proposed Determination, and Final Determination.\267\
The ATK1 engine model is based directly on the 2.0L 2014 Mazda
SkyActiv-G (ATK) engine. The ATK2 represents an Atkinson engine concept
based on the Mazda engine, adding cEGR, cylinder deactivation, and an
increased compression ratio (14:1). In this analysis, Eng24 and Eng25
correspond to the HCR1 and HCR2 technologies.
---------------------------------------------------------------------------
\267\ 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.
---------------------------------------------------------------------------
We used the same HCR2 engine map model application in this analysis
as we used in the 2020 final rule.\268\ The agency believes the use of
HCR0, HCR1, and the new addition of HCR1D reasonably represents the
application of Atkinson Cycle engine technologies within the current
light-duty fleet and the anticipated applications of Atkinson Cycle
technology in the MY 2024-2026 timeframe. We sought comment on whether
and how to change our engine maps for HCR2 in the analysis for the
final rule.
---------------------------------------------------------------------------
\268\ 85 FR 24425-27 (April 30, 2020).
---------------------------------------------------------------------------
[[Page 25796]]
ICCT, among others supported the use of the HCR2 engine map model
stating that: 269 270 271 272
---------------------------------------------------------------------------
\269\ NRDC, at p. 47.
\270\ UCS, at p. 6.
\271\ CARB, at p. 4.
\272\ California Attorney General et al., A2, at p. 33.
Not only does EPA's proposed rule allow HCR2 technology to be
used in their modeling, but comments previously submitted and
previous EPA documentation provide extensive justification for HCR
technology benefits beyond just HCR1D. Also, both cooled EGR and
cylinder deactivation have been in production since 2018. Thus, it
is not credible to assume no further advances in HCR technology
prior to 2027. Further, the manufacturer claim of ``diminishing
returns to additional conventional engine technology improvements''
is also not credible, given the discussion in the Appendix Section 1
of extensive engine technologies under development that can reduce
GHG emissions by over 30 [percent]. ICCT certainly supports
developing an updated family of HCR engine map models that
incorporate many of the technologies discussed in Section 1 for
future rulemakings. But in the interim, HCR2 should be allowed in
the Final Rule using EPA's engine map for HCR2 developed in the
Technical Support Documents for EPA's Proposed and 2017 Final
Determination.\273\
---------------------------------------------------------------------------
\273\ ICCT, at p. 11.
Other commenters were opposed to the use of the HCR2 engine map
model in the analysis. Toyota provided comment on both the NHTSA and
---------------------------------------------------------------------------
EPA analysis, stating that:
HCR2 Atkinson engine technology has returned to EPA's compliance
modeling. EPA now defines HCR2 as ``the addition of dynamic cylinder
deactivation and cooled EGR within non-HEV Atkinson Cycle engine
applications''. However, the cost, technology effectiveness, and
underlying engine map used for modeling HCR2 technology appears
identical to that used for the SAFE 2 Final Rule which is
represented by the simulated and experimental effectiveness of the
2014 2.0L SKYACTIV engine with the addition of cooled Exhaust Gas
Recirculation (cEGR), 14:1 compression ratio (CR), and cylinder
deactivation. There is still no U.S. production vehicle that
incorporates this definition of HCR2 technology because the 14:1 CR
requires higher octane than currently available in U.S. regular
grade gasoline. Further, there are more cost-effective pathways than
combining cylinder deactivation with Atkinson cycle engines which
have inherently low pumping loss characteristics.
EPA compliance modeling applies HCR2 engine technology to over
40 percent of Toyota's fleet by 2026 model year. For example, Camry
receives HCR2 along with engine friction reduction (EFR) in 2024
model year. The resulting 51.7 mpg fuel economy is about a 9
[percent] improvement over Toyota's current generation Camry powered
by a 2.5L Atkinson engine which has a world-best 40 [percent]
thermal efficiency. The modeled [CO2] and fuel economy
are closer to hybrid Camry performance and are unreasonably large
for the technologies involved. First, cylinder deactivation is the
only practical distinction between HCR2 and Toyota's 2.5L Dynamic
Force Atkinson engine. NHTSA's evaluation has determined applying
only cylinder deactivation to Atkinson cycle engines (HCR1) nets an
incremental improvement of roughly 2 percent. Second, the 2.5L
Dynamic Force engine already encompasses EFR as explained in past
comments under CBI. Finally, IACC and EFR benefits appear to be
double counted on top of ERF already being included in the Camry
2.5L Atkinson engine. This is because IACC and EFR are both fully
included in the simulated HCR2 engine map, yet both technologies are
added again in the CAFE Model runs.
EPA modeling sequentially adds enhanced technology to a 2017
baseline fleet until compliance with the proposed standards is
achieved. The 2017 model year fleet is outdated because it fails to
capture more recent state-of-the-art technologies in the U.S. fleet
and requires the [CO2] reduction effectiveness of those
technologies to be assumed or simulated. An example is Toyota's 2.5L
Atkinson engine technology which has been in the market since 2018
model year. The Camry example above could largely be avoided using a
more recent baseline. A 2020 model year baseline fleet is more
appropriate and provides a more accurate performance assessment, and
with fewer product redesign cycles available, there is less chance
for technology effectiveness errors to propagate through the fleet.
The 2017 baseline has resulted in more Atkinson technology being
assumed in the 2018 through 2021 model year fleets than really
exists in the market.
Toyota further stated,
For compliance modeling of gasoline powertrains, EPA is
extensively relying on the HCR2 classification of Atkinson engine
technology for which the assumed efficacy remains unproven and
highly unlikely as previously explained. NHTSA effectively deploys
only to the HCR1 level of Atkinson engines which better reflects the
state of technology in the fleet today and identifies HCR1D as a
more advanced future pathway that while not cost-effective has a
considerably more reasonable assumed technology effectiveness than
HCR2.\274\
---------------------------------------------------------------------------
\274\ Toyota, at pp. 3-4.
The Auto Innovators also provided information and comment on the
---------------------------------------------------------------------------
HCR2 engine map model:
In the GHG NPRM [86 FR 43726, August 10, 2021], EPA resurrected
highly optimistic effectiveness estimates for future Atkinson cycle
engines based on a speculative engine map, and used the results as
``HCR2'' technology. The use of this technology package can diminish
the integrity of the analysis and distort discussions of
technological feasibility and economic practicability of future
standards. We recommend against the inclusion of this technology
package in the CAFE Model at this time.
While some organizations have asserted that EPA's 2016
characterization of HCR2 is a reasonable characterization of engines
in the market today, like Toyota's 2.5L on the Camry and RAV4, or
Mazda's 2.5L on the CX-5, history has shown that the HCR2
assumptions used in EPA's analysis significantly and unreasonably
overestimate the real-world fuel saving capability of state-of-the-
art Atkinson engine technology in these applications. The EPA HCR2
engine map assumes engine accessory drive improvements (``IACC'')
and engine friction reduction (``EFR'') have already been used to
the maximum extent possible, so reapplying these technologies again
in the modeling (as the EPA analysis does) incorrectly double counts
the potential effectiveness of these technologies. EPA incorrectly
states that HCR2 technology, as modeled, exists in the fleet and is
widely available for adoption.\275\
---------------------------------------------------------------------------
\275\ Auto Innovators, at pp. 49-51.
After review of the comments provided, we continue to believe HCR
engine technology shows promise for future ICE fuel economy
improvements and we continue with testing and validation for the IAV-
generated HCR engine map model family so that those engine map models
can be used in future analyses. However, we also believe that this
specific engine map model presents several problems when considered in
the context of this analysis. First, we believe that the technology
combination modeled by the HCR2 engine map is unlikely to be utilized
in the rulemaking timeframe based on comments received from the
industry leaders in HCR technology application. Second, as illustrated
by the Auto Innovators, this specific engine map model provides an
excessive jump in effectiveness when compared to the other IAV-based
engine map models used in this analysis. As a result, we have decided
to continue to exclude the HCR2 engine map model from our central
analysis. We will continue to expand the HCR engine map model family of
technologies in future analyses. This is consistent with EPA's current
assessment of their own model and choice to exclude the HCR2 engine in
their final rule analysis.\276\
---------------------------------------------------------------------------
\276\ See U.S. EPA, ``Revised 2023 and Later Model Year Light-
Duty Vehicle GHG Emissions Standards: Regulatory Impact Analysis.''
December 2021. EPA-420-R-21-028. https://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=P1013ORN.pdf. (Accessed: March 9, 2022)
---------------------------------------------------------------------------
(2) Analogous Engine Effectiveness Improvements and Fuel Economy
Improvement Values
For some technologies, the effectiveness for applying an
incremental engine technology is determined by using the effectiveness
values for applying the same engine technology to a reasonably similar
base
[[Page 25797]]
engine. An example of this can be seen in the determination of the
application of SGDI to the baseline SOHC engine. Currently there is no
engine map model for the SOHC+VVT+SGDI engine configuration. To create
the effectiveness data required as an input to the CAFE Model, first, a
pairwise comparison between technology configurations that included the
DOHC+VVT engine (Eng1) and the DOHC+VVT+SGDI (Eng18) engine was
conducted. Then, the results of that comparison were used to generate a
data set of emulated performance values for adding the SGDI technology
to the SOHC+VVT engine (Eng5b) systems.
The pairwise comparison is performed by finding the difference in
fuel consumption performance between every technology configuration
using the analogous base technology (e.g., Eng1) and every technology
configuration that only changes to the analogous technology (e.g.,
Eng18). The individual changes in performance between all the
technology configurations are then added to the same technology
configurations that use the new base technology (e.g., Eng5b) to create
a new set of performance values for the new technology (e.g.,
SOHC+VVT+SGDI). Table III-10 shows the engine technologies where
analogous effectiveness values were used.
[GRAPHIC] [TIFF OMITTED] TR02MY22.069
The agency received a comment about the use of analogous estimation
from ICCT. ICCT stated,
The modeled benefit of adding cylinder deactivation to
turbocharged and HCR1 vehicles is only about 25 [percent] of the
benefit from adding DEAC or ADEAC to a basic engine. While adding
DEAC to a turbocharged or HCR1 engine has smaller pumping loss
reductions than for base naturally aspirated engines, DEAC still has
significant pumping loss reductions and has the additional benefit
of enabling the engine to operate in a more thermal efficient region
of the engine fuel map. The agencies also failed to provide even the
most basic information supporting their effectiveness estimates for
TURBOD. Further compounding the problem, NHTSA based the
effectiveness of adding DEAC to HCR engines on the TURBOD estimate,
without any further justification.\277\
---------------------------------------------------------------------------
\277\ ICCT, at pp. 4-5.
We disagree with ICCT's characterization of the TURBOD engine map
model as ``not having information supporting its creation.'' A
discussion of the creation of the TURBOD engine map model, along with
all the engine map models, is provided in Chapter 3.1.3.1 of the TSD.
Furthermore, as discussed in Chapter 3.1.3.2.1 of the TSD, the HCR1D
effectiveness values are based on application of the DEAC technology to
a similar technology model (TURBO1) where there is a reduced pumping
loss benefit. Additionally, commenters did not indicate what
effectiveness values they would consider reasonable or plausible, and
NHTSA has no new data to support the ICCT position. As a result, we
will continue to use the effectiveness values from the NPRM for the
final rule analysis.
We also developed a static fuel efficiency improvement factor to
simulate applying an engine technology for some technologies where
there is either, no appropriate analogous technology, or there are not
enough data to create a full engine map model. The improvement factors
are developed based on a literature review or confidential business
information (CBI) provided by stakeholders. Table III-11 provides a
summary of the technology effectiveness values simulated using
improvement factors, and the value and rules for how the improvement
factors are applied. Advanced cylinder deactivation (ADEAC, TURBOAD,
DSLIAD), advanced diesel engines (DSLIA) and engine friction reduction
(EFR) are the three technologies modeled using improvement factors.
The application of the advanced cylinder deactivation is
responsible for three of the five technologies using an improvement
factor in this analysis. The initial review of the advanced cylinder
deactivation technology is based on a technical publication that used a
MY 2010 SOHC VVT basic engine.\278\ Additional information about the
technology effectiveness came from a benchmarking analysis of pre-
production 8-cylinder OHV prototype systems.\279\ However, at the time
of the
[[Page 25798]]
analysis no studies of production versions of the technology are
available, and the only available technology effectiveness came from
existing studies, not operational information. Thus, only estimates of
effect can be developed and not a full model of operation. No engine
map model can be developed, and no other technology pairs are
analogous.
---------------------------------------------------------------------------
\278\ Wilcutts, M., Switkes, J., Shost, M., and Tripathi, A.,
``Design and Benefits of Dynamic Skip Fire Strategies for Cylinder
Deactivated Engines,'' SAE Int. J. Engines 6(1):278-288, 2013,
available at https://doi.org/10.4271/2013-01-0359 (Accessed:
February 17, 2022); Eisazadeh-Far, K. and Younkins, M., ``Fuel
Economy Gains through Dynamic-Skip-Fire in Spark Ignition Engines,''
SAE Technical Paper 2016-01-0672, 2016, available at https://doi.org/10.4271/2016-01-0672. (Accessed: February 17, 2022).
\279\ EPA, 2018. ``Benchmarking and Characterization of a Full
Continuous Cylinder Deactivation System.'' Presented at the SAE
World Congress, April 10-12, 2018. Retrieved from https://www.regulations.gov/document?D=EPA-HQ-OAR-2018-0283-0029. (Accessed:
February 17, 2022).
---------------------------------------------------------------------------
To model the effects of advanced cylinder deactivation, an
improvement factor is determined based on the information referenced
above and applied across the engine technologies. The effectiveness
values for naturally aspirated engines are predicted by using full
vehicle simulations of a basic engine with DEAC, SGDI, VVL, and VVT,
and adding 3 percent or 6 percent improvement based on engine cylinder
count: 3 percent for engines with 4 cylinders or less and 6 percent for
all other engines. Effectiveness values for turbocharged engines are
predicted using full vehicle simulations of the TURBOD engine and
adding 1.5 percent or 3 percent improvement based on engine cylinder
count: 1.5 percent for engines with 4 cylinders or less and 3 percent
for all other engines. For diesel engines, effectiveness values are
predicted by using the DSLI effectiveness values and adding 4.5 percent
or 7.5 percent improvement based on vehicle technology class: 4.5
percent improvement is applied to small and medium non-performance
cars, small performance cars, and small non-performance SUVs. 7.5
percent improvement is applied to all other vehicle technology classes.
The analysis models advanced engine technology application to the
baseline diesel engine by applying an improvement factor to the ADSL
engine technology combinations. A 12.8 percent improvement factor is
applied to the ADSL technology combinations to create the DSLI
technology combinations. The improvement in performance is based on the
application of a combination of low pressure and high pressure EGR,
reduced parasitic loss, advanced friction reduction, incorporation of
highly integrated exhaust catalyst with low temp light off
temperatures, and closed loop combustion
control.280 281 282 283
---------------------------------------------------------------------------
\280\ 2015 NAS Report, at p. 104.
\281\ Hatano, J., Fukushima, H., Sasaki, Y., Nishimori, K.,
Tabuchi, T., Ishihara, Y. ``The New 1.6L 2-Stage Turbo Diesel Engine
for HONDA CR-V.'' 24th Aachen Colloquium--Automobile and Engine
Technology 2015.
\282\ Steinparzer, F., Nefischer, P., Hiemesch, D., Kaufmann,
M., Steinmayr, T. ``The New Six-Cylinder Diesel Engines from the BMW
In-Line Engine Module.'' 24th Aachen Colloquium--Automobile and
Engine Technology 2015.
\283\ Eder, T., Weller, R., Spengel, C., B[ouml]hm, J., Herwig,
H., Sass, H. Tiessen, J., Knauel, P. ``Launch of the New Engine
Family at Mercedes-Benz.'' 24th Aachen Colloquium--Automobile and
Engine Technology 2015.
---------------------------------------------------------------------------
As discussed above, the application of the EFR technology does not
simulate the application of a specific technology, but the application
of an array of potential improvements to an engine. All reciprocating
and rotating components in the engine are potential candidates for
friction reduction, and small improvements in several components can
add up to a measurable fuel economy
improvement.284 285 286 287 Because of the incremental
nature of this analysis, a range of 1-2 percent improvement was
identified initially, and narrowed further to a specific 1.39 percent
improvement. The final value is likely representative of a typical
value industry may be able to achieve in future years.
---------------------------------------------------------------------------
\284\ ``Polyalkylene Glycol (PAG) Based Lubricant for Light- &
Medium-Duty Axles,'' 2017 DOE Annual Merit Review. Ford Motor
Company, Gangopadhyay, A., Ved, C., Jost, N. https://energy.gov/sites/prod/files/2017/06/f34/ft023_gangopadhyay_2017_o.pdf.
\285\ ``Power-Cylinder Friction Reduction through Coatings,
Surface Finish, and Design,'' 2017 DOE Annual Merit Review. Ford
Motor Company. Gangopadhay, A. Erdemir, A. https://energy.gov/sites/prod/files/2017/06/f34/ft050_gangopadhyay_2017_o.pdf. (Accessed:
February 17, 2022).
\286\ ``Nissan licenses energy-efficient engine technology to
HELLER,'' https://newsroom.nissan-global.com/releases/170914-01-e?lang=en-US&rss&la=1&downloadUrl=%2Freleases%2F170914-01-e%2Fdownload (accessed: February 17, 2022).
\287\ ``Infiniti's Brilliantly Downsized V-6 Turbo Shines,''
https://wardsauto.com/engines/infiniti-s-brilliantly-downsized-v-6-turbo-shines (accessed: February 17, 2022).
[GRAPHIC] [TIFF OMITTED] TR02MY22.070
(3) Engine Effectiveness Values
The effectiveness values for the engine technologies, for all ten
vehicle technology classes, are shown in Figure III-8. Each of the
effectiveness values shown are representative of the improvements seen
for upgrading only the listed engine technology for a given combination
of other technologies. In other words, the range of effectiveness
values seen for each specific technology (e.g., TURBO1) represents the
addition of the TURBO1 technology to every technology combination that
could select the addition of TURBO1. See Table III-12 for several
specific examples. It must be emphasized, the change in fuel
consumption values between entire technology keys are
[[Page 25799]]
used,\288\ and not the individual technology effectiveness values.
Using the change between whole technology keys captures the
complementary or non-complementary interactions among technologies.
---------------------------------------------------------------------------
\288\ Technology key is the unique collection of technologies
that constitutes a specific vehicle, see Section III.C.4.c).
[GRAPHIC] [TIFF OMITTED] TR02MY22.071
Some of the advanced \289\ engine technologies have values that
indicate seemingly low effectiveness. Investigation of these values
shows the low effectiveness is a result of applying the advanced
engines to existing SHEVP2 architectures. This effect is expected and
illustrates the importance of using the full vehicle modeling to
capture interactions between technologies and capture instances of both
complimentary technologies and non-complimentary technologies. In this
instance, the SHEVP2 powertrain improves fuel economy, in part, by
allowing the engine to spend more time operating at efficient engine
speed and load conditions. This reduces the advantage of adding
advanced engine technologies, which also improve fuel economy, by
broadening the range of speed and load conditions for the engine to
operate at high efficiency. This redundancy in fuel savings mechanism
results in a lower effectiveness when the technologies are added to
each other.
---------------------------------------------------------------------------
\289\ The full data set we used to generate this example can be
found in the FE_1 Improvements file.
---------------------------------------------------------------------------
[[Page 25800]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.072
(e) Engine \290\ Costs
---------------------------------------------------------------------------
\290\ The box shows the inner quartile range (IQR) of the
effectiveness values and whiskers extend out 1.5 x IQR. The dots
outside this range show effectiveness values outside those
thresholds. The data used to create this figure can be found in the
FE_1 Improvements file.
---------------------------------------------------------------------------
We consider both cost and effectiveness in the CAFE Model when
selecting any technology changes. As discussed in detail in TSD Chapter
3.1.8, the engine costs we use in this analysis build on estimates from
the 2015 NAS Report, from agency-funded teardown studies, and from work
performed by non-government organizations.\291\
---------------------------------------------------------------------------
\291\ FEV prepared several cost analysis studies for EPA on
subjects ranging from advanced 8-speed transmissions to belt
alternator starters or start/stop systems. NHTSA contracted
Electricore, EDAG, and Southwest Research for teardown studies
evaluating mass reduction and transmissions. The 2015 NAS Report
also evaluated technology costs developed based on these teardown
studies.
---------------------------------------------------------------------------
We use the absolute costs of the engine technology in this
analysis, instead of relative costs used prior to the 2020 final rule.
We use absolute costs to ensure the full cost of the IC engine is
removed when electrification technologies are applied, specifically for
transition to BEVs. In this analysis, we model the cost of adopting BEV
technology by first removing the costs associated with IC powertrain
systems, then applying the BEV systems costs. Relative costs can still
be determined through comparison of the absolute costs for the initial
technology combination and the new technology combination.
As discussed in detail in TSD Chapter 3.1.8, we assigned engine
costs based on the number of cylinders in the engine and whether the
engine is naturally aspirated or turbocharged and downsized. Table III-
13 below shows an example of absolute costs for engine technologies in
2018$. The example costs are shown for a straight 4-cylinder DOHC
engine and V-6-cylinder DOHC engine. The table shows costs declining
across successive years due to the learning rate we applied to each
engine technology. For a full list of all absolute engine costs we used
in the analysis across all model years, see the Technologies file.
[[Page 25801]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.073
We received several comments regarding engine technology costs.
ICCT provided several cost comments for technologies including direct
injection, cool exhaust gas recirculation, cylinder deactivation and
turbo charging, that all took issue with the agency for not using cost
data from a 2015 FEV teardown study.\292\
---------------------------------------------------------------------------
\292\ FEV 2015--David Blanco-Rodriguez, 2025 Passenger car and
light commercial vehicle powertrain technology analysis. FEV GmbH.
September 2015. https://theicct.org/sites/default/files/publications/PV-LCV-Powertrain-Tech-Analysis_FEV-ICCT_2015.pdf.
(Accessed: February 16, 2022).
---------------------------------------------------------------------------
As we explained in the 2020 final rule, we do not believe that the
FEV report referenced by ICCT is an appropriate source to use for this
analysis for a few reasons. First, the primary focus of the FEV study
``is the European Market according to the EU6b regulation as well as
the consideration of emissions under both the NEDC and WLTP test
procedures.'' Components designed for use in Europe will have alternate
constraints from parts designed for use in the U.S., such as octane
limits, which can result in different designs and costs. This final
rule analysis specifically considered the U.S. automotive market during
the rulemaking timeframe based on U.S.-specific regulatory test cycles.
Accordingly, the costs reflect incremental technology effectiveness for
achieving improvements as measured through U.S. regulatory test
methods. We discuss these test cycles and methods further in Section
III.C.4.
Second, FEV did not conduct original teardown studies for this
report, as indicated by project tasks, but rather used engineering
judgement and external studies in assessing incremental costs.\293\ The
FEV report did not provide sources for each individual cost and it is
unclear how costs in many scenarios were developed since no teardowns
were used. Note that for this final rule analysis, we used previously
conducted FEV cost teardown studies and the referenced 2015 NAS costs
that also references FEV teardowns. As a result of this assessment we
are not concluding that FEV as a whole is a source on which NHTSA
should not rely, but we do want to make sure the baseline assumptions
of costing data, and how they are collected, are consistent with the
baseline assumptions of our analysis.
---------------------------------------------------------------------------
\293\ FEV EU Costs Tasks: ``Definition of reference hardware or
description made by experience of development and design engineers
as well as additional research as base for cost analysis (no
purchase of hardware).''
---------------------------------------------------------------------------
Finally, the cost for different vehicle classes identified by the
FEV study does not line up with the vehicle classes discussed in the
NPRM and this final rule analysis. FEV stated specifically, ``the
configuration of the vehicles has not been optimized for the [U.S.]
market and may not be representative of this market.'' \294\ We have
discussed the importance of aligning the CAFE vehicle models with the
U.S. market earlier in
[[Page 25802]]
Sections III.C.2 and III.C.4. All of these factors make it difficult to
compare directly our estimates and estimates presented in the FEV
report cited by ICCT in their comments.
---------------------------------------------------------------------------
\294\ Id. at p. 141.
---------------------------------------------------------------------------
ICCT's comment regarding the cost of the HCR engine technology
costs, unlike the costs discussed above, did not originate with the
2015 FEV report. ICCT stated that ``DMC costs for HCR in the SAFE rule,
which are unchanged in NHTSA's proposed rule, were about $200 more than
in EPA's 2016 TAR. This is a clear case where the agencies appear to
have not used the best available data from EPA.''
We used the same DMCs established by the 2015 NAS Report for the
Atkinson cycle technologies in both the NPRM analysis and the final
rule analysis. However, because there are many various engine
configurations in the market, we do not use the same fixed costs that
were set for each type of vehicle described in the 2015 NAS Report,
such as pickup and sedan. We have expanded costs by considering the
type of technology in the baseline, like SGDI, and the configuration of
the engine, such as SOHC versus DOHC. In addition, the cost used in the
NPRM also included updated dollar year, learning rate, and RPE in
comparison to the 2016 TAR. Although EPA also used costs from the 2015
NAS Report for the Proposed Determination analysis, they used a
different approach to account for components.
After review of the provided comments, we continue to rely on the
costs developed from the data provided by NAS and used for the NPRM
analysis. Engine technology costs often exist as a range of values
across manufacturers, and we work to try and find the best
representative value of that range, avoiding either maximum or minimum
values.
Transmission Paths
For this analysis, we classify all light duty vehicle transmission
technologies into discrete transmission technology paths. We use these
paths to model the most representative characteristics, costs, and
performance of the fuel-economy improving transmissions most likely
available during the rulemaking time frame, MYs 2024-2026.
In the following sections we discuss how we define transmission
technologies in this analysis, the general technology categories we use
in the CAFE Model, and the transmission technologies' relative
effectiveness and costs. In the following sections we also provide an
overview of how we assign transmission technologies to the baseline
fleet, as well as the adoption features, we apply to the transmission
technologies.
We only received comments regarding the costs assigned to eCVT
technology for power-split strong hybrid (i.e., SHEVPS) systems. Our
model only uses the eCVT technology as part of the SHEVPS technology
package, and the eCVT is not modeled as a standalone transmission
technology. As a result, we have responded to comments on eCVT costs in
Section III.D.3. For all other transmission technologies, we use the
same NPRM transmission technologies inputs and costs for the final rule
analysis.
(a) Transmission Modeling in the CAFE Model
We model two categories of transmissions for this analysis:
Automatic and manual. We characterize automatic transmissions as
transmissions that automatically select and shift between transmission
gears for the driver during vehicle operation. We further subdivide
automatic transmissions into four subcategories: Traditional automatic
transmissions (AT), dual clutch transmissions (DCT), continuously
variable transmissions (CVT), and direct drive transmissions (DD).
We model both the DD transmission and eCVT as part of electrified
powertrain technology packages, and not as independently selectable
technologies. As a result, we do not explicitly include either
technology in the transmission paths, and the technologies are
discussed further in Section III.D.3.
We employ different levels of high efficiency gearbox (HEG)
technology in the ATs and CVTs. HEG improvements for transmissions
represent incremental advancement in technology that improve
efficiency, such as reduced friction seals, bearings and clutches,
super finishing of gearbox parts, and improved lubrication. These
advancements are aimed at reducing frictional and other parasitic loads
in transmissions, to improve efficiency. We consider three levels of
HEG improvements in this analysis, based on 2015 NAS Report and CBI
data.\295\ We apply HEG efficiency improvements to ATs and CVTs,
because those transmissions inherently have higher friction and
parasitic loads related to hydraulic control systems and greater
component complexity, compared to MTs and DCTs. We note HEG technology
improvements in the transmission technology pathways by increasing
``levels'' of a transmission technology; for example, the baseline 8-
speed automatic transmission is termed ``AT8'', while an AT8 with level
2 HEG technology is ``AT8L2'' and an AT8 with level 3 HEG technology is
``AT8L3.''
---------------------------------------------------------------------------
\295\ 2015 NAS Report, at p. 191.
---------------------------------------------------------------------------
AT: Conventional planetary gear automatic transmissions are the
most popular transmission.\296\ ATs typically contain three or four
planetary gear sets that provide the various gear ratios. Gear ratios
are selected by activating solenoids which engage or release multiple
clutches and brakes as needed. ATs are packaged with torque converters,
which provide a fluid coupling between the engine and the driveline and
provide a significant increase in launch torque. When transmitting
torque through this fluid coupling, energy is lost due to the churning
fluid. These losses can be eliminated by engaging the torque convertor
clutch to directly connect the engine and transmission (``lockup'').
For the Draft TAR and 2020 final rule, EPA and DOT surveyed automatic
transmissions in the market to assess trends in gear count and
purported fuel economy improvements.\297\ Based on that survey, and
also EPA's 2021 Automotive Trends Report,\298\ we concluded that
modeling ATs with a range of 5 to 10 gears, with three levels of HEG
technology for this analysis was reasonable.
---------------------------------------------------------------------------
\296\ 2021 EPA Automotive Trends Report, at pp. 62-66.
\297\ Draft TAR at 5-50, 5-51; Final Regulatory Impact Analysis
accompanying the 2020 final rule, at 549.
\298\ 2021 EPA Automotive Trends Report, at pp. 62-66.
---------------------------------------------------------------------------
CVT: Conventional continuously variable transmissions consist of
two cone-shaped pulleys, connected with a belt or chain. Moving the
pulley halves allows the belt to ride inward or outward radially on
each pulley, effectively changing the speed ratio between the pulleys.
This ratio change is smooth and continuous, unlike the step changes of
other transmission varieties.\299\ We include two types of CVT systems
in the selectable transmission paths, the baseline CVT and a CVT with
HEG technology applied.
---------------------------------------------------------------------------
\299\ 2015 NAS Report, at p. 171.
---------------------------------------------------------------------------
DCT: Dual clutch transmissions, like automatic transmissions,
automate shift and launch functions. DCTs use separate clutches for
even-numbered and odd-numbered gears, allowing the next gear needed to
be pre-selected, resulting in faster shifting. The use of multiple
clutches in place of a torque converter results in lower parasitic
[[Page 25803]]
losses than ATs.\300\ Because of a history of limited
appeal,301 302 we constrain application of additional DCT
technology to vehicles already using DCT technology, and only model two
types of DCTs in this analysis.
---------------------------------------------------------------------------
\300\ 2015 NAS Report, at p. 170.
\301\ 2020 EPA Automotive Trends Report, at p. 57.
\302\ 2021 NAS Report, at 56.
---------------------------------------------------------------------------
MT: Manual transmissions are transmissions that require direct
control by the driver to operate the clutch and shift between gears. In
a manual transmission, gear pairs along an output shaft and parallel
layshaft are always engaged. Gears are selected via a shift lever,
operated by the driver. The lever operates synchronizers, which speed
match the output shaft and the selected gear before engaging the gear
with the shaft. During shifting operations (and during idle), a clutch
between the engine and transmission is disengaged to decouple engine
output from the transmission. Automakers today offer a minimal
selection of new vehicles with manual transmissions.\303\ As a result
of reduced market presence, we only include three variants of manual
transmissions in the analysis.
---------------------------------------------------------------------------
\303\ 2020 EPA Automotive Trends Report, at p. 61.
---------------------------------------------------------------------------
The transmission model paths used in this analysis are shown in
Figure III-10. Baseline-only technologies (MT5, AT5, AT7L2, AT9L2, and
CVT) are grayed and can only be assigned as initial vehicle
transmission configurations. Further details about transmission path
modeling can be found in TSD Chapter 3.2.
[GRAPHIC] [TIFF OMITTED] TR02MY22.074
(b) Transmission Analysis Fleet Assignments
The wide variety of transmissions on the market are classified into
discrete transmission technology paths for this analysis. These paths
are used to model the most representative characteristics, costs, and
performance of the fuel economy-improving technologies most likely
available during the rulemaking time frame.
To generate the analysis fleet, we gather data on transmissions
from manufacturer mid-model year CAFE compliance submissions and
publicly available manufacturer specification sheets. We use the data
to assign transmissions in the analysis fleet and determine which
platforms share transmissions.
We specify transmission type, number of gears, and high-efficiency
gearbox (HEG) level for the baseline fleet assignment. The number of
gears in the assignments for automatic and manual transmissions usually
match the number of gears listed by the data sources, with some
exceptions. We did not model four-speed transmissions in Autonomie for
this analysis due to their rarity and low likelihood of being used in
the future, so we assigned MY 2020 vehicles with an AT4 or MT4 to an
AT5 or MT5 baseline, respectively. Some dual-clutch transmissions were
also an exception; dual-clutch transmissions with seven gears were
assigned to DCT6.
For automatic and continuously variable transmissions, the
identification of the most appropriate transmission path model required
additional steps; this is because high-efficiency gearboxes are
considered in the analysis but identifying HEG level from specification
sheets alone was not always straightforward. We conducted a review of
the age of the transmission design, relative performance versus
previous designs, and technologies incorporated and used the
information obtained to assign an HEG level. No automatic transmissions
in the analysis fleet were determined to be at HEG Level 3. In
addition, no six-speed automatic transmissions were assigned HEG Level
2. However, we found all 7-speed, all 9-speed, all 10-speed, and some
8-speed automatic transmissions to be advanced transmissions operating
at HEG Level 2 equivalence. Eight-speed automatic transmissions
developed after MY 2017 are assigned HEG Level 2. All other
transmissions are assigned to their respective transmission's baseline
level. The baseline (HEG level 1) technologies available include AT6,
AT8, and CVT.
We assigned any vehicle in the analysis fleet with an electric
powertrain a direct drive (DD) transmission. This designation is for
informational purposes; if specified, the transmission will not be
replaced or updated by the model. Similarly, we assigned any power-
split hybrid vehicle
[[Page 25804]]
an eCVT transmission. As with the direct drive (DD) transmission, this
designation is for informational purposes.
In addition to technology type, gear count, and HEG level,
transmissions are characterized in the analysis fleet by drive type and
vehicle architecture. Drive types considered in the analysis include
front-, rear-, all-, and four-wheel drive. Our definition of drive
types in the analysis does not always align with manufacturers' drive
type designations; see the end of this subsection for further
discussion. These characteristics, supplemented by information such as
gear ratios and production locations, showed that manufacturers use
transmissions that are the same or similar on multiple vehicle models.
Manufacturers have told the agency they do this to control component
complexity and associated costs for development, manufacturing,
assembly, and service. If multiple vehicle models share technology
type, gear count, drive configuration, internal gear rations, and
production location, the transmissions are treated as a single group
for the analysis. Vehicles in the analysis fleet with the same
transmission configuration adopt additional fuel-saving transmission
technology together, as described in Section III.C.2.a).
Shared transmissions are designated and tracked in the CAFE Model
input files using transmission codes. Transmission codes are six-digit
numbers that are assigned to each transmission and encode information
about them. This information includes the manufacturer, drive
configuration, transmission type, and number of gears. TSD Chapter
3.2.4 includes more information on the transmission codes designated in
the analysis fleet.
We assigned different transmission codes to variants of a
transmission that may have appeared to be similar based on the
characteristics considered in the analysis but are not mechanically
identical. We distinguish among transmission variants by comparing
their internal gear ratios and production locations. For example,
several Ford nameplates carry a rear-wheel drive, 10-speed automatic
transmission. These nameplates comprise a wide variety of body styles
and use cases, and so we assigned different transmission codes to these
different nameplates. Because we assigned different transmission codes,
we are not treating them as ``shared'' for the purposes of the analysis
and the transmission models have the opportunity to adopt transmission
technologies independently.
Note that when we determine the drive type of a transmission, the
assignment of all-wheel drive (AWD) versus four-wheel drive (4WD) is
determined by vehicle architecture. Our assignment does not necessarily
match the drive type used by the manufacturer in specification sheets
and marketing materials. We assigned vehicles with a powertrain capable
of providing power to all wheels and a transverse engine (front-wheel
drive architecture), AWD. We assigned vehicles with power to all four
wheels and a longitudinal engine (rear-wheel drive architecture), 4WD.
(c) Transmission Adoption Features
We designated transmission technology pathways to prevent ``branch
hopping''--changes in transmission type that would correspond to
significant changes in transmission architecture--for vehicles that are
relatively advanced on a given pathway. The CAFE Model prevents
``branch hopping'' recognizing that stranded capital associated with
moving from one transmission architecture to another is relevant and
not entirely feasible when making technology selections. Stranded
capital is discussed in Section III.C.6. For example, a vehicle with an
automatic transmission with more than five gears cannot adopt a dual-
clutch transmission. For a more detailed discussion of path logic
applied in the analysis, including technology supersession logic and
technology mutual exclusivity logic, please see CAFE Model
Documentation S4.5 Technology Constraints (Supersession and Mutual
Exclusivity).
Some technologies modeled in the analysis are not yet in
production, and therefore are not assigned in the baseline fleet.
Nonetheless, we made these technologies available for future adoption
because, they are projected to be available in the analysis timeframe.
For instance, we did not observe an AT10L3 in the baseline fleet, but
it is plausible that manufacturers that employ AT10L2 technology may
improve the efficiency of those AT10L2s in the rulemaking timeframe.
In the following sections we discuss specific adoption features
applied to each type of transmission technology.
When we adopt electrification technologies, the transmissions
associated with those technologies will supersede the existing
transmission on a vehicle. We superseded the transmission technology
when P2 hybrids, plug-in hybrids, or battery electric vehicle
technologies are applied. For more information, see Section III.D.3.c).
We preclude adoption of other transmission types once a platform
progresses past an AT6 on the automatic transmission path. We use this
restriction to avoid the significant level of stranded capital loss
that could result from adopting a completely different transmission
type shortly after adopting an advanced transmission, which would occur
if a different transmission type were adopted after AT6 in the
rulemaking timeframe.
We do not allow vehicles that do not start with AT7L2 or AT9L2
transmissions to adopt those technologies during simulation. We
observed that MY 2020 vehicles with those technologies were primarily
luxury performance vehicles and concluded that other vehicles would
likely not adopt those technologies. We concluded that this was also a
reasonable assumption for the analysis fleet because vehicles that have
moved to more advanced automatic transmissions have overwhelmingly
moved to 8-speed and 10-speed transmissions.\304\
---------------------------------------------------------------------------
\304\ 2020 EPA Automotive Trends Report, at 64, figure 4.18.
---------------------------------------------------------------------------
We limited CVT adoption by technology path logic. We do not allow
CVTs to be adopted by vehicles that do not originate with a CVT or by
vehicles with multispeed transmissions beyond AT6 in the baseline
fleet. Once on the CVT path, we only allow the platform to apply
improved CVT technologies. We restrict application of CVT technology on
larger vehicles because of the higher torque (load) demands of those
vehicles and CVT torque limitations based on durability constraints.
Additionally, we use this restriction to avoid the loss of significant
level of stranded capital.
We allow vehicles in the baseline fleet that have DCTs to apply an
improved DCT and allows vehicles with an AT5 to consider DCTs.
Drivability and durability issues with some DCTs have resulted in a low
relative adoption rate over the last decade; this is also broadly
consistent with manufacturers' technology choices.\305\
---------------------------------------------------------------------------
\305\ Ibid.
---------------------------------------------------------------------------
We only allow vehicles with MTs to adopt more advanced manual
transmissions for this analysis, because other transmission types do
not provide a similar driver experience (utility). We do not allow
vehicles with MTs to adopt ATs, CVTs, or DCT technologies under any
circumstance. We do not allow vehicles with other transmissions to
adopt MTs in recognition of the low
[[Page 25805]]
customer demand for manual transmissions.\306\
---------------------------------------------------------------------------
\306\ Ibid.
---------------------------------------------------------------------------
(d) Transmission Effectiveness Modeling
For this analysis, we use the Autonomie full vehicle simulation
tool to model the interaction between transmissions and the full
vehicle system to improve fuel economy, and how changes to the
transmission subsystem influence the performance of the full vehicle
system. Our full vehicle simulation approach clearly defines the
contribution of individual transmission technologies and separates
those contributions from other technologies in the full vehicle system.
Our modeling approach follows the recommendations of the 2015 NAS
Report to use full vehicle modeling supported by application of
collected improvements at the sub-model level.\307\ See TSD Chapter
3.2.4 for more details on transmission modeling inputs and results.
---------------------------------------------------------------------------
\307\ 2015 NAS Report, at p. 292.
---------------------------------------------------------------------------
The only technology effectiveness results that were not directly
calculated using the Autonomie simulation results were for the AT6L2.
We determined the model for this specific technology was inconsistent
with the other transmission models and overpredicted effectiveness
results. Evaluation of the AT6L2 transmission model revealed an
overestimated efficiency map was developed for the AT6L2 model. The
high level of efficiency assigned to the transmission surpassed
benchmarked advanced transmissions.\308\ To address the issue, we
replaced the effectiveness values of the AT6L2 model. We replaced the
effectiveness for the AT6L2 technology with analogous effectiveness
values from the AT7L2 transmission model. For additional discussion on
how analogous effectiveness values are determined please see Section
III.D.1.d)(2).
---------------------------------------------------------------------------
\308\ Autonomie model documentation, Chapter 5.3.4, Transmission
Performance Data.
---------------------------------------------------------------------------
The effectiveness values for the transmission technologies, for all
ten vehicle technology classes, are shown in Figure III-11. Each of the
effectiveness values shown is representative of the improvements seen
for upgrading only the listed transmission technology for a given
combination of other technologies. In other words, the range of
effectiveness values we show for each specific technology, e.g.,
AT10L3, represents the addition of the AT10L3 technology to every
technology combination that could select the addition of AT10L3. We
must emphasize that the graph shows the change in fuel consumption
values between entire technology keys,\309\ and not the individual
technology effectiveness values. Using the change between whole
technology keys captures the complementary or non-complementary
interactions among technologies. In the graph, the box shows the inner
quartile range (IQR) of the effectiveness values and whiskers extend
out 1.5 x IQR. The dots outside of the whiskers show values for
effectiveness that are outside these bounds.
---------------------------------------------------------------------------
\309\ Technology key is the unique collection of technologies
that constitutes a specific vehicle, see Section III.C.4.c).
---------------------------------------------------------------------------
BILLING CODE 4510-59-P
[GRAPHIC] [TIFF OMITTED] TR02MY22.075
[[Page 25806]]
We also want to note the effectiveness for the MT5, AT5, eCVT and
DD technologies are not shown. The DD and eCVT do not have standalone
effectiveness values because they are only implemented as part of
electrified powertrains. The MT5 and AT5 also have no effectiveness
values because both technologies are baseline technologies against
which all other technologies are compared.
---------------------------------------------------------------------------
\310\ The data used to create this figure can be found the FE_1
Improvements file.
---------------------------------------------------------------------------
(e) Transmission Costs
We use transmission costs drawn from several sources, including the
2015 NAS Report and NAS-cited studies for this analysis. TSD Chapter
3.2.7 provides a detailed description of the cost sources used for each
transmission technology. In Table III-14 we show an example of absolute
costs for transmission technologies in 2018$ across select model years,
which demonstrates how we applied cost learning to the transmission
technologies over time. Note, because transmission hardware is often
shared across vehicle classes, transmission costs are the same for all
vehicle classes. For a full list of all absolute transmission costs
used in the analysis across all model years, see the Technologies file.
[GRAPHIC] [TIFF OMITTED] TR02MY22.076
3. Electrification Paths
The electric paths include a large set of technologies that share
the common element of using electrical power for certain vehicle
functions that were traditionally powered mechanically by IC engines.
Electrification technologies thus can range from electrification of
specific accessories (for example, electric power steering to reduce
engine loads by eliminating parasitic losses) to electrification of the
entire powertrain (as in the case of a battery electric vehicle).
The following subsections discuss how we define each
electrification technology in the CAFE Model and the electrification
pathways down which a vehicle can travel in the compliance simulation.
The subsections also discuss how we assigned electrified vehicle
technologies to vehicles in the analysis fleet, any limitations on
electrification technology adoption, and the specific effectiveness and
cost assumptions that we use in the Autonomie and CAFE Model analysis.
We received many comments on electrification technologies, and
specifically on technology costs. Commenters were generally supportive
of our use of Argonne's BatPaC battery cost model to determine costs of
batteries for different electrified powertrains.\311\ In contrast, we
received several comments indicating that we overstated the cost for
hybrid vehicles and batteries,\312\ in particular due to non-battery
electrification component costs. These comments and our approach to
addressing them for this final rule are discussed in the following
sections.
---------------------------------------------------------------------------
\311\ Auto Innovators, Docket No. NHTSA-2021-0053-0021, at 55;
Kia, Docket No. NHTSA-2021-0053-1525, at p. 5.
\312\ Tesla, Inc. (Tesla), Docket No. NHTSA-2021-0053-1480, at
9-10; Toyota, Docket No. NHTSA-2021-0053-1568, at 7; ICCT, Docket
No. NHTSA-2021-0053-1581, at p. 10.
---------------------------------------------------------------------------
Electrification technologies are a complex set of systems that each
manufacturer individually optimizes based on cost, performance,
reliability, durability, customer acceptance and other metrics. We
attempted to capture these complexities to provide a reasonable
assessment of the costs and
[[Page 25807]]
benefits of more stringent fuel economy standards. We expect that there
will be future opportunities to improve upon this work as more
substantiated data on electrification technologies becomes available.
(a) Electrification Modeling in the CAFE Model
The CAFE Model defines the technology pathway for each type of
electrification grouping in a logical progression. Whenever the CAFE
Model converts a vehicle model to one of the available electrified
systems, both effectiveness and costs are updated according to the
specific components' modeling algorithms. Additionally, all
technologies on the electrification paths are mutually exclusive and
are evaluated in parallel. For example, the model may evaluate PHEV20
technology prior to having to apply SS12V or strong hybrid technology.
The specific set of algorithms and rules are discussed further in the
sections below, and more detailed discussions are included in the CAFE
Model Documentation. The specifications for each electrification
technology that we include in the analysis is discussed below.
The technologies that we include on the three vehicle-level paths
pertaining to the electrification and electric improvements defined
within the modeling system are illustrated in Figure III-12. As shown
in the Electrification path, the baseline-only CONV technology is
grayed out. This technology is used to denote whether a vehicle comes
in with a conventional powertrain (i.e., a vehicle that does not
include any level of hybridization) and to allow the model to properly
map to the Autonomie vehicle simulation database results. If multiple
technologies from different pathways come together on single technology
set, then those previous technology pathways are disabled. This avoids
unrealistic adoption of legacy technologies as the simulation
progresses from model year to model year. For example, in the Figure
III-12 PHEVs converge on to BEVs then all the PHEVs are disabled from
adoption.
[GRAPHIC] [TIFF OMITTED] TR02MY22.077
[[Page 25808]]
SS12V: 12-volt stop-start (SS12V), sometimes referred to as start-
stop, idle-stop, or a 12-volt micro hybrid system, is the most basic
hybrid system that facilitates idle-stop capability. In this system,
the integrated starter generator is coupled to the internal combustion
(IC) engine. When the vehicle comes to an idle-stop the IC engine
completely shuts off, and, with the help of the 12-volt battery, the
engine cranks and starts again in response to throttle to move the
vehicle, application or release of the brake pedal to move the vehicle.
The 12-volt battery used for the start-stop system is an improved unit
compared to a traditional 12-volt battery, and is capable of higher
power, increased life cycle, and capable of minimizing voltage drop on
restart. This technology is beneficial to reduce fuel consumption and
emissions when the vehicle frequently stops, such as in city driving
conditions or in stop and go traffic. SS12V can be applied to all
vehicle technology classes. As discussed further below, for this final
rule analysis we lowered the cost of the battery used in the SS12V
system to reflect a more widely utilized SS12V battery chemistry.
Next, mild and strong hybrid systems, discussed in the following
paragraphs, can be classified based on the location of the electric
motor in the system. Depending on the location of the electric machine,
the hybrid technologies are classified as follows:
P0: Motor located at the primary side of the engine,
P1: Motor located at the flywheel side of the engine,
P2: Motor located between engine and transmission,
P3: Motor located at the transmission output, and
P4: Motor located on the axle.
BISG: The belt integrated starter generator, sometimes referred to
as a mild hybrid system or P0 hybrid, provides idle-stop capability and
uses a higher voltage battery with increased energy capacity over
conventional automotive batteries. These higher voltages allow the use
of a smaller, more powerful, and efficient electric motor/generator to
replace the standard alternator. In BISG systems, the motor/generator
is coupled to the engine via belt (similar to a standard alternator).
In addition, these motor/generators can assist vehicle braking and
recover braking energy while the vehicle slows down (regenerative
braking) and in turn can propel the vehicle at the beginning of launch,
allowing the engine to be restarted later. Some limited electric assist
is also provided during acceleration to improve engine efficiency. Like
micro hybrids, BISG can be applied to all vehicles in the analysis
except for Engine 26a (VCR). We assume all mild hybrids are fixed
battery capacity 48-volt systems with engine belt-driven motor/
generators.
ICCT commented that we should consider another type of mild hybrid
system that has a higher power output, which leads to an increased
efficiency compared to the 48V mild hybrid assumed in the NPRM
analysis. The increased benefit from this higher power output mild
hybrids is due to its placement in the powertrain in P1 and P2
positions rather than P0.313 314
---------------------------------------------------------------------------
\313\ ICCT, at p. 2.
\314\ Autonomie assumes a P0 position for mild hybrid 48-volt
systems.
---------------------------------------------------------------------------
We agree with ICCT that mild hybrids in configurations other than
the P0 position offer higher improvements compared to mild hybrids
configured in the P0 position. However, this inherently increases the
cost of the system and makes the system less cost effective compared to
traditional strong hybrids for a few reasons. First, like a mild hybrid
CISG system,\315\ non-P0 mild hybrid architecture requires significant
changes to the area of the powertrain where the electric machine
components are installed compared to P0 BISG systems. Second, these
system's higher power output will also require a higher battery pack
capacity, which could also increase costs. Separately, no manufacturer
has indicated that they will adopt this type of mild hybrid
configuration in the rulemaking time frame. For MYs 2024-2026, the CAFE
Model estimates that a significant penetration of strong hybrids and
plug-in hybrids is required to meet the analyzed alternatives. Similar
to what we observed in past rulemakings with the CISG system, the non-
P0 mild hybrid is not a cost-effective way for manufacturers to meet
standards in the rulemaking time frame. Accordingly, we did not add an
additional mild hybrid technology for this final rule.
---------------------------------------------------------------------------
\315\ We discuss challenges with CISG mild hybrids, a system
that is similar to the P2 hybrid system, further in TSD Chapter
3.3.1.2.
---------------------------------------------------------------------------
SHEVP2/SHEVPS: A strong hybrid vehicle is a vehicle that combines
two or more propulsion systems, where one uses gasoline (or diesel),
and the other captures energy from the vehicle during deceleration or
braking, or from the engine and stores that energy for later used by
the vehicle. This analysis evaluated the following strong hybrid
systems: hybrids with P2 parallel drivetrain architectures (SHEVP2),
and hybrids with power-split architectures (SHEVPS). Both strong hybrid
types provide start-stop or idle-stop functionality, regenerative
braking capability, and vehicle launch assist. A SHEVPS has a higher
potential for fuel economy improvement than a SHEVP2, although it costs
more and has a lower power density.\316\
---------------------------------------------------------------------------
\316\ Kapadia, J., Kok, D., Jennings, M., Kuang, M. et al.,
``Powersplit or Parallel--Selecting the Right Hybrid Architecture,''
SAE Int. J. Alt. Power. 6(1):2017, doi:10.4271/2017-01-1154.
---------------------------------------------------------------------------
P2 parallel hybrids (SHEVP2) are a type of hybrid vehicle that use
a transmission-integrated electric motor placed between the engine and
a gearbox or CVT, with a clutch that allows decoupling of the motor/
transmission from the engine. Disengaging the clutch allows all-
electric operation and more efficient brake-energy recovery. Engaging
the clutch allows coupling of the engine and electric motor and, when
combined with a transmission, reduces gear-train losses relative to
power-split or 2-mode hybrid systems. P2 hybrid systems typically rely
on the internal combustion engine to deliver high, sustained power
levels. Electric-only mode is used when power demands are low or
moderate.
An important feature of the SHEVP2 system is that it can be applied
in conjunction with most engine technologies. Accordingly, once a
vehicle is converted to a SHEVP2 powertrain in the compliance
simulation, the CAFE Model allows the vehicle to adopt the conventional
engine technology that is most cost effective, regardless of relative
location of the existing engine on the engine technology path. This
means a vehicle could adopt a lower technology engine when the CAFE
Model converts it to a SHEVP2 strong hybrid. For example, a vehicle in
the analysis fleet that starts with a TURBO2 engine could adopt a
TURBO1 engine with the SHEVP2 system, if that TURBO1 engine allows the
vehicle to meet fuel economy standards more cost effectively.
The power-split hybrid (SHEVPS) is a more advanced electrified
system than SHEVP2 hybrid. The SHEVPS electric drive replaces the
traditional transmission with a single planetary gear set (the power-
split device) and a motor/generator.\317\
---------------------------------------------------------------------------
\317\ For more discussion of SHEVPS operation and
characteristics, see TSD Section 3.3.
---------------------------------------------------------------------------
Table III-15 below shows the configuration of conventional engines
and transmissions used with strong hybrids for this analysis. The
SHEVPS powertrain configuration is paired with a planetary transmission
(eCVT) and Atkinson engine (Eng26). This configuration is designed to
maximize efficiency at the cost of reduced towing
[[Page 25809]]
capability and real-world acceleration performance.\318\ In contrast,
SHEVP2 powertrains are paired with an advanced 8-speed automatic
transmission (AT8L2) and can be paired with most conventional
engines.\319\
---------------------------------------------------------------------------
\318\ Kapadia, J., D, Kok, M. Jennings, M. Kuang, B. Masterson,
R. Isaacs, A. Dona. 2017. Powersplit or Parallel--Selecting the
Right Hybrid Architecture. SAE International Journal of Alternative
Powertrains 6 (1): 68-76. https://doi.org/10.4271/2017-01-1154
(accessed: Feb. 11, 2022).
\319\ We did not model SHEVP2s with VTGe (Eng23c) and VCR
(Eng26a).
[GRAPHIC] [TIFF OMITTED] TR02MY22.078
PHEV: Plug-in hybrid electric vehicles are hybrid electric
vehicles with the means to charge their battery packs from an outside
source of electricity (usually the electric grid). These vehicles have
larger battery packs than strong HEVs with more energy storage and a
greater capability to be discharged than other non-plug-in hybrid
electric vehicles. PHEVs also generally use a control system that
allows the battery pack to be substantially depleted under electric-
only or blended mechanical/electric operation and batteries that can be
cycled in charge-sustaining operation at a lower state of charge than
non-plug-in hybrid electric vehicles. These vehicles generally have a
greater all-electric range than typical strong HEVs. Depending on how
these vehicles are operated, they can use electricity exclusively,
operate like a conventional hybrid, or operate in some combination of
these two modes.
---------------------------------------------------------------------------
\320\ Twenty-one different engines are evaluated with SHEVP2
hybrid architecture: Engine 01, 02, 03, 04, 5b, 6a, 7a, 8a, 12, 12-
DEAC, 13, 14, 17, 18, 19, 20, 21, 22b, 23b, 24, 24-Deac. See Section
III.D.1 for these engine specifications.
---------------------------------------------------------------------------
There are four PHEV architectures included in this analysis that
reflect combinations of two levels of all-electric range (AER) and two
engine types. We use 20 miles AER and 50 miles AER to reasonably span
the various PHEV AERs in the market, and their effectiveness and cost.
We use an Atkinson engine and a turbocharged downsized engine to span
the variety of engines available in the market.
PHEV20/PHEV20H and PHEV50/PHEV50H are essentially a SHEVPS with a
larger battery and the ability to drive with the engine turned off. In
the CAFE Model, the designation ``H'' in PHEVxH could represent another
type of engine configuration, but for this analysis we use the same
effectiveness values as PHEV20 and PHEV50 to represent PHEV20H and
PHEV50H, respectively. The PHEV20/PHEV20H represents a ``blended-type''
plug-in hybrid that can operate in all-electric (engine off) mode only
at light loads and low speeds, and must blend electric motor and engine
power together to propel the vehicle at medium or high loads and
speeds. The PHEV50/PHEV50H represents an extended range electric
vehicle (EREV) that can travel in all-electric mode even at higher
speeds and loads. Engine sizing, batteries, and motors for these PHEVs
are discussed further in Section III.D.3.d).
PHEV20T and PHEV50T are 20 mile and 50 mile AER vehicles based on
the SHEVP2 engine architecture. The PHEV versions of these
architectures include larger batteries and motors to meet performance
metrics in charge sustaining mode at higher speeds and loads as well as
similar performance and range in all electric mode in city driving and
at higher speeds and loads. For this analysis, the CAFE Model considers
these PHEVs to have an advanced 8-speed automatic transmission (AT8L2)
and TURBO1 (Eng12) in the powertrain configuration. Further discussion
of engine sizing, batteries, and motors for these PHEVs is discussed in
Section III.D.3.d).
Table III-16 shows the different PHEV configurations used in this
analysis.
[[Page 25810]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.079
BEV: Battery electric vehicles are equipped with all-electric drive
systems powered by energy-optimized batteries charged primarily by
electricity from the grid. BEVs do not have a combustion engine or
traditional transmission. Instead, BEVs rely on all electric
powertrains with a single speed gear reduction in place of an advanced
transmission. Battery electric vehicle range varies by vehicle and
battery pack size.
We simulate BEVs with ranges of 200, 300, 400 and 500 miles in the
CAFE Model. BEV range is measured pursuant to EPA test procedures and
guidance.\321\ The CAFE Model assumes a BEV direct drive transmission
is unique to each vehicle (i.e., the transmissions are not shared by
any other vehicle) and that no further improvements to the transmission
are available.
---------------------------------------------------------------------------
\321\ BEV electric ranges are determined per EPA guidance
Document. ``EPA Test Procedure for Electric Vehicles and Plug-in
Hybrids.'' https://fueleconomy.gov/feg/pdfs/EPA%20test%20procedure%20for%20EVs-PHEVs-11-14-2017.pdf. November
14, 2017. (Accessed: May 3, 2021)
---------------------------------------------------------------------------
An important note about the BEVs offered in this analysis is that
the CAFE Model does not account for vehicle range when considering
additional BEV technology adoption. That is, the CAFE Model does not
have an incentive to build BEV 300, 400, and 500s, because the BEV200
is just as efficient as those vehicles and counts the same toward
compliance, but at a significantly lower cost because of the smaller
battery.\322\ While manufacturers have been building 200-mile range
BEVs, those vehicles have generally been passenger cars. Manufacturers
have told us that greater range is important for meeting the needs of
broader range of consumers and to increase consumer demand. More
recently, there has been a trend towards manufacturers building higher
range BEVs in the market, and manufacturers building CUV/SUV and pickup
truck BEVs.\323\ To simulate the potential relationship of BEV range to
consumer demand, we have included several adoption features for BEVs.
These are discussed further in Section III.D.3.c).
---------------------------------------------------------------------------
\322\ See section III.D.3.d Electrification Effectiveness
Modeling for effectiveness of different rage BEVs.
\323\ 2021 EPA Automotive Trends Report, at p. 58.
---------------------------------------------------------------------------
FCEV: Fuel cell electric vehicles are equipped with an all-electric
drivetrain, but unlike BEVs, FCEVs do not solely rely on batteries;
rather, electricity to run the FCEV electric motor is mainly generated
by an onboard fuel cell system. FCEV architectures are similar to
series hybrids,\324\ but with the engine and generator replaced by a
fuel cell. Commercially available FCEVs consume hydrogen to generate
electricity for the fuel cell system, with most automakers using high
pressure gaseous hydrogen storage tanks. FCEVs are currently produced
in limited numbers and are available in limited geographic areas where
hydrogen refueling stations are accessible. For reference, in MY 2020,
only four FCEV models were offered for sale, and since 2014 only 12,081
FCEVs have been sold.325 326 327
---------------------------------------------------------------------------
\324\ Series hybrid architecture is a strong hybrid that has the
engine, electric motor and transmission in series. The engine in a
series hybrid drives a generator that charges the battery.
\325\ Argonne National Laboratory, ``Light Duty Electric Drive
Vehicles Monthly Sales Update.'' Energy Systems Division, https://www.anl.gov/es/light-duty-electric-drive-vehicles-monthly-sales-updates. (Accessed: Dec. 15, 2021)
\326\ See the MY 2020 Market Data file. The four vehicles are
the Honda Clarity, Hyundai Nexo and Nexo Blue, and Toyota Mirai.
\327\ These are majority leased vehicles that are returned back
to the manufacturer rather than resold as a used vehicle.
---------------------------------------------------------------------------
For this analysis, the CAFE Model simulates a FCEV with a range of
320 miles. Any powertrain type can adopt a FCEV powertrain; however, to
account for limited market penetration and unlikely increased adoption
in the rulemaking timeframe, technology phase in caps are used to
control how many FCEVs a manufacturer can build. The details of this
concept are further discussed in Section III.D.3.c).
(b) Electrification Analysis Fleet Assignments
We use electrification technologies assigned in the baseline fleet
as the starting point for regulatory analysis. These assignments are
based on manufacturer-submitted CAFE compliance information, publicly
available technical specifications, marketing brochures, articles from
reputable media outlets, and data from Wards Intelligence.\328\
---------------------------------------------------------------------------
\328\ ``U.S. Car and Light Truck Specifications and Prices, '20
Model Year.'' Wards Intelligence, 3 Aug. 2020,
wardsintelligence.informa.com/WI964244/US-Car-and-Light-Truck-Specifications-and-Prices-20-Model-Year (accessed: Feb. 11, 2022).
---------------------------------------------------------------------------
Table III-17 gives the penetration rates of electrification
technologies eligible to be assigned in the baseline fleet. Over half
of the fleet had some level of electrification, with the vast majority
of these being micro hybrids. PHEVs represented 0.5 percent of the MY
2020 baseline fleet. BEVs represented less than 2 percent of MY 2020
baseline fleet; BEV300 was the most common BEV technology, while no
BEV500s were observed.
[[Page 25811]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.080
BILLING CODE 4910-59-C
Micro and mild hybrids refer to the presence of SS12V and BISG,
respectively. The data sources discussed above are used to identify the
presence of these technologies on vehicles in the fleet. Vehicles are
assigned one of these technologies only if its presence can be
confirmed with manufacturer brochures or technical specifications.
Strong hybrid technologies include SHEVPS and SHEVP2. Note that
P2HCR0, P2HCR1, P2HCR1D, and P2HCR2 are not assigned in the fleet and
are only available to be applied by the model. When possible,
manufacturer specifications are used to identify the strong hybrid
architecture type. In the absence of more sophisticated information,
hybrid architecture is determined by number of motors. Hybrids with one
electric motor are assigned P2, and those with two motors are assigned
PS. We sought comment in the NPRM on additional ways the agency could
perform initial hybrid assignments based on publicly available
information or technical publications. We did not receive any
substantive comments regarding baseline fleet strong hybrid
assignments. Accordingly, this final rule analysis uses the same
approach to assigning SHEVPS and SHEVP2 in the baseline fleet.
Plug-in hybrid technologies PHEV20/20T and PHEV50/50T are assigned
in the baseline fleet. PHEV20H and PHEV50H are not assigned in the
fleet and are only available to be applied by the model. Vehicles with
an electric-only range of 40 miles or less are assigned PHEV20;
vehicles with a range above 40 miles are assigned PHEV50. They are
respectively assigned PHEV20T/50T if the engine is turbocharged (i.e.,
if it would qualify for one of technologies on the turbo engine
technology pathway). We also calculate baseline fuel economy values for
PHEV technologies as part of the PHEV analysis fleet assignments; that
process is described in detail in TSD Chapter 3.3.2.
Battery electric vehicle and fuel cell technologies include BEV200/
300/400/500 and FCEV with a 320-mile range. The BEV technologies are
assigned to vehicles based on range thresholds that best account for
vehicles' existing range capabilities while allowing room for the model
to potentially apply more advanced electrification technologies.
Vehicles with all-electric powertrains that use hydrogen fuel are
assigned FCEV.
For more detail about the electrification analysis fleet assignment
process, see TSD Chapter 3.3.2.
(c) Electrification Adoption Features
Multiple types of adoption features apply to the electrification
technologies. The hybrid/electric technology path logic dictates how
different vehicle types can adopt different levels of electrification
technology. Broadly speaking, more advanced levels of hybridization or
electrification supersede all prior levels, with certain technologies
within each level being mutually exclusive.
As discussed further below, SKIP logic--restrictions on the
adoption of certain technologies--apply to plug-in (PHEV) and strong
hybrid vehicles (SHEV). Some technologies on these pathways are
``skipped'' if a vehicle is high performance, requires high towing
capabilities as a pickup truck, or belongs to certain manufacturers who
have demonstrated that their future product plans will more than likely
not include the technology. The specific criteria for SKIP logic for
each applicable electrification technology is expanded on later in this
section.
This section also discusses the supersession of engines and
transmissions on vehicles that adopt SHEV or PHEV powertrains. To
manage the complexity of the analysis, these types of hybrid
powertrains are modeled with several specific engines and
transmissions, rather than in multiple configurations. Therefore, the
cost and effectiveness values SHEV and PHEV technologies consider these
specific engines and transmissions.
Finally, phase-in caps limit the adoption rates of battery electric
(BEV) and fuel cell electric vehicles (FCEV). We set the phase-in caps
to account for current market share, scalability, and reasonable
consumer adoption rates of each technology. TSD Chapter 3.3.3 discusses
the electrification phase-in caps and the reasoning behind them in
detail.
[[Page 25812]]
The only adoption feature applicable to micro and mild hybrid
technologies is path logic. The pathway consists of a linear
progression starting with a conventional powertrain with no
electrification at all, which is superseded by SS12V, which in turn is
superseded by BISG. Vehicles can only adopt micro and mild hybrid
technology if the vehicle does not already have a more advanced level
of electrification.
The adoption features that apply to strong hybrid technologies
include path logic, powertrain substitution, and vehicle class
restrictions. Per the defined technology pathways, SHEVPS, SHEVP2, and
the P2HCR technologies are considered mutually exclusive. In other
words, when the model applies one of these technologies, the others are
immediately disabled from future application. However, all vehicles on
the strong hybrid pathways can still advance to one or more of the
plug-in hybrid technologies.
When the model applies any strong hybrid technology to a vehicle,
the transmission technology on the vehicle is superseded. Regardless of
the transmission originally present, P2 hybrids adopt an 8-speed
automatic transmission (AT8L2), and PS hybrids adopt an electronic
continuously variable transmission (eCVT).
When the model applies SHEVP2 technology, the model can consider
various engine options to pair with the SHEVP2 architecture according
to existing engine path constraints, considering relative cost
effectiveness. For SHEVPS technology, the existing engine is replaced
with Eng26, which is a full Atkinson cycle engine.
SKIP logic is also used to constrain adoption for SHEVPS, P2HCR0,
P2HCR1, and P2HCR1D. These technologies are ``skipped'' for vehicles
with engines \329\ that met one of the following conditions:
---------------------------------------------------------------------------
\329\ This refers to the engine assigned to the vehicle in the
2020 baseline fleet.
---------------------------------------------------------------------------
The engine belongs to an excluded manufacturer; \330\
---------------------------------------------------------------------------
\330\ Excluded manufacturers included BMW, Daimler, and Jaguar
Land Rover.
---------------------------------------------------------------------------
The engine belongs to a pickup truck (i.e., the engine is on a
vehicle assigned the ``pickup'' body style);
The engine's peak horsepower is more than 405 HP; or if
The engine is on a non-pickup vehicle but is shared with a pickup.
No SKIP logic is applied to SHEVP2, however P2HCR2 is not used in
this analysis, as discussed further in Section III.D.1.
The reasons for these conditions are similar to those applied to
HCR engine technologies, discussed in more detail above. In the real
world, pickups and performance vehicles with certain powertrain
configurations cannot adopt the technologies listed above and maintain
vehicle performance without redesigning the entire powertrain. SKIP
logic is put in place to prevent the model from pursuing compliance
pathways that are ultimately unrealistic.
Auto Innovators in their comments for the NPRM, also to the 2018
NPRM, discussed issues with HCR technologies.\331\ Ford had similarly
provided comments in opposition of high dependency on HCR
technologies.\332\ For further discussion of HCR, see Section
III.D.1.c).
---------------------------------------------------------------------------
\331\ Auto Innovators, Docket No. NHTSA-2018-0067-12073-A1, at
p. 139.
\332\ Ford, Docket No. NHTSA-2018-0067-11928-A1, at p. 8.
---------------------------------------------------------------------------
PHEV technologies supersede the micro, mild, and strong hybrids,
and can only be replaced by full electric technologies. Plug-in hybrid
technology paths are also mutually exclusive, with the PHEV20
technologies able to progress to the PHEV50 technologies.
The engine and transmission technologies on a vehicle are
superseded when PHEV technologies are applied to a vehicle. For all
plug-in technologies, the model applies an AT8L2 transmission. For
PHEV20/50 and PHEV20H/50H, the vehicle receives a full Atkinson cycle
engine, Eng26, and for PHEV20T/50T, the vehicle receives a TURBO1
engine, Eng12.
SKIP logic applies to PHEV20/20H and PHEV50/50H under the same four
conditions listed for the strong hybrid technologies in the previous
section, for the same reasons previously discussed.
The adoption of BEVs and FCEVs is limited by both path logic and
phase in caps. BEV200/300/400/500 and FCEV are applied as end-of-path
technologies that superseded previous levels of electrification.
The main adoption feature applicable to BEVs and FCEVs is phase-in
caps, which are defined in the CAFE Model input files as percentages
that represent the maximum rate of increase in penetration rate for a
given technology. They are accompanied by a phase-in start year, which
determines the first year the phase-in cap applies. Together, the
phase-in cap and start year determine the maximum penetration rate for
a given technology in a given year; the maximum penetration rate equals
the phase-in cap times the number of years elapsed since the phase-in
start year. Note that phase-in caps do not inherently dictate how much
a technology is applied by the model. Rather, they represent how much
of the fleet could have a given technology by a given year. Because
BEV200 costs less and has higher effectiveness values than other
advanced electrification technologies,\333\ the model will have
vehicles adopt it first, until it is restricted by the phase-in cap.
---------------------------------------------------------------------------
\333\ This is because BEV200 uses fewer batteries and weighs
less than BEVs with greater ranges.
---------------------------------------------------------------------------
Table III-18 shows the phase-in caps, phase-in year, and maximum
penetration rate through 2050 for BEV and FCEV technologies. For
comparison, the actual penetration rate of each technology in the
baseline fleet is also listed in the fourth column from the left.
[[Page 25813]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.081
The BEV200 phase-in cap is informed by manufacturers' tendency to
move away from low-range vehicle offerings, in part because of consumer
hesitancy to adopt this technology. The advertised range on most
electric vehicles does not reflect extreme cold and hot real-world
driving conditions that affect the utility of already low-range
vehicles.\334\ Many manufacturers have told us that the portion of
consumers willing to accept a vehicle with our lowest range model which
is less than 250 miles of electric range is small, and many
manufacturers do not plan to offer vehicles with less 250 miles of
electric range.\335\
---------------------------------------------------------------------------
\334\ AAA. ``AAA Electric Vehicle Range Testing.'' February
2019. https://www.aaa.com/AAA/common/AAR/files/AAA-Electric-Vehicle-Range-Testing-Report.pdf (accessed: Feb. 11, 2022).
\335\ See also, e.g., Baldwin, Roberto. ``Tesla Model Y Standard
Range Discontinued; CEO Musk Tweets Explanation.'' Car and Driver,
30 Apr. 2021, www.caranddriver.com/news/a35602581/elon-musk-model-y-discontinued-explanation. (Accessed: May 20, 2020)
---------------------------------------------------------------------------
Furthermore, the average BEV range has steadily increased over the
past decade,\336\ perhaps in part as batteries have become more cost
effective. EPA observed in its 2021 Automotive Trends Report that ``the
average range of new EVs has climbed substantially. In model year 2020
the average new EV is projected to have a 286-mile range, or about four
times the range of an average EV in 2011. This difference is largely
attributable to higher production of new EVs with much longer ranges.''
\337\ The maximum growth rate for BEV200 in the model is set
accordingly low to less than 0.1 percent per year. While this rate is
significantly lower than that of the other BEV technologies, the BEV200
phase-in cap allows the penetration rate of low-range BEVs to grow by a
multiple of what is currently observed in the market.
---------------------------------------------------------------------------
\336\ 2021 EPA Automotive Trends Report, at 56, figure 4.17.
\337\ 2021 EPA Automotive Trends Report, at p. 58.
---------------------------------------------------------------------------
For BEV300, 400, and 500, phase-in caps are intended to
conservatively reflect potential challenges in the scalability of BEV
manufacturing, and implementing BEV technology on many vehicle
configurations, including larger vehicles. In the short term, the
penetration of BEVs is largely limited by battery availability. For
example, Tesla is not yet producing electric vans because of cell
production constraints, and it remains a bottleneck in the company's
expansion into new product lines.\338\ Incorporating battery packs that
provide greater amounts of electric range into vehicles also poses its
own engineering challenges. Heavy batteries and large packs may be
difficult to integrate for many vehicle configurations, and require
structural vehicle modifications. Pickup trucks and large SUVs, in
particular, require higher levels of energy as the number of passengers
and/or payload increases, for towing and other high-torque
applications. The BEV400 and 500 phase-in caps reflect these
transitional challenges.
---------------------------------------------------------------------------
\338\ Hyatt, Kyle. ``Tesla Will Build an Electric Van
Eventually, Elon Musk Says.'' Roadshow, CNET, 28 Jan. 2021,
www.cnet.com/roadshow/news/tesla-electric-van-elon-musk/. (Accessed
May 20, 2021)
---------------------------------------------------------------------------
The phase-in cap for FCEVs is based on existing market share as
well as historical trends in FCEV production. FCEV production share in
the past five years has been extremely low, and we set the phase-in cap
accordingly.\339\ As with BEV200, however, the phase-in cap still
allows for the market share of FCEVs to grow several times over.
---------------------------------------------------------------------------
\339\ 2020 EPA Automotive Trends Report, at 52, figure 4.13.
---------------------------------------------------------------------------
We received limited comments on the NPRM referring to how we apply
electrification adoption features for the analysis. In its comments to
EPA's NPRM, submitted to our docket as a courtesy, Auto Innovators
stated they expect that consumers are likely to be more accepting of
longer BEV ranges,\340\ which generally agrees with our expectations
and reasoning in support of why we set the BEV200 phase-in cap.
---------------------------------------------------------------------------
\340\ Auto Innovators, at p. 56.
---------------------------------------------------------------------------
In contrast, ICCT stated that ``there is no engineering or
technical reason to limit application of strong hybrids in the fleet.
Powersplit hybrids may have torque limits, but there is no limitation
for parallel hybrid systems, whether P0, P1, P2, P3, or P4
architecture, as the engine output is routed separately from the motor
output. This is demonstrated by the 2021 Ford F150 pickup truck with a
P2 strong hybrid and the upcoming 2022 Toyota Tundra full-size pickup
truck with a strong hybrid and a conventional 10-speed automatic.''
\341\ ICCT also included examples of hybrid applications in support of
its comment that all vehicles can benefit from hybrid technology that
included the Porsche 918 plug-in hybrid, 2019 Dodge Ram 1500 pickup
truck, and 2021 Ford F150 pickup truck. Similarly, Tesla stated that we
artificially constrained the level of electrification, pointing to the
phase-in caps placed on BEVs.
---------------------------------------------------------------------------
\341\ ICCT, at p. 10.
---------------------------------------------------------------------------
[[Page 25814]]
Regarding ICCT's comment, the NPRM analysis only limited adoption
of SHEVPS and P2HCR combinations for a small number of applications
like pickups, large SUVs that shared pickup engines, and performance-
oriented vehicles. All other conventional vehicles can adopt P2 hybrid
powertrains; for example, the Toyota Tundra, which has a turbocharged
engine paired with a 10-speed automatic transmission is allowed to
adopt P2 hybrid. Additionally, most vehicles can adopt a PS hybrid
system, like the Toyota Highlander. ICCT's other example, the Porsche
918, an $845,000 4.6 liter V8 plug-in P2 hybrid with total 887 hp and
944 lb.-ft of torque, is an example of a vehicle that we could model in
our analysis as a SHEVP2 plug-in hybrid.\342\ However, it is unclear to
what extent the hybrid technology on the Porsche 918 could apply to the
mass market fleet. Other U.S. market Porsche plug-in hybrids, like the
Cayenne E-Hybrid and Panamera E-Hybrid, are modeled as SHEVP2 plug-
hybrids in our analysis.\343\ In all cases, the examples provided by
ICCT were modeled in accordance with their comments.344 345
---------------------------------------------------------------------------
\342\ Porsche. ``The Super Sportscar.'' https://newsroom.porsche.com/en/products/918-spyder-10713.html. (Accessed:
Dec. 17, 2021); Cnet Road and Show. ``Porsche 918 Spyder: Plug-in
hybrid does 94mpg, 198mph.'' https://www.cnet.com/roadshow/pictures/porsche-918-spyder-plug-in-hybrid-does-94mpg-198mph/. (Accessed:
Dec. 17, 2021)
\343\ See the market_data file vehicle codes 4212003, 4212004,
4212009, 4212010, 4222003, 4222004, 4222005, 4222015, 4222016, and
4222017 in the vehicles tab.
\344\ 2022 Toyota Tundra Product Information.
2022_Toyota_Tundra_Product_Information_FINAL.pdf; Buchholz, K.,
``2022 Toyota Tundra: V8 out, twin-turbo hybrid takes over'', SAE.
September 22, 2021. https://www.sae.org/news/2021/09/2022-toyota-tundra-gains-twin-turbo-hybrid-power. (Accessed: Dec. 20, 2021);
Macaulay, S., ``Engineering the 2022 Toyota Tundra'', SAE. October
10, 2021. https://www.sae.org/news/2021/10/engineering-the-2022-toyota-tundra. (Accessed: Dec. 20, 2021)
\345\ ICCT, at p. 8.
---------------------------------------------------------------------------
For both the NPRM and the final rule analysis, BEVs have phase-in
cap limitations applied based on an analysis market availability,
battery costs, and consumer acceptance in the rule making time
frame.\346\ The BEV200 is limited to a greater extent than the BEV300
and BEV400 to account for anticipated market demand for shorter-range
BEVs. As discussed earlier, the 2021 EPA Trends Report that showed that
the average range of BEVs has increased beyond 200 miles to an average
of 286 miles. As such, 300-mile range BEVs and up will most likely
become the status quo for the fleet in the rulemaking time frame.\347\
In addition, the BEV300 and BEV400 caps were not met in either the NPRM
or this final rule analysis for any of the alternatives considered.
This means that even with the market caps in place, the alternatives
did not require manufacturers to increase BEV production because the
standards were met with other cost-effective technologies. Accordingly,
for the final rule analysis, we continued to use the same adoption
features as used in the NPRM to reflect what we believe will
foreseeably occur in the market in the rulemaking time frame.
---------------------------------------------------------------------------
\346\ John Elkin, MIT finds that it might take a long time for
EVs to be as affordable as you want, Digital Trends (November 23,
2019), https://www.digitaltrends.com/cars/mit-study-finds-ev-market-will-stall-in-the-2020s/.
\347\ 20210 EPA Automotive Trends Report, at 536, figure 4.174.
---------------------------------------------------------------------------
(d) Electrification Effectiveness Modeling
For this analysis, we consider a range of electrification
technologies which, when modeled, result in varying levels of
effectiveness at reducing fuel consumption. As discussed above, the
modeled electrification technologies include micro hybrids, mild
hybrids, two different strong hybrids, two different plug-in hybrids
with two separate all electric ranges, full battery electric vehicles,
and fuel cell electric vehicles. Each electrification technology
consists of many complex sub-systems with unique component
characteristics and operational modes. As discussed further below, the
systems that contribute to the effectiveness of an electrified
powertrain in the analysis include the vehicle's battery, electric
motors, power electronics, and accessory loads. Procedures for modeling
each of these sub-systems are broadly discussed in this section and the
Autonomie model documentation.
Argonne uses data from their Advanced Mobility Technology
Laboratory (AMTL) to develop Autonomie's electrified powertrain models.
The modeled powertrains are not intended to represent any specific
manufacturer's architecture but are intended to act as surrogates
predicting representative levels of effectiveness for each
electrification technology.
Autonomie determines the effectiveness of each electrified
powertrain type by modeling the basic components, or building blocks,
for each powertrain, and then combining the components modularly to
determine the overall efficiency of the entire powertrain. Autonomie
identifies components for each electrified powertrain type, and then
interlinks those components to create a powertrain architecture.
Autonomie then models each electrified powertrain architecture and
provides an effectiveness value for each. For example, Autonomie
determines a BEV's overall efficiency by considering the efficiencies
of the battery, the electric traction drive system (the electric
machine and power electronics), and mechanical power transmission
devices. Or, for a SHEVP2, Autonomie combines a very similar set of
components to model the electric portion of the hybrid powertrain, and
then also includes the combustion engine and related power for
transmission components. See TSD Chapter 3.3.4 and the Autonomie model
documentation for a complete discussion of electrification component
modeling.
As discussed earlier in Section III.C.4, Autonomie applies
different powertrain sizing algorithms depending on the type of vehicle
considered because different types of vehicles not only contain
different powertrain components to be optimized, but they must also
operate in different driving modes. While the conventional powertrain
sizing algorithm must consider only the power of the engine, the more
complex algorithm for electrified powertrains must simultaneously
consider multiple factors, which could include the engine power,
electric machine power, battery power, and battery capacity. Also,
while the resizing algorithm for all vehicles must satisfy the same
performance criteria, the algorithm for some electric powertrains must
also allow those electrified vehicles to operate in certain driving
cycles, like the US06 cycle, without assistance of the combustion
engine, and ensure the electric motor/generator and battery can handle
the vehicle's regenerative braking power, all-electric mode operation,
and intended range of travel.
To establish the effectiveness of the technology packages,
Autonomie simulates the vehicles' performance on compliance test
cycles, as discussed in Section III.C.4.348 349 350 The
range of effectiveness for the electrification technologies in this
analysis is a result of the interactions between the components listed
above and how the modeled vehicle operates on its respective test
cycle.
---------------------------------------------------------------------------
\348\ See U.S. EPA, ``How Vehicles are Tested.'' https://www.fueleconomy.gov/feg/how_tested.shtml. (Accessed: May 6, 2021)
\349\ See Autonomie model documentation, Chapter 6, Test
Procedures and Energy Consumption Calculations.
\350\ EPA Guidance Letter. ``EPA Test Procedures for Electric
Vehicles and Plug-in Hybrids.'' Nov. 14, 2017. https://www.fueleconomy.gov/feg/pdfs/EPA%20test%20procedure%20for%20EVs-PHEVs-11-14-2017.pdf. (Accessed: May 6, 2021)
---------------------------------------------------------------------------
[[Page 25815]]
This range of values will result in some modeled effectiveness values
being close to real-world measured values, and some modeled values that
will depart from real-world measured values, depending on the level of
similarity between the modeled hardware configuration and the real-
world hardware and software configurations. This modeling approach
comports with NAS's 2015 recommendation to use full vehicle modeling
supported by application of lumped improvements at the sub-model
level.\351\ In addition, the more recent 2021 NAS Report modeled
electrification technologies with Argonne's Autonomie model using a
similar approach to our analysis.\352\
---------------------------------------------------------------------------
\351\ 2015 NAS Report, at p. 292.
\352\ 2021 NAS Report, at p. 189.
---------------------------------------------------------------------------
We received limited comments regarding electrification
effectiveness modeling. ICCT commented that the agency's strong hybrid
effectiveness data are outdated, because we rely on older powertrain
data like engine maps from the 2010 Toyota Prius, and we do not allow
this engine and other hybrid technologies to improve.\353\ Similarly,
ICCT recommended that further research should be considered to improve
hybrid power management and engines for strong hybrids.\354\ Another
commenter, Walter Kreucher, stated that the electric ranges for
electrified vehicles are lower than what we are modeling. Specifically,
Mr. Kreucher stated that extreme cold, hot, and aggressive driving
conditions have reduced all-electric range anywhere from 39 to 51
percent, based on a study from AAA.\355\
---------------------------------------------------------------------------
\353\ ICCT, at p. 5.
\354\ ICCT, in Appendices at p. 2.
\355\ Walt Kreucher, Docket No. NHTSA-2021-0053-0015, at p. 6.
---------------------------------------------------------------------------
We disagree with ICCT that the electrification technology
represented in this analysis is outdated. The majority of the
technologies were developed specifically to support analysis for this
rulemaking time frame. For example, the hybrid Atkinson engine peak
thermal efficiency was updated based on 2017 Toyota Prius engine
data.356 357 Toyota stated that their current hybrid engines
achieve 41 percent thermal efficiency for their current product line up
which aligns with our modeling.\358\ Similarly, the electric machine
peak efficiency for FCEVs and BEVs is 98 percent and based on the 2016
Chevy Bolt.\359\ Accordingly, we have made no changes to the electric
machine efficiency maps for this final rule analysis.
---------------------------------------------------------------------------
\356\ Atkinson Engine Peak Efficiency is based on 2017 Prius
Peak Efficiency and scaled up to 41 percent. Autonomie Model
Documentation at p. 138.
\357\ Docketed supporting material. ANL--All
Assumptions_Summary_NPRM_022021.xlsx, ANL--Summary of Main Component
Performance Assumptions_NPRM_022021.xlsx, Argonne Autonomie Model
Documentation_NPRM.pdf and ANL--Data Dictionary_NPRM_022021.XLSX.
\358\ Carney, D. ``Toyota unveils more new gasoline ICEs with
40% thermal efficiency''. SAE. April 4, 2018. https://www.sae.org/news/2018/04/toyota-unveils-more-new-gasoline-ices-with-40-thermal-efficiency. (Accessed Dec. 21, 2021)
\359\ F. Momen, K. Rahman, Y. Son and P. Savagian, ``Electrical
propulsion system design of Chevrolet Bolt battery electric
vehicle,'' 2016 IEEE Energy Conversion Congress and Exposition
(ECCE), 2016, pp. 1-8, doi: 10.1109/ECCE.2016.7855076.
---------------------------------------------------------------------------
We agree with Mr. Kreucher that extreme cold and hot conditions
impact electrified vehicle range. We use the latest compliance testing
procedures to appropriately evaluate the effectiveness and range of
electrified technologies, as discussed earlier in this section.
However, there are some extreme conditions, which may impact electric
vehicle range, which may not be captured by the Federal test cycle. The
selection of a phase-in cap for BEV200 is based in part on
consideration of differences in utility, including the potential for
temperature-based (among other things) variations in driving range,
that may affect consumer adoption of shorter-range BEVs. For more
details, see Section III.D.3.c) of this preamble, Electrification
Adoption Features.
The range of effectiveness values for the electrification
technologies, for all ten vehicle technology classes, is shown in
Figure III-13. In the graph, the box shows the inner quartile range
(IQR) of the effectiveness values and whiskers extend out 1.5 x IQR.
The dots outside of the whiskers show values outside these bounds.
BILLING CODE 4910-59-P
[[Page 25816]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.082
BILLING CODE 4910-59-C
(e) Electrification Costs
---------------------------------------------------------------------------
\360\ The data used to create this figure can be found in the
FE_1 Adjustments file.
---------------------------------------------------------------------------
The total cost to electrify a vehicle in this analysis is based on
the battery the vehicle requires, the non-battery electrification
component costs the vehicle requires, and the traditional powertrain
components that must be added or removed from the vehicle to build the
electrified powertrain.
We work collaboratively with the experts at Argonne National
Laboratory to generate battery costs using BatPaC, which is a model
designed to calculate the cost of a vehicle battery for a specified
battery power, energy, and type. For this analysis, Argonne used BatPaC
v4.0 (October 2020 release) to create lookup tables for battery cost
and mass that the Autonomie simulations reference when a vehicle
receives an electrified powertrain. The BatPaC battery cost estimates
for mild hybrids, strong hybrids, plug-in hybrids, and full battery
electric vehicles are generated for a base year, in this case for MY
2020. Accordingly, our BatPaC inputs characterize the state of the
market in MY 2020 and employ a widely utilized cell chemistry
(NMC622),\361\ average estimated battery pack production volume per
plant (25,000), and a plant efficiency or plant cell yield value of 95
percent.
---------------------------------------------------------------------------
\361\ Autonomie model documentation, Chapter 5.9. Argonne
surveyed A2Mac1 and TBS teardown reports for electrified vehicle
batteries and of the five fully electrified vehicles surveyed, four
of those vehicles used NMC622 and one used NMC532. See also Georg
Bieker, A Global Comparison of the Life-Cycle Greenhouse Gas
Emissions of Combustion Engine and Electric Passenger Cars,
International Council on Clean Transportation (July 2021), https://theicct.org/sites/default/files/publications/Global-LCA-passenger-cars-jul2021_0.pdf (``For cars registered in 2021, the GHG emission
factors of the battery production are based on the most common
battery chemistry, NMC622-graphite batteries . . . .''); 2021 NAS
Report, at 87 (``. . . NMC622 is the most common cathode chemistry
in 2019. . . .'').
---------------------------------------------------------------------------
For this final rule, we use a lower SS12V micro hybrid battery cost
that was not developed in BatPaC. The NPRM SS12V fixed battery pack
direct manufacturing cost was $237, across all vehicle classes. For
this final rule analysis, the agency conducted additional research
regarding battery types used in typical SS12V systems yielding a
battery cost that reflects the cost of a more common battery chemistry.
Specifically, absorbed-glass-mat (AGM) batteries are more common in
SS12V systems than the Li-ion-based chemistry used in the NPRM
analysis.362 363 364 The battery pack direct manufacturing
cost for SS12V systems is now $113, across all vehicle classes. This
cost also more closely aligns with the estimated cost of the SS12V
system presented in the 2015 NAS Report.\365\
---------------------------------------------------------------------------
\362\ EPA-HQ-OAR-2021-0208-0144, p. 5-73.
\363\ USABC, ``United States Advanced Battery Consortium Battery
Test Manual For 12 Volt Start/Stop Vehicles.'' January 2018.
Revision 2. Contract DE-AC07-05ID14517.
\364\ H. Tataria; O. Gross; C. Bae; B. Cunningham; J.A. Barnes;
J. Deppe; J. Neubauer. ``USABC Development of 12 Volt Battery for
Start-Stop Application: Preprint'': 10 pp. 2015. https://www.nrel.gov/docs/fy15osti/62680.pdf.
\365\ 2015 NAS Report, at 158.
---------------------------------------------------------------------------
For BEV400 and BEV500, we did not use BatPaC to generate battery
pack costs. Rather, we scaled the BatPaC-generated BEV300 costs to
match the range of BEV400 and BEV500 vehicles to compute a direct
manufacturing cost for those vehicles' batteries. We explained in the
NPRM that we initially examined using BatPaC to model the
[[Page 25817]]
cost and weight of BEV400 and BEV500 packs, however, initial values
from the model could not be validated and were based on assumptions for
smaller sized battery packs. We stated that the initial results
provided cost and weight estimates for BEV400 battery packs out of
alignment with current examples of BEV400s in the market, and there are
currently no examples of BEV500 battery packs in the market against
which to validate the pack results.
Although one example of a BEV500 has entered the market since
publication of the NPRM, it is for a low volume passenger vehicle, and
it is not representative of some pack characteristics and costs for
vehicles in this analysis.366 367 In particular, BatPaC
weights for the BEV400 and BEV500 pickup truck classes often made the
vehicle exceed the light duty 8,500 lb. curb weight threshold for light
duty vehicles, pushing the vehicles into the next weight class. While
this may be representative of what could happen with vehicles that have
more significant range and towing requirements (for example, the 2022
GMC Hummer EV will be a class 2b vehicle \368\), we also believe that
manufacturers will employ different weight saving strategies to keep
heavier vehicles in the light-duty fleet. For this final rule analysis,
we determined that keeping the battery pack mass a more consistent
percentage of vehicle curb weight using the scaling method was a
reasonable assumption, and we will explore how to model this concept
more in future analyses.
---------------------------------------------------------------------------
\366\ CarAndDriver. ``2022 Lucid Air Lucid Air EV's Battery Will
Be a Big 113.0 kWh, Topping Tesla's Best.'' September 2, 2020.
https://www.caranddriver.com/news/a33797162/2021-lucid-air-517-mile-range-113-kwh-battery. Last accessed March 28, 2022.
\367\ Fueleconomy.gov, 2022 Lucid Air. https://www.fueleconomy.gov/feg/Find.do?action=sbs&id=44495&id=44493 (last
accessed: January 23, 2022).
\368\ CarAndDriver. ``2022 GMC Hummer EV EPA Documents Reveal
MPGe, Weight, Other Details.'' Feb 15, 2022. https://www.caranddriver.com/news/a39049358/2022-gmc-hummer-ev-pickup-epa-specs. Last accessed March 28, 2022.
---------------------------------------------------------------------------
Finally, we apply a learning rate to the direct manufacturing cost
to reflect how we expect battery costs could fall over the timeframe
considered in the analysis. For most electrification technologies, the
learning rate that we apply reflects ``midrange'' year-over-year
improvements until MY 2032. Post 2032, the learning rates incrementally
become shallower as battery technology is expected to mature in MY 2033
and beyond. Applying learning curves to the battery pack DMC in
subsequent analysis years reduces costs such that battery pack costs
are believed to represent the manufacturing costs for any future pack,
regardless of cell chemistry, cell format, or production volume.
Unlike the rest of the electrification technologies, however, the
SS12V micro hybrid system uses a shallower learning curve, as shown in
TSD Chapter 3.3.5.2. This shallow curve reflects the maturity of the
technology; as we discuss in TSD Chapter 3.3.2, 50 percent of the MY
2020 fleet utilizes a SS12V micro hybrid system.
TSD Chapter 3.3.5.1 includes more detail about the process to
develop battery costs for this analysis. In addition, all BatPaC-
generated direct manufacturing costs for all technology keys can be
found in the CAFE Model's Battery Costs file, and the Argonne BatPaC
Assumptions file includes the assumptions used to generate the costs,
pack costs, pack mass, cell capacity, $/kW at the pack level, and W/kg
at the pack level for all vehicle classes.
A range of parameters can ultimately influence battery pack
manufacturing costs, including other vehicle improvements (e.g., mass
reduction technology, aerodynamic improvements, or tire rolling
resistance improvements all affect the size and energy of a battery
required to propel a vehicle where all else is equal), and the
availability of materials required to manufacture the
battery.369 370 Or, if manufacturers adopt more
electrification technology than projected in this analysis, increases
in battery pack production volume will likely lower actual battery pack
costs.
---------------------------------------------------------------------------
\369\ The cost of raw material also has a meaningful influence
on the future cost of the battery pack. As the production volume
goes up, the demand for battery critical raw materials also goes up,
which has an offsetting impact on the efficiency gains achieved
through economies of scale, improved plant efficiency, and advanced
battery cell chemistries, at least while supply is readjusting to
demand. We do not consider future battery raw material price
fluctuations for this analysis, however that may be an area for
further exploration in future analyses.
\370\ See, e.g., Jacky Wong, EV Batteries: The Next Victim of
High Commodity Prices?, The Wall Street Journal (July 22, 2021),
https://www.wsj.com/articles/ev-batteries-the-next-victim-of-high-commodity-prices-11626950276.
---------------------------------------------------------------------------
In the NPRM, we compared our battery pack costs in future years to
battery pack costs from a non-exhaustive list of other sources that may
or may not account for some of these additional parameters, including
varying potential future battery chemistry and learning rates. As
discussed in TSD Chapter 3.3.5.1.4, our battery pack costs in 2025 and
2030 fell fairly well in the middle of other sources' cost projections,
with Bloomberg New Energy Finance (BNEF) projections presenting the
highest year-over-year cost reductions, and one scenario in MIT's
Insights into Future Mobility report providing an upper bound of
potential future costs of the studies surveyed to create this
comparison.371 372 ICCT presented a similar comparison of
costs from several sources in its 2019 working paper and predicted
battery pack costs in 2025 and 2030 would drop to approximately $104/
kWh and $72/kWh, respectively, which put their projections slightly
higher than BNEF's 2019 projections.\373\ BNEF's 2020 Electric Vehicle
Outlook projected average pack cost to fall below $100/kWh by 2024,
while the 2021 NAS Report projected pack costs to reach $90-115/kWh by
2025.374 375 Since the NPRM, BNEF released its 2021 Electric
Vehicle Outlook, which estimated average pack prices in 2021 at $132/
kwh.\376\ In addition, Bloomberg weighed in on recent supply chain
impacts on battery materials availability, which is discussed in more
detail below.
---------------------------------------------------------------------------
\371\ See Logan Goldie-Scot, A Behind the Scenes Take on
Lithium-ion Battery Prices, Bloomberg New Energy Finance (March 5,
2019), https://about.bnef.com/blog/behind-scenes-take-lithium-ion-battery-prices/.
\372\ MIT Energy Initiative. 2019. Insights into Future
Mobility. Cambridge, MA: MIT Energy Initiative. Available at http://energy.mit.edu/insightsintofuturemobility.
\373\ Nic Lutsey and Michael Nicholas, ``Update on electric
vehicle costs in the United States through 2030'', ICCT (April 2,
2019), available at https://theicct.org/publications/update-US-2030-electric-vehicle-cost.
\374\ Bloomberg New Energy Finance (BNEF), ``Electric Vehicle
Outlook 2020,'' https://about.bnef.com/electric-vehicle-outlook/,
last accessed July 29, 2021.
\375\ 2021 NAS Report, at 114. The 2021 NAS Report assumed a 7
percent cost reduction per year from 2018 through 2030.
\376\ BloombergNEF. ``Battery Pack Prices Fall to an Average of
$132/kWh, But Rising Commodity Prices Start to Bite.'' November 30,
2021. https://about.bnef.com/blog/battery-pack-prices-fall-to-an-average-of-132-kwh-but-rising-commodity-prices-start-to-bite/#_ftn1.
(Last accessed: January 10, 2022)
---------------------------------------------------------------------------
We concluded in the NPRM that our projected costs seemed to fall
between several projections, giving confidence that the costs used in
the analysis could reasonably represent future battery pack costs
across the industry during the rulemaking time frame. We emphasized
that battery technology is currently under intensive development, and
that characteristics such as cost, and capability are rapidly changing.
These advances are reflected in recent aggressive projections, like
those from ICCT, BNEF, and the 2021 NAS Report.
We sought comment on several elements of the battery modeling
analysis in the NPRM, including on battery direct manufacturing costs,
or DMCs (and inputs and assumptions
[[Page 25818]]
used in BatPaC to estimate those costs), battery learning curves, and
other battery-related materials. More specifically, we first sought
comments on DMC assumptions, including comments supported by data
elements on different assumptions for battery chemistry, plant
manufacturing volume, or plant efficiency in MY 2020.\377\ To align
with our guiding principle that each technology model employed in the
analysis be representative of a wide range of specific technology
applications used in the industry, we requested that commenters explain
how these assumptions reasonably represent applications across the
industry in MY 2020.\378\ This is important to ensure that the CAFE
Model's simulation of manufacturer compliance pathways results in
impacts that we would reasonably expect to see in the real world. In
addition, we sought comment on the scaling used to generate direct
manufacturing costs for BEV400 and BEV500 technologies; in particular,
we were interested in any additional data or information on the
relationship between cost and weight for heavier battery packs used for
these higher-range BEV applications, particularly in light truck
vehicle segments.
---------------------------------------------------------------------------
\377\ Note that stakeholders had commented on the 2020 final
rule that batteries using NMC811 chemistry had either recently come
into the market or was imminently coming into the market, and
therefore DOT should have selected NMC811 as the appropriate
chemistry for modeling battery pack costs. Similar to the other
technologies considered in this analysis, DOT endeavors to use
technology that is a reasonable representation of what the industry
could achieve in the model year or years under consideration, in
this case the base DMC year of 2020, as discussed above. At the time
of this current analysis, the referenced A2Mac1 teardown reports and
other reports provided the best available information about the
range of battery chemistry actually employed in the industry. At the
time of writing for this final rule, DOT still has not found
examples of NMC811 in commercial application across the industry in
a way that DOT believes selecting NMC811 would have represented
industry average performance in MY 2020. As discussed in TSD Chapter
3.3.5.1.4, DOT did analyze the potential future cost of NMC811 in
the composite learning curve generated to ensure the battery
learning curve projections are reasonable.
\378\ Again, some vehicle manufacturer's systems may perform
better and cost less than our modeled systems and some may perform
worse and cost more. However, employing this approach will ensure
that, on balance, the analysis captures a reasonable level of costs
and benefits that would result from any manufacturer applying the
technology.
---------------------------------------------------------------------------
We also sought comment on the learning rates applied to battery
pack costs and on battery pack costs in future years. We recognized
that any battery pack cost projections for future years from our
analysis or external analyses will involve assumptions that may or may
not come to pass and stated that it would be most helpful if commenters
thoroughly explained the basis for any recommended learning rates,
including references to publicly available data or models (and if such
models are peer reviewed) where appropriate. We also noted that it
would be helpful for commenters to note where external analyses may or
may not take into account certain parameters in their battery pack cost
projections, and whether we should attempt to incorporate those
parameters in our analysis. For example, as discussed above, our
analysis does not consider long-term trends in raw material prices;
however, the price of raw materials may put a lower bound on NMC-based
battery prices.\379\
---------------------------------------------------------------------------
\379\ See, e.g., MIT Energy Initiative. 2019. Insights into
Future Mobility. Cambridge, MA: MIT Energy Initiative. Available at
http://energy.mit.edu/insightsintofuturemobility, at pp. 78-79.
---------------------------------------------------------------------------
We also stated that it would also be helpful if commenters
explained how learning rates or future cost projections could represent
the state of battery technology across the industry. Like other
technologies considered in this analysis, some battery and vehicle
manufacturers have more experience manufacturing electric vehicle
battery packs, and some have less, meaning that different manufacturers
will be at different places along the learning curve in future years.
We also stated that comments should specify whether their referenced
costs, either for MY 2020 or for future years, are for the battery cell
or the battery pack. We requested the information to ensure our
learning rates encompass these diverse parameters and to ensure that
the analysis best predicts the costs and benefits associated with
standards.
Tesla commented that the battery pack costs we projected in the
SAFE rule were too high, citing lower estimates published in the UBS-
sponsored Volkswagen ID 3 teardown report, among other studies.\380\
Tesla also commented that we unnecessarily constrained the analysis by
assuming that the drivetrain and other components are unique to each
vehicle and not shared by another vehicle.\381\
---------------------------------------------------------------------------
\380\ Tesla, at p. 9; DNV-GL, Tesla's Battery Day and the Energy
Transition (Oct. 26, 2020); BNEF, Electric Vehicle Outlook 2021
(June 9, 2021).; BNEF, Hitting the Inflection Point: Electric
Vehicle Price Parity and Phasing Out Combustion Vehicle Sales in
Europe (May 5, 2021); 2021 NAS Report; UBS, EVs Shifting into
Overdrive: VW ID.3 teardown--How will electric cars re-shape the
auto industry? (March 2, 2021).
\381\ Tesla, at p. 10.
---------------------------------------------------------------------------
To be clear, the battery pack DMCs used in our 2021 proposal and
this final rule are different than the battery pack DMCs used in the
SAFE rule that Tesla refers to in their comments. While our battery
pack DMCs have decreased since the 2020 final rule, our projected costs
are still higher than the sources that Tesla identifies. In the NPRM,
we provided a detailed explanation of how we developed those costs
using the BatPaC model and the specific inputs and assumptions used to
do so. We explained that we also expected those costs to represent the
range of costs across the industry. We acknowledged that each
manufacturer has different strategies associated with each vehicle line
based on several factors such as performance, costs, technology class,
utility among others, and this affects manufacturers strategy on
sourcing only certain components of battery pack or the complete
battery pack. We acknowledge that the cost of the battery pack as
measured in $/kWh can vary for each manufacturer with different form,
fit, and function requirements.\382\ BatPaC's inputs and assumptions,
including those developed specifically to support this rule,\383\ are
based on various and extended teardown reports available to the public
for predominant batteries that use robust and safe battery
chemistries.\384\ We understand that some mass market and premium
luxury BEVs have already achieved $/kWh values that are lower than our
projected costs, however others have not. To investigate the
sensitivity of our analysis to this cost we performed additional
analyses considering a 20 percent reduction in battery direct
manufacturing costs. And as discussed further below, this additional
cost reduction had a minimal impact on the overall vehicle cost and
increased electrification technology penetration. Therefore, we believe
the cost estimates from the BatPaC model represent a reasonable average
across all manufacturers for all vehicle technology classes.
---------------------------------------------------------------------------
\382\ Form, fit, and function is the identification and
description of characteristics of a part or assembly. Each defines a
specific aspect of the part to help engineers match parts to needs.
\383\ See Autonomie Model Documentation.
\384\ Ahmed, S., Nelson, P., Kubal, J., Liu, Z., Knehr, K. Dees,
D., ``Estimated cost of EV Batteries.'' Argonne. August 12, 2021.
https://www.anl.gov/cse/batpac-model-software. Last accessed January
20, 2022.
---------------------------------------------------------------------------
In contrast, the Auto Innovators submitted extensive comments on
our assumptions that the costs of battery electric vehicles will
continue to decline because of decreases in costs to produce battery
packs and other non-battery electrification components.\385\ Auto
Innovators stated that ``the traditional method of accounting for
possible future changes in battery-pack
[[Page 25819]]
costs is to apply a learning curve in future years based on production
volume, and then make a somewhat arbitrary assumption about when the
rate of decline decelerates or stops (technological maturity).'' Auto
Innovators identified that we characterized our learning curve as a
proxy for changes in battery chemistry, changes in energy density,
further gains in plant efficiency, and additional economies of scale in
production due to higher production volumes, but stated that we and NAS
do not ``confront the real possibility that counteracting, unanalyzed
factors could work to restrain the future decline in battery-pack
costs.'' \386\
---------------------------------------------------------------------------
\385\ Auto Innovators, at pp. 94-121.
\386\ Id., at pp. 94-95.
---------------------------------------------------------------------------
Auto Innovators and also the Alliance for Vehicle Efficiency (AVE)
requested that we consider potential impacts to battery raw materials
costs in the analysis.\387\ Auto Innovators provided a lengthy
qualitative survey of the state of raw materials extraction issues,
including their perspective on political and environmental obstacles to
further supply development. Auto Innovators also provided estimates of
battery materials costs that assumed a doubling of raw materials prices
and stated that ``a pre-2032 doubling of raw material prices could
substantially erode the `learning-curve' cost reductions assumed in the
RIAs.'' Auto Innovators stated that the battery sensitivity cases
presented in the PRIA are not large enough to account for simultaneous
increases in several raw materials prices, and that ``there is no basis
for believing that raw material prices will decline for a sustained
period prior to 2032.'' Accordingly, Auto Innovators stated that much
more careful analysis of raw material prices is necessary in the final
RIAs.
---------------------------------------------------------------------------
\387\ AVE, NHTSA-2021-0053-1488, at pp. 6-7.
---------------------------------------------------------------------------
With respect to analytical tools available to perform such an
analysis, Auto Innovators stated that ``less than a handful of the
dozens of published battery-forecasting models include any formal
analysis of global trends in raw material prices'' and stated that
``none of the published battery-forecasting models have accounted for
the surge in material price experienced in 2021.'' \388\ Auto
Innovators stated that ``BatPaC does not include a formal global model
of the market for each raw material used in battery packs,'' and
instead provides a best estimate of raw materials prices at the time of
version release.\389\ Auto Innovators stated that the version of BatPaC
we used did not account for the 2021 surge in raw material prices. Auto
Innovators stated that the MIT's Insights into Future Mobility report
took an important step to forecasting battery pack costs by using a
two-stage model, one for the cost of materials and the second for the
costs to manufacture the battery pack.\390\ However, Auto Innovators
stated that we erroneously characterized MIT's estimate as an ``upper
bound'' of battery pack costs, while the report actually provides best
estimates based on different scenarios.
---------------------------------------------------------------------------
\388\ Auto Innovators, at pp. 97-98.
\389\ Id., at pp. 119-121.
\390\ Insights into Future Mobility, MIT Energy Initiative
(2019), Cambridge, MA: MIT Energy Initiative, https://energy.mit.edu/research/mobilityofthefuture/ at p. 76. Accessed
January 19, 2022.
---------------------------------------------------------------------------
Auto Innovators made three explicit requests in regards to future
battery materials costs and chemistry impacts; first, Auto Innovators
stated that we should work with National Laboratories, DOE, and others
to produce sensitivity cases for raw and processed material costs,
material efficiency in battery construction, and other considerations;
next, Auto Innovators stated that we should remove changes in battery
chemistry from the near-term learning factor and analyze it separately
and explicitly in our RIA; and finally, Auto Innovators stated that
``instead of choosing one battery chemistry as representative of the
entire industry, as the [a]gencies do with the Argonne battery model,
the [a]gencies should forecast the penetration of different battery
chemistries in the fleet from 2021 to 2032 and estimate applicable
costs for each of them.''
As a reminder, the learning rate that we used in the NPRM and this
final rule, carried forward from work done for the 2018 NPRM, is based
on an assessment of cost reductions due to production volume increases.
As we described in the TSD, we identified the change in cost for the
estimated changes in production volumes linked to model years and used
this rate to develop the learning curves used out to MY 2032, which
resulted in an approximately 4.5 percent year over year cost reduction.
For MYs 2033 to 2050, we scaled down the learning rate in steps based
on literature values and market research.
The parametric analysis presented in the NPRM TSD was meant to
confirm that looking at any one potential factor that could have an
impact on the battery pack direct manufacturing costs would not have
significantly changed this original near-term (i.e., through MY 2032)
4.5 percent production-volume-based learning rate. The parametric
analysis showed that considering two factors by themselves--increasing
production volume and improving manufacturing plant efficiency--would
result in a slightly shallower learning curve (3.26 and 3.5 percent
near-term, year-over-year reductions in cost), while changing battery
chemistry by itself would result in a steeper learning curve (5.15
percent near-term, year-over-year cost reductions). Constructing a
composite learning curve to consider these three factors in tandem,
assuming that the predominant battery chemistry will change over the
course of this decade, and also that battery manufacturing plants will
become better at producing battery cells--two widely accepted
assumptions--confirmed that our original learning curve based on year-
over-year production volume increases could reasonably encompass these
changes.\391\ Furthermore, while Auto Innovators asserted that our
production-based learning curve could miss several important factors,
as discussed in Section III.C.6 above and in recent literature,\392\ a
production-volume-based learning curve is an accepted and reasonable
method for projecting future costs.
---------------------------------------------------------------------------
\391\ See, e.g., MIT Insights into Future Mobility Report, at 77
(``A clear trend within the EV LIB industry is to increase nickel
content to boost energy density (for increased driving range) while
reducing the amount of expensive cobalt required.'').
\392\ Lukas Mauler, Fabian Duffner, Wolfgang G Zeier, Jens
Leker, ``Battery Cost Forecasting: A Review of Methods and Results
with an Outlook to 2050,'' Energy and Environmental Science, 14
(2021) at p. 4724.
---------------------------------------------------------------------------
Regarding Auto Innovators' extensive comments about the impact of
materials availability on battery costs, we are aware that the outlook
for battery materials has remained uncertain since we released the
NPRM. At this time, studies and organizations have provided projections
about the impact of battery materials price increases due to supply
chain factors and the consensus seems to be that the overall impact on
prices will be minimal for the predominant battery chemistries.\393\
Our estimated future battery costs are fairly conservative compared to
leading analysis firms, even accounting for materials price impacts
since the
[[Page 25820]]
NPRM.394 395 This makes us confident that our projected
battery costs, presented in this final rule, still fall within the
scope of reasonable projections for the near-term model years covered
by this analysis.
---------------------------------------------------------------------------
\393\ Lukas Mauler, Fabian Duffner, Wolfgang G Zeier, Jens
Leker, ``Battery Cost Forecasting: A Review of Methods and Results
with an Outlook to 2050,'' Energy and Environmental Science, 14
(2021) at p. 4734 (``Every single study that provides time-based
projections expects LIB cost to fall, even if increasing raw and
battery material prices are taken into account.''); Henze, V.,
``Battery Pack Prices Fall to an Average of $132/kWh, But Rising
Commodity Prices Start to Bite''. BloombergNEF. November 30, 2021.
https://about.bnef.com/blog/battery-pack-prices-fall-to-an-average-of-132-kwh-but-rising-commodity-prices-start-to-bite/. Last accessed
January 23, 2022.
\394\ See NPRM TSD at 296, Table 3-86--Battery Cost Estimates
from Other Sources.
\395\ Henze, V., ``Battery Pack Prices Fall to an Average of
$132/kWh, But Rising Commodity Prices Start to Bite''. BloombergNEF.
November 30, 2021. https://about.bnef.com/blog/battery-pack-prices-fall-to-an-average-of-132-kwh-but-rising-commodity-prices-start-to-bite/. Last accessed January 23, 2022.
---------------------------------------------------------------------------
Nonetheless, we do appreciate Auto Innovators' data and analysis
submitted on raw materials cost impacts on battery pack costs. We also
appreciate the enormity of the task of integrating forecasts of global
trends in raw materials prices in our analysis, given that only a
minority of the dozens of published battery-forecasting models include
any formal analysis of global trends in raw materials prices and none
of the published forecasting models have accounted for the increase in
material price experienced in 2021. MIT's two-stage model, and
multidimensional mathematical models are more refined than single
dimensional models due to the use of numerous parameters. However, this
comes at the expense of needing to obtain high quality and accurate
data for these parameters, potentially at the cost of reduced
transparency. For example, MIT's two-stage model requires data from
mining companies, materials producers, cell producers, and battery pack
producers.\396\ However, detailed data on these specifics are not
readily publicly available.\397\ \398\ \399\
---------------------------------------------------------------------------
\396\ Insights into Future Mobility, MIT Energy Initiative
(2019), Cambridge, MA: MIT Energy Initiative, https://energy.mit.edu/research/mobilityofthefuture/ at p. 77. Accessed
January 19, 2022.
\397\ S. Matteson and E. Williams, Learning dependent subsidies
for lithium-ion electric vehicle batteries, Technol. Forecast. Soc.
Change, 2015, 92, 322-331.
\398\ B. Nykvist, F. Sprei and M. Nilsson, Assessing the
progress toward lower priced long range battery electric vehicles,
Energy Policy, 2019, 124, 144-155.
\399\ Lukas Mauler, Fabian Duffner, Wolfgang G Zeier, Jens
Leker, ``Battery Cost Forecasting: A Review of Methods and Results
with an Outlook to 2050,'' Energy and Environmental Science, 14
(2021) at p. 4715 (``However, details on company-specific prices,
costs and profit margins are not publicly available and differences
are difficult to assess.'').
---------------------------------------------------------------------------
Developing a multi-stage model that can perform the calculations we
need for the number of large-scale simulations required by our
analysis, with data and assumptions that are transparent and can be
made publicly available, would be a difficult task. As discussed above,
BatPaC is a publicly available model and the inputs and assumptions
used to develop and populate BatPaC are publicly available. More
specifically, we included detailed data from teardown reports that we
used to generate the battery pack inputs for this analysis in the TSD
and Argonne Model Documentation. The battery pack designs and cell
chemistry that we modeled in BatPaC represented the most common battery
pack parameters in the market in MY 2020, our base year for calculating
direct manufacturing costs. This approach reflects the same approach we
use across our analysis; we do not currently model, for example, the
penetration rate of Toyota's HCR engine separately from Mazda's HCR
engine. Again, modeling an industry-average system will ensure that, on
balance, the analysis captures a reasonable level of costs and benefits
that would result from any manufacturer applying the technology. In
addition, while Auto Innovators presents important points about the
uncertainty regarding the predominant battery chemistry beyond MY 2027,
the battery chemistries that we analyzed--NMC622 and NMC811--are still
expected to be the dominant chemistries in this rulemaking timeframe.
The sensitivity analyses presented in the TSD accompanying the NPRM and
this final rule show that analyzing both chemistries separately results
in only a small difference in cost between the two options. We see only
a small difference in costs because we consider a narrow range of
battery pack power and energy sizes in the respective vehicle
technology classes.
At this time, we believe that our battery pack costs in this final
rule still could reasonably represent costs to the industry during the
model years under consideration taking into account the factors
mentioned by Auto Innovators. In addition, as discussed further below,
our sensitivity cases show that BEV prices remain within a fairly
narrow range in the rulemaking timeframe considering potentially higher
direct manufacturing costs or shallower learning rates.
We will continue to investigate further refinements to input data
and models that we use to assess battery costs as the input data and
models continue to develop. We understand that battery technologies and
manufacturing processes are undergoing significant development and we
will continue to monitor and evaluate battery cost and performance, and
how to reflect those trends in our modeling.
For future actions, we would welcome any additional information on
the impact of raw materials prices on battery pack costs, including
information on a CBI or public basis on the impact of long-term supply
contracts on battery costs.\400\ In particular, we would be interested
in more information on whether manufacturers that had contracted for
battery packs prior to the 2021 materials supply chain disruptions were
insulated from materials cost increases and if there is a contractual
or other mechanism within the vehicle manufacturer's control through
which vehicle manufacturers could insulate themselves from such
disruptions moving forward.\401\
---------------------------------------------------------------------------
\400\ C. Xu, et al., Future material demand for automotive
lithium-based batteries, Commun. Mater., 2020, 1, 99.; H. Hao, et
al., Impact of transport electrification on critical metal
sustainability with a focus on the heavy-duty segment, Nat. Commun.,
2019, 10, 5398.; Reuters. ``Stellantis, LG Energy Solution to form
battery JV for North America.'' Automotive News. October 18, 2021.
https://www.autonews.com/manufacturing/stellantis-lg-energy-solution-form-battery-jv-north-america. Last accessed 01/20/2022.;
``Daimler, Stellantis enter agreement with battery maker Factorial
Energy.'' Automotive News. November 30, 2021. https://www.autonews.com/suppliers/why-daimler-stellantis-are-investing-battery-maker. Last accessed January 20, 2022.; ``FORD COMMITS TO
MANUFACTURING BATTERIES, TO FORM NEW JOINT VENTURE WITH SK
INNOVATION TO SCALE NA BATTERY DELIVERIES.' Ford Media Center. May
20, 2021. https://media.ford.com/content/fordmedia/fna/us/en/news/2021/05/20/ford-commits-to-manufacturing-batteries.html. Last
accessed January 20, 2022.; ``Toyota Selects North Carolina for New
U.S. Automotive Battery Plant.'' Toyota Newsroom. December 7, 2021.
https://global.toyota/en/newsroom/corporate/36418723.html. Last
accessed January 20, 2022.
\401\ See, e.g., Lukas Mauler, Fabian Duffner, Wolfgang G Zeier,
Jens Leker, ``Battery Cost Forecasting: A Review of Methods and
Results with an Outlook to 2050,'' Energy and Environmental Science,
14 (2021) at p. 4724; (``In the battery industry-prices are further
influenced by strategic pricing, long-term contracts and rebates to
utilize excess production capacity.'').
---------------------------------------------------------------------------
As in any large-scale analysis, uncertainties exist. Recognizing
that there could be additional factors that constrain battery learning
rates, as Auto Innovators suggests, we performed four sensitivity
studies around battery pack costs that are described in FRIA Chapter
7.2.2.3. The sensitivity studies examined the impacts of increasing and
decreasing the direct cost of batteries and battery learning costs by
20 percent from central analysis levels, based on our survey of
external analyses' battery pack cost projections that fell generally
within 20 percent of our central analysis costs. The
average difference in vehicle cost between the reference case and four
battery sensitivity cases ranged from -$52 to $128. This means that,
even accounting for potential unanalyzed factors related to battery
prices, we expect battery electric vehicle prices to remain within a
fairly narrow range in the rulemaking timeframe. These sensitivity
outcomes are similar
[[Page 25821]]
to those we showed in the NPRM sensitivity analysis. Although Auto
Innovators showed how an increase in individual raw material cost could
impact the final cost, we believe that at the total pack level the 20
percent high sensitivity case encompasses these situations in the
rulemaking time frame. Again, these results, in addition to the
consensus in literature regarding the impact of rising materials prices
on future costs described above, make us comfortable that our approach
to estimating battery costs is a reasonable approach for this final
rule analysis.
After pointing out the BatPaC model's limitations regarding future
potential increases in materials costs, Auto Innovators commented that
we should use BatPaC to estimate battery pack costs for BEV400 and
BEV500 technologies instead of scaling up BEV300 battery pack
costs.\402\ Beyond the request to do so, we received no updated real-
world data on the cost and weight of battery packs used in 400- and
500-mile range electric vehicles. As discussed above, and as originally
stated in the NPRM, initial values from BatPaC could not be validated
by real-world data, leading us to continue using the scaled values for
the final rule.
---------------------------------------------------------------------------
\402\ Auto Innovators, at p. 119.
---------------------------------------------------------------------------
Auto Innovators identified other costs related to electric vehicles
(EVs) that they stated our analysis does not consider; specifically,
they stated that our battery-price estimates are industry averages that
do not exclude supply chains that fail environmental, social, and
governance (ESG) tests. Auto Innovators stated that ``for the major
global automakers that operate in the [U.S.] auto market, the RIAs
should not assume that low-cost suppliers with poor ESG profiles can be
utilized in EV supply chains.'' Auto Innovators also identified the
shift from recycling engines and transmissions to recycling EV
batteries, as well as the price of electricity to produce EV batteries,
as costs that we do not currently account for. In addition, Auto
Innovators stated that the BEVs and PHEVs are a new technology type for
many drivers and, as a result, drivers may incur some costs and
inconveniences that we should consider as part of our analysis.\403\
They provided three examples of costs to the user beyond the purchase
price: (1) Costs of charging stations for BEVs and PHEVs; (2) costs to
the user of a vehicle that has a shorter driving range than the typical
conventional IC engine and that requires a long time to charge, and (3)
the time spent charging.
---------------------------------------------------------------------------
\403\ Id., at pp. 119-121.
---------------------------------------------------------------------------
We applaud Auto Innovators members for including serious ESG
considerations in their planning for developing battery supply chains.
However, like the issues surrounding raw materials impacts discussed
above, we currently do not have a specific mechanism to account for the
cost of supply chains that pass basic ESG tests, as Auto Innovators
suggests. To the extent that Auto Innovators members have already
entered into contracts with battery suppliers and have included ESG
terms in those contracts, and have data or other information on how
that increases the costs for EV production over and above an industry
average that we would project quantitatively, we welcome that
information for future analysis. We will continue to research these
factors and consider whether to include them in the cost-benefit
analysis. We support Auto Innovators and any individual component or
vehicle manufacturer providing the agency with supporting material for
these specific topics.
As a reminder, our analysis considers technology costs that vehicle
manufacturers ultimately pass to the buyer separately from the user
costs for a technology, like fueling from either gasoline or
electricity. We consider many externalities that accrue cost for the
consumer in the analysis, and these are discussed in Section III.E. We
specifically identified a cost to the user for time spent charging an
EV, which is discussed further in that section. However, regardless of
where we account for those costs in the analysis, we believe those
costs would be minimal in the timeframe of this rulemaking considering
the standard-setting projections of EV and PHEV penetration rates,
which are discussed further in FRIA Chapter 6.3.1. That said, for
future rules we appreciate any new data Auto Innovators and other
stakeholders can provide to develop more precise electric vehicle user
costs.
Next, ICCT commented that we ``erroneously inflated battery costs
by applying the retail price equivalent (RPE) markup to base costs that
already include indirect costs.'' \404\ We disagree. The indirect costs
represented in BatPaC output are those that apply to the battery
supplier, and do not represent the indirect costs experienced by the
OEM who purchases the battery and integrates it into the vehicle. NHTSA
has always considered RPE markup to be applicable to purchased items.
---------------------------------------------------------------------------
\404\ ICCT, at p. 8.
---------------------------------------------------------------------------
We also believe that the warranty costs are appropriately marked up
with the BatPaC outputs. The RPE markup factor is based on an
examination of historical financial data contained in 10-K reports
filed by manufacturers with the Securities and Exchange Commission. It
represents the ratio between the retail price of motor vehicles and the
direct costs of all activities that manufacturers engage in, including
the design, development, manufacturing, assembly, and sales of new
vehicles, refreshed vehicle designs, and modifications to meet safety
or fuel economy standards. An RPE of 1.5 does not imply that
manufacturers automatically mark up each vehicle by exactly 50 percent.
Rather, it means that, over time, the competitive marketplace has
resulted in pricing structures that average out to this relationship
across the entire industry. Prices for any individual model may be
marked up at a higher or lower rate depending on market demand. The
consumer who buys a popular vehicle may, in effect, subsidize the
installation of a new technology in a less marketable vehicle. But, on
average, over time and across the vehicle fleet, the retail price paid
by consumers has risen by about $1.50 for each dollar of direct costs
incurred by manufacturer.
The direct costs associated with any specific technology will
change over time as some combination of learning and resource price
changes occurs. Resource costs, such as the price of steel, can
fluctuate over time and can experience real long-term trends in either
direction, depending on supply and demand. However, the normal learning
process generally reduces direct production costs as manufacturers
refine production techniques and seek out less costly parts and
materials for increasing production volumes. By contrast, this learning
process does not generally influence indirect costs. To be consistent
with the basis for the RPE multiplier, we apply learning to direct
costs, and then mark up the resulting learned direct costs using the
RPE multiplier.
We consulted Argonne and the BatPaC manual and as shown in the
BatPaC documentation, the final cost provided by the BatPaC model
includes two-part variable costs (what we consider direct costs) and
fixed expenses (what we consider indirect costs). Table 8.7 in the
BatPaC Model Documentation shows the breakdown of the costs and the
approximate percentage of each cost.
These costs combine to provide the overall cost of the battery pack
from the supplier to the OEM. The cost of the battery pack from the
supplier to the OEM is considered a direct cost to the OEM, like any
other part that an OEM
[[Page 25822]]
acquires from other suppliers. In turn, while using the battery pack in
the finished vehicle, the OEM will incur indirect costs including
research and development (R&D), general sales and administrative costs
(GSA), as well as warranty and profit. Thus, the indirect costs
associated with components or subsystems incurred by the automotive
suppliers should not be conflated with vehicle manufacturer indirect
costs.
Supplier warranty costs should reflect losses they experience to
replace defective battery packs or parts. Likewise, OEM warranty costs
should reflect actual losses they incur in replacing defective parts.
OEM losses are partially reimbursed by supplier warranties. Both OEM
warranty costs and supplier warranty costs should thus represent the
net loss to each business due to warranty coverage. OEM warranty costs
should thus already reflect reimbursement to OEMs from supplier
warranties, implying that reflecting warranty costs within the direct
cost of the product and separate warranty costs at the OEM level is not
double counting. Accordingly, we did not make any changes to how
indirect cost markups are applied to the BatPaC costs for this final
rule.
In sum, after considering the comments received on how we modeled
battery pack costs, we determined that it was appropriate to use the
same battery costs for this final rule. We will perform additional
research and update our analysis accordingly for future analyses.
Turning to electrification costs that are non-battery related, each
vehicle powertrain type receives different non-battery electrification
components. When researching costs for different non-battery
electrification components, we found that different reports vary in
components considered and cost breakdown. This is not surprising, as
vehicle manufacturers use different non-battery electrification
components in different vehicle's systems, or even in the same vehicle
type, depending on the application.\405\ We use costs for the major
non-battery electrification components on a dollar per kilowatt basis
based on the costs presented in two reports. We use a $/kW cost metric
for non-battery components to align with the normalized costs for a
system's peak power rating as presented in U.S. DRIVE's Electrical and
Electronics Technical Team (EETT) Roadmap report.\406\ This approach
captures components in some manufacturer's systems, but not all
systems; however, we believe this is a reasonable metric and approach
to use for this analysis given the differences and complexities in non-
battery electrification systems. This approach allows us to scale the
cost of non-battery electrification components based on the
requirements of the system to meet vehicle utility and performance
requirements. We also rely on a MY 2016 Chevrolet Bolt teardown study
for some categories of strong hybrid component costs and all other PHEV
and BEV non-battery component costs that were not explicitly estimated
in the EETT Roadmap report.\407\
---------------------------------------------------------------------------
\405\ For example, the MY 2020 Nissan Leaf does not have an
active cooling system whereas Chevy Bolt uses an active cooling
system.
\406\ U.S. DRIVE, Electrical and Electronics Technical Team
Roadmap (Oct. 2017), available at https://www.energy.gov/sites/prod/files/2017/11/f39/EETT%20Roadmap%2010-27-17.pdf.
\407\ Hummel et al., UBS Evidence Lab Electric Car Teardown--
Disruption Ahead?, UBS (May 18, 2017), https://neo.ubs.com/shared/d1wkuDlEbYPjF/ (accessed: Feb. 11, 2022).
---------------------------------------------------------------------------
We received several comments specific to strong hybrid non-battery
electrification technology costs, in particular regarding the costs of
eCVTs and high voltage cables.
Tesla stated that it believes that non-battery electrification
components that add to the total cost required to electrify a vehicle
continue to decrease in price and are utilized across vehicle types and
EVs are rapidly approaching price parity with ICE technology.\408\
---------------------------------------------------------------------------
\408\ Tesla, at pp. 9-10.
---------------------------------------------------------------------------
American Council for an Energy-Efficient Economy (ACEEE) commented
that the cost to manufacture hybrid vehicles has fallen significantly
in recent years, more so than NHTSA's analysis assumes.\409\ They
stated that the incremental hybridization costs used in this rule are
significantly higher than those assessed by the 2021 NAS Report.
Specifically, they stated that when accounting for differing
assumptions, the costs assumed by this rule are 20 percent higher.
---------------------------------------------------------------------------
\409\ ACEEE, Docket No. NHTSA-2021-0053-0074, at p. 5.
---------------------------------------------------------------------------
Toyota commented that ``NHTSA's estimated costs are significantly
higher than Toyota's understanding based on our current products and
experience developing and marketing hybrids systems over the last two
decades. The estimated costs for power split hybrids used as an input
to compliance modeling for the proposed standards are more than twice
the cost estimates in the National Academies of Science Engineering and
Medicine (NASEM) 2025-2035 CAFE Study.'' \410\ They added ``NHTSA's
projected power split system costs are always significantly higher than
P2 system costs for the same vehicle class. Toyota's experience is that
the relative cost of the power split and P2 systems depends on vehicle
class and operational requirements, and that for many applications
power split and P2 system costs are much more similar than NHTSA's
estimates suggest.'' They further added ``Once adjusted for future cost
savings, NHTSA's 2020 hybrid costs are still typically double the NASEM
estimates. Further, the NASEM committee estimates the incremental cost
of midsize and crossover strong hybrids in 2020 model year to be $2,000
to 3,000 more than a conventional vehicle which is well below NHTSA's
2020 power split estimate,'' and ``Toyota believes the NASEM 2025 model
year cost values are more representative of hybrid vehicle costs
through the 2026 model year, including any accompanying engine
developments and normalization for differences in component sizes and
assessment methodologies. We disagree that engine upgrades should
account for a large portion of the difference between the NASEM and
NHTSA cost estimates. Such a significant cost difference does not exist
for Toyota's 2.5L Dynamic Force engine used in the hybrid and non-
hybrid versions of the 2021 model year Camry referenced by NHTSA.''
---------------------------------------------------------------------------
\410\ Toyota, at pp. 7-8.
---------------------------------------------------------------------------
ICCT also commented on cost estimates for the power-split hybrid,
stating that ``NHTSA has substantially overestimated the costs of full
hybrid vehicles, as eCVT costs are far lower than the CVTL2 costs
assumed by NHTSA; NHTSA's high-voltage cable cost is more than twice
that of both NAS and FEV; NHTSA's battery size and cost are overstated,
as they do not take into account power density improvements that cut
the size and cost of strong hybrid battery packs in half; and NHTSA's
analysis has $432 for power electronics and thermal management that
appear to be already be included in motor/inverter/generator/regen
brake costs for NAS and FEV.'' \411\
---------------------------------------------------------------------------
\411\ ICCT, at p. 10.
---------------------------------------------------------------------------
We agree with Tesla that there are many non-battery components that
are shared across different vehicle lines, and this provides an
opportunity for cost reductions over time from economies of scale. We
capture cost reductions for non-battery electrification components
through a learning curve Section III.C.6. We will continue to monitor
trends and other information related to non-battery components.
Based on the comments specific to hybrid vehicle non-battery
component costs, as well as data from the 2021 NAS Report, we
reexamined the costs for
[[Page 25823]]
non-battery components. For this final rule, we updated the cost of an
eCVT for SHEVPS vehicles, as well as the costs of high voltage cables
for all strong hybrid vehicles.
Previously, we had used the cost of a CVTL2 as a proxy for the
eCVT; for this final rule, the eCVT cost comes from data in the EPA-
sponsored teardown study of a 2011 Ford Fusion strong hybrid,\412\ and
has been adjusted to 2018$. This cost also aligns with the eCVT cost
presented in the 2021 NAS Report.
---------------------------------------------------------------------------
\412\ EPA. ``Light Duty Technology Cost Analysis, Power-Split
and P2 HEV Case Studies.' November 2011. EPA-420-R-11-015. https://nepis.epa.gov/Exe/ZyPDF.cgi/P100EG1R.PDF?Dockey=P100EG1R.PDF.
(Accessed: Dec. 3, 2021)
---------------------------------------------------------------------------
We also used data from the 2011 Ford Fusion teardown study to
adjust the cost of SHEVP2 and SHEVPS high voltage cables. This
adjustment brought our high voltage cable costs in closer proximity to
the 2021 NAS Report high voltage cable costs. More details about the
updated costs can be found in TSD Chapter 3.3.5.3. The resulting cost
differences between the SHEVP2 and SHEVPS hybrid systems is mainly
associated with the fact that our analysis considers two motors/
generators for SHEVPS and one motor/generator for SHEVP2. We discuss
how SHEVPS and SHEVP2 are characterized in our analysis in Section
III.D.3.a).
As a reminder, the assumptions that we use to model and simulate
strong hybrid vehicles in Autonomie are not specific to any one
manufacturer's vehicle type. The engines and/or electric motors are
sized to meet different characteristics like utility, performance, and
other key designs to provide the highest system efficiency. These key
characteristics and attributes are discussed in detail in Section
III.C.4. This results in costs that may not match one specific vehicle
teardown. However, we still believe that on average the system cost
estimates are appropriate.
We agree with Toyota that in some cases a vehicle's engine does not
change when going from a conventional powertrain to hybrid powertrain,
like Toyota's example of the 2.5L naturally aspirated engine in the
RAV4 and RAV4 hybrid. However, the analysis fleet consists of vehicles
with an assortment of engines that are as basic as VVT-only to as
advanced as VCR. In some cases, a vehicle that starts with a basic
conventional engine that adopts SHEVP2 system could also adopt a more
advanced engine. For example, the 2022 Hyundai Tucson base engine is a
2.5L naturally aspirated engine and its hybrid version engine is a
downsized turbocharged engine.\413\ We allow the CAFE Model to both
upgrade and downgrade the engine associated with SHEVP2 powertrains to
apply the ICE engine that is most cost effective with the hybrid
system. The details of these scenarios discussed further in Sections
III.D.3.a) and III.D.3.c) for SHEVs.
---------------------------------------------------------------------------
\413\ Lorio, J., ''Tested: 2022 Hyundai Tucson Hybrid Aids
Mileage and Performance.'' Car and Driver. December 22, 2021.
https://www.caranddriver.com/reviews/a38591574/2022-hyundai-tucson-hybrid-by-the-numbers/. (Accessed: Dec. 29, 2021)
---------------------------------------------------------------------------
Finally, we use Autonomie and BatPaC to model the size and cost of
batteries used in strong hybrid vehicles. More details on the sizing
algorithm and battery costs can be found in the Argonne model
documentation as well as in TSD Chapter 3.3.5.1.
We received another comment from ICCT stating that ``for 2018 Mid
Term Evaluation, non-battery BEV and PHEV costs were updated based on
more recent teardown data from California Air Resources Board, UBS, and
other references, but these updated costs were not used in the proposed
NHTSA rule.'' \414\
---------------------------------------------------------------------------
\414\ ICCT, at pp. 7-8.
---------------------------------------------------------------------------
Although ICCT references multiple studies in their comment, they do
not provide any specific BEV and PHEV component costs that they believe
are estimated incorrectly in our analysis. As discussed earlier and in
TSD Chapter 3.3.5.2, we have used the most recent public data available
to estimate the cost of non-battery electrification components. In
particular, we rely on the UBS teardown study that ICCT references for
some BEV and PHEV components.
To develop the learning curves for non-battery electrification
components, we used cost information from Argonne's 2016 Assessment of
Vehicle Sizing, Energy Consumption, and Cost through Large-Scale
Simulation of Advanced Vehicle Technologies report.\415\ The report
provided estimated cost projections from the 2010 lab year to the 2045
lab year for individual vehicle components.416 417 We
considered the component costs used in electrified vehicles, and
determined the learning curve by evaluating the year over year cost
change for those components. Argonne published a 2020 version of the
same report that included high and low-cost estimates for many of the
same components, that also included a learning rate.\418\ Our learning
estimates generated using the 2016 report fall fairly well in the
middle of these two ranges, and therefore we decided that continuing to
apply the learning curve estimates based on the 2016 report was
reasonable. There are many sources that we could have picked to develop
learning curves for non-battery electrification component costs,
however given the uncertainty surrounding the complexity of the systems
and extrapolating costs out to MY 2050, we believe these learning
curves provide a reasonable estimate.
---------------------------------------------------------------------------
\415\ Moawad, Ayman, Kim, Namdoo, Shidore, Neeraj, and Rousseau,
Aymeric. Assessment of Vehicle Sizing, Energy Consumption and Cost
Through Large Scale Simulation of Advanced Vehicle Technologies
(ANL/ESD-15/28). United States (2016). Available at https://www.autonomie.net/pdfs/Report%20ANL%20ESD-1528%20-%20Assessment%20of%20Vehicle%20Sizing,%20Energy%20Consumption%20and%20Cost%20through%20Large%20Scale%20Simulation%20of%20Advanced%20Vehicle%20Technologies%20-%201603.pdf, (accessed: Feb. 11, 2022).
\416\ ANL/ESD-15/28, at p. 116.
\417\ DOE's lab year equates to five years after a model year,
e.g., DOE's 2010 lab year equates to MY 2015.
\418\ Islam, E., Kim, N., Moawad, A., Rousseau, A. ``Energy
Consumption and Cost Reduction of Future Light-Duty Vehicles through
Advanced Vehicle Technologies: A Modeling Simulation Study Through
2050'', Report to the U.S. Department of Energy, Contract ANL/ESD-
19/10, June 2020 https://www.autonomie.net/pdfs/ANL%20-%20Islam%20-%202020%20-%20Energy%20Consumption%20and%20Cost%20Reduction%20of%20Future%20Light-Duty%20Vehicles%20through%20Advanced%20Vehicle%20Technologies%20A%20Modeling%20Simulation%20Study%20Through%202050.pdf, (accessed: Feb.
11, 2022).
---------------------------------------------------------------------------
Table III-19 shows an example of how the non-battery
electrification component costs are computed for the Medium Car and
Medium SUV non-performance vehicle classes for the final rule analysis.
BILLING CODE 4910-59-P
[[Page 25824]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.083
[[Page 25825]]
TSD Chapter 3.3.5.2 contains more information about the non-battery
electrification components relevant to each specific electrification
technology and the sources used to develop these costs.
Finally, the cost of electrifying a vehicle depends on the other
powertrain components that must be added or removed from a vehicle with
the addition of the electrification technology. Table III-20 below
provides a breakdown of each electrification component included for
each electrification technology type, as well as where to find the
costs in each CAFE Model input file.
[GRAPHIC] [TIFF OMITTED] TR02MY22.084
The following example in Table III-21 shows how the costs are
computed for a vehicle that progresses from a lower level to a higher
level of electrified powertrain. The table shows the components that
are removed and the components that are added as a GMC Acadia
progresses from a MY 2024 vehicle with only SS12V electrification
technology to a BEV300 in MY 2025.\420\ The total cost in MY 2025 is a
net cost addition to the vehicle. The same methodology could be used
for any other technology advancement in the electric technology tree
path. For the final rule analysis, the cost of the SS12V battery was
updated as discussed earlier, and this example has been updated to show
the new cost.
---------------------------------------------------------------------------
\419\ As discussed in section 3.3.5.3 of the TSD, we no longer
use the BatPaC SS12V battery cost and use a cheaper AGM battery
instead, and the updated cost is reflected in the battery_costs.csv
file.
\420\ Vehicle code 11001008 in the Vehicle Report output file.
---------------------------------------------------------------------------
[[Page 25826]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.085
TSD Chapter 3.3.5.3 includes more details about how the costs
associated with the internal combustion engine, transmission, electric
machine(s), non-battery electrification components, and battery pack
for each electrified technology type are combined to create a full
electrification system cost.
---------------------------------------------------------------------------
\421\ Please note that in this calculation the CAFE Model
accounts for the air conditioning and off-cycle technologies (g/
mile) applied to each vehicle model. The cost for the AC/OC
adjustments are located in the CAFE Model Scenarios file. The air
conditioning and off-cycle cost values are discussed further in TSD
Chapter 3.8.
---------------------------------------------------------------------------
Mass Reduction
Mass reduction is a relatively cost-effective means of improving
fuel economy, and vehicle manufacturers are expected to apply various
mass reduction technologies to meet fuel economy standards. Reducing
vehicle mass is accomplished through several different techniques, such
as modifying and optimizing vehicle component and system designs, part
consolidation, and adopting lighter weight materials (advanced high
strength steel, aluminum, magnesium, and plastics including carbon
fiber reinforced plastics).
The cost for mass reduction depends on the type and amount of
materials used, the manufacturing and assembly processes required, and
the degree to which changes to plants and new manufacturing and
assembly equipment is needed. In addition, manufacturers may develop
expertise and invest in certain mass reduction strategies that may
affect the approaches for mass reduction they consider and the
associated costs. Manufacturers may also consider vehicle attributes
like noise-vibration-harshness (NVH), ride quality, handling, crash
safety and various acceleration metrics when considering how to
implement any mass reduction strategy. These are considered to be
aspects of performance, and for this analysis any identified pathways
to compliance are intended to maintain performance neutrality.
Therefore, mass reduction via elimination of, for example, luxury items
such as climate control, or interior vanity mirrors, leather padding,
etc., is not considered in the mass reduction pathways for this
analysis.
The automotive industry uses different metrics to measure vehicle
weight. Some commonly used measurements are vehicle curb weight,\422\
gross vehicle weight (GVW),\423\ gross vehicle weight rating
(GVWR),\424\ gross combined weight (GCVW),\425\ and equivalent test
weight (ETW),\426\ among others. The vehicle curb weight is the most
commonly used
[[Page 25827]]
measurement when comparing vehicles. A vehicle's curb weight is the
weight of the vehicle including fluids, but without a driver,
passengers, and cargo. A vehicle's glider weight, which is vehicle curb
weight minus the powertrain weight, is used to track the potential
opportunities for weight reduction not including the powertrain. A
glider's subsystems may consist of the vehicle body, chassis, interior,
steering, electrical accessory, brake, and wheels systems. The
percentage of weight assigned to the glider will remain constant for
any given rule but may change overall. For example, as electric
powertrains including motors, batteries, inverters, etc. become a
greater percent of the fleet, glider weight percentage will change
compared to earlier fleets with higher dominance of ICE powertrains.
---------------------------------------------------------------------------
\422\ This is the weight of the vehicle with all fluids and
components but without the drivers, passengers, and cargo.
\423\ This weight includes all cargo, extra added equipment, and
passengers aboard.
\424\ This is the maximum total weight of the vehicle,
passengers, and cargo to avoid damaging the vehicle or compromising
safety.
\425\ This weight includes the vehicle and a trailer attached to
the vehicle, if used.
\426\ For the EPA two-cycle regulatory test on a dynamometer, an
additional weight of 300 lbs. is added to the vehicle curb weight.
This additional 300 lbs. represents the weight of the driver,
passenger, and luggage. Depending on the final test weight of the
vehicle (vehicle curb weight plus 300 lbs.), a test weight category
is identified using the table published by EPA according to 40 CFR
1066.805. This test weight category is called ``Equivalent Test
Weight'' (ETW).
---------------------------------------------------------------------------
For this analysis, NHTSA considers six levels of mass reduction
technology that include increasing amounts of advanced materials and
mass reduction techniques applied to the glider. NHTSA accounts for
changes in mass associated with powertrain changes separately. The
following sections discuss the assumptions for the six mass reduction
technology levels, the process used to assign initial analysis fleet
mass reduction assignments, the effectiveness for applying mass
reduction technology, and mass reduction costs.
(a) Mass Reduction in the CAFE Model
The CAFE Model considers six levels of mass reduction technologies
that manufacturers could use to comply with CAFE standards. The
magnitude of mass reduction in percent for each of these levels is
shown in Table III-22 for mass reductions for light trucks, passenger
cars and for gliders.
BILLING CODE 4910-59-C
[GRAPHIC] [TIFF OMITTED] TR02MY22.086
For this analysis, NHTSA considers mass reduction opportunities
from the glider subsystems of a vehicle first, and then consider
associated opportunities to downsize the powertrain, which are
accounted for separately.\427\ As explained below, in the Autonomie
simulations, the glider system includes both primary and secondary
systems from which a percentage of mass is reduced for different glider
weight reduction levels; specifically, the glider includes the body,
chassis, interior, electrical accessories, steering, brakes and wheels.
In this analysis, NHTSA assumes the glider share is 71 percent of
vehicle curb weight. The Autonomie model sizes the powertrain based on
the glider weight and the mass of some of the powertrain components in
an iterative process. The mass of the powertrain depends on the
powertrain size. Therefore, the weight of the glider impacts the weight
of the powertrain.\428\
---------------------------------------------------------------------------
\427\ When the mass of the vehicle is reduced by an appropriate
amount, the engine may be downsized to maintain performance. See
Section III.C.4 for more details.
\428\ Since powertrains are sized based on the glider weight for
the analysis, glider weight reduction beyond a threshold amount
during a redesign will lead to re-sizing of the powertrain. For the
analysis, the glider was used as a base for the application of any
type of powertrain. A conventional powertrain consists of an engine,
transmission, exhaust system, fuel tank, radiator, and associated
components. A hybrid powertrain also includes a battery pack,
electric motor(s), generator, high voltage wiring harness, high
voltage connectors, inverter, battery management system(s), battery
pack thermal system, and electric motor thermal system.
---------------------------------------------------------------------------
NHTSA uses glider weight to apply non-powertrain mass reduction
technology in the CAFE Model and use Autonomie simulations to determine
the size of the powertrain and corresponding powertrain weight for the
respective glider weight. The combination of glider weight (after mass
reduction) and re-sized powertrain weight equal the vehicle curb
weight.
While there are a range of specific mass reduction technologies
that may be applied to vehicles to achieve each of the six mass
reduction levels, there are some general trends that are helpful to
illustrate some of the more widely used approaches. Typically, MR0
reflects vehicles with widespread use of mild steel structures and body
panels, and very little or no use of high strength steel or aluminum.
MR0 reflects materials applied to average vehicles in the MY 2008
timeframe. MR1-MR3 can be achieved with a steel body structure. In
going from MR1 to MR3, expect that mild steel to be replaced by high
strength and then advanced high strength steels. In going from MR3 to
MR4 aluminum is required. This will start at using aluminum closure
panels and then to get to MR4 the vehicle's primary structure will need
to be mostly made from aluminum. In the vast majority of cases, carbon
fiber technology is necessary to reach MR5, perhaps with a mix of some
aluminum. MR6 requires nearly every primary structural component of the
vehicle, like body structure and closure panels, be made from carbon
fiber. There may be some use of aluminum in the suspension components.
TSD Chapter 3.4 includes more discussion of the challenges involved
with adopting large amounts of carbon fiber in the vehicle fleet.
Arconic Corporation commented that ``the NPRM makes specific
references to aluminum, which are accurate and consistent with
practical automotive industry experience and future program
expectations. Mass reduction utilizing advanced materials like aluminum
is recognized as one of the technology options to achieve safe, fuel-
efficient
[[Page 25828]]
and cost-effective vehicles that meet or exceed consumer demands.''
\429\
---------------------------------------------------------------------------
\429\ Arconic, Docket No. NHTSA-2021-0053-1560, at p. 1.
---------------------------------------------------------------------------
The American Chemistry Council (ACC) commented on the agency's
statements about vehicle light-weighting in several respects, but
particularly disagreeing with our analysis of mass reduction technology
levels.\430\ Specifically, ACC stated that ``as written, the NPRM could
be construed as NHTSA discouraging the use of carbon fiber composites
as well as an endorsement for utilizing steel and aluminum-based
designs to achieve mass reduction.'' \431\ ACC also provided updated
data on carbon fiber costs from DOE ORNL studies that they asked the
agency to consider in the final rule.
---------------------------------------------------------------------------
\430\ ACC, Docket No. NHTSA-2021-0053-1564, at p. 5.
\431\ Id.
---------------------------------------------------------------------------
To be clear, our analysis does not endorse any specific technology
solution or pathway over another. However, our analysis does need to
accurately reflect trends that are developing in the real-world
automotive marketplace and potential fuel economy improving technology
to appropriately estimate the costs and benefits of more stringent
standards. It also does need to consider what could reasonably occur in
the future of the market given automotive development timelines for
implementing new technology into real passenger vehicles. Precursor
materials technologies that potentially offer game-changing dry carbon
fiber cost reductions are still under development and therefore we
would not expect them to end up in a production vehicle program beyond
what our adoption features allow in the rulemaking timeframe.
In addition, while carbon fiber composites are considered a
potential pathway to compliance, wholly carbon fiber primary structure,
which is what is necessary to reduce mass enough to achieve the highest
mass reduction levels in the analysis, simply have not materialized.
While the number and mass of discrete applications of carbon fiber has
expanded the fleet--for example, adding carbon fiber decorative
interior trim pieces or carbon fiber roof panels to medium and high-end
luxury cars--these discrete applications do not contribute to
substantial mass reduction required to meet the highest levels of mass
reduction in this analysis. The price to apply carbon fiber technology
to produce wholly carbon fiber composite primary structure with the
precursor material available today has not yet dropped to a price that
would make it cost-effective for the industry to apply to meet more
stringent fuel economy standards. This fact is corroborated by the 2021
NAS Report, which provided updated data for carbon fiber composite
costs that show the technology has not yet dropped to a price that
would make it cost-effective for the industry to apply to meet more
stringent fuel economy standards. This is discussed further in Section
III.D.4.c) below. We also appreciate ACC's inclusion of the DOE ORNL
technoeconomic analysis on carbon fiber and discuss the study further
in Section III.D.4.e) below.
As discussed further below, the cost studies used to generate the
cost curves assume mass can be reduced in levels that require utilizing
different materials and modifying different components, in a specific
order. NHTSA's mass reduction levels are loosely based on what
materials and component modifications are required for each percent of
mass reduction, based on the conclusions of those studies.
(b) Mass Reduction Analysis Fleet Assignments
To assign baseline mass reduction levels (MR0 through MR6) for
vehicles in the MY 2020 analysis fleet, NHTSA uses previously developed
regression models to estimate curb weight for each vehicle based on
observable vehicle attributes. NHTSA uses these models to establish a
baseline (MR0) curb weight for each vehicle, and then determines the
existing mass reduction technology level by finding the difference
between the vehicles actual curb weight to the estimated regression-
based value, and comparing the difference to the values in Table III-
22. NHTSA originally developed the mass reduction regression models
using MY 2015 fleet data; for this analysis, NHTSA used MY 2016 and
2017 analysis fleet data to update the models.
NHTSA believes the regression methodology is a technically sound
approach for estimating mass reduction levels in the analysis fleet.
For a detailed discussion about the regression development and use
please see TSD Chapter 3.4.2.
Manufacturers generally apply mass reduction technology at a
vehicle platform level (i.e., using the same components across multiple
vehicle models that share a common platform) to leverage economies of
scale and to manage component and manufacturing complexity, so
conducting the regression analysis at the platform level leads to more
accurate estimates for the real-world vehicle platform mass reduction
levels. The platform approach also addresses the impact of potential
weight variations that might exist for specific vehicle models, as all
the individual vehicle models are aggregated into the platform group,
and are effectively averaged using sales weighting, which minimizes the
impact of any outlier vehicle configurations.
(c) Mass Reduction Adoption Features
Given the degree of commonality among the vehicle models built on a
single platform, manufacturers do not have complete freedom to apply
unique mass reduction technologies to each vehicle model that shares
the platform. While some technologies (e.g., low rolling resistance
tires) are very nearly ``bolt-on'' technologies, others involve
substantial changes to the structure and design of the vehicle, and
therefore affect all vehicle models that share a platform. In most
cases, mass reduction technologies are applied to platform level
components and therefore the same design and components are used on all
vehicle models that share the platform.
Each vehicle in the analysis fleet is associated with a specific
platform. Similar to the application of engine and transmission
technologies, the CAFE Model defines a platform ``leader'' as the
vehicle variant of a given platform that has the highest level of
observed mass reduction present in the analysis fleet. If there is a
tie, the CAFE Model begins mass reduction technology on the vehicle
with the highest sales volume in MY 2020. If there remains a tie, the
model begins by choosing the vehicle with the highest manufacturer
suggested retail price (MSRP) in MY 2020. As the model applies
technologies, it effectively levels up all variants on a platform to
the highest level of mass reduction technology on the platform. For
example, if the platform leader model is already at MR3 in MY 2020, and
a ``follower'' platform model starts at MR0 in MY 2020, the follower
platform model will get MR3 at its next redesign, assuming no further
mass reduction technology is applied to the leader model before the
follower model's next redesign.
In addition to the platform-sharing logic employed in the model,
NHTSA applies phase-in caps for MR5 and MR6 (15 percent and 20 percent
reduction of a vehicle's curb weight, respectively), based on the
current state of mass reduction technology. As discussed above, for
nearly every type of vehicle, a manufacturer's strategy to achieve mass
reduction consistent with MR5 and MR6 will require extensive use of
carbon fiber technologies in the vehicles' primary structures. For
example, one way of using carbon fiber
[[Page 25829]]
technology to achieve MR6 is to develop a carbon fiber monocoque
structure.\432\
---------------------------------------------------------------------------
\432\ A monocoque structure is one where the outer most skins
support the primary loads of the vehicle. For example, they do not
have separate non-load bearing aero surfaces. All of the vehicle's
primary loads are supported by the monocoque. In the most
structurally efficient automotive versions, the monocoque is made
from multiple well-consolidated plies of carbon fiber infused with
resin. Such structures would likely require a few hundred kilograms
of carbon fiber for most passenger vehicles.
---------------------------------------------------------------------------
High CAFE stringency levels will push the CAFE Model to select
compliance pathways that include these higher levels of mass reduction
for vehicles produced in the mid and high hundreds of thousands of
vehicles per year. NHTSA assumes, based on material costs and
availability, that achieving MR6 levels of mass reduction will cost
over ten thousand dollars per car. The cost of achieving MR6 in the
CAFE Model is consistent with our understanding of the real-world costs
to produce a carbon fiber monocoque structure.\433\ Therefore,
application of such technology to high volume vehicles is unrealistic
today and will, with certainty, remain so for the next several years.
---------------------------------------------------------------------------
\433\ In simplest terms, the cost to produce a component made
from carbon fiber composite materials is the sum of the cost of dry
carbon fiber, resin, amortized tooling, direct labor, and factor
overhead. A BMW i3 monocoque contains between 100 and 150 kg of
carbon fiber composite material depending on source (see article on
https://www.marklines.com/en/report_all/rep1419_201506, (accessed:
Feb. 11, 2022). ``Recent Trends in CFRP Development: Increased Usage
in European Vehicles, July 2015, and see book: ``Lightweight and
Sustainable Materials for Automotive Applications,'' Chapter 8,
2017, CRC Press). Assuming a very typical 60/40 mix of carbon fiber
to resin, and assuming the price of dry carbon fiber is $20-$40 per
kilogram and the price of resin is $5-$10 per kilogram, the cost of
direct materials alone in an i3's carbon fiber monocoque is already
approaching $4,200. Adding direct labor, factory overhead (which
scales with cycle time) and the amortized cost of tooling can easily
bring the cost for components made from composite materials in the
i3 to a higher level. Note that the BMW i3 is on the small end of
the size spectrum in the 2020 fleet and these costs would increase
faster than proportional to vehicle footprint because of the mass
compounding effect. Therefore, the cost of a monocoque for a large
sedan (the current BMW 7-series has a foot-print that is 30 percent
higher than that of the i3) could easily cost over ten thousand
dollars.
---------------------------------------------------------------------------
The CAFE Model applies technologies to vehicles that provide a
cost-effective pathway to compliance. In some cases, the direct
manufacturing cost, indirect costs, and applied learning factor do not
capture all the considerations that make a technology more or less
costly for manufacturers to apply in the real world. For example, there
are direct labor, R&D overhead, manufacturing overhead and tooling
costs. Due to the complexities of manufacturing composite components,
many of these are more expensive for manufacturing carbon fiber
components than for manufacturing metal components. Next, as of yet, no
one has found an effective way to recycle carbon fiber composites,
which means there saving money through re-using material is a
challenge. In addition, R&D overhead will also increase because of the
knowledge base for composite materials in automotive applications is
simply not as deep as it is for steel and aluminum.
ACC commented on this characterization of potential costs for
carbon fiber technology, using it as an example of where, as discussed
above, they believed the NPRM could be construed as NHTSA discouraging
the use of carbon fiber composites.\434\ However, the views stated in
the previous paragraph explaining why carbon fiber technologies are not
widespread are not indicative of NHTSA discouraging the use of or
further development carbon fiber technologies. Rather, they reflect
what has actually occurred in the automotive market and views shared by
others. In fact, BMW decided that a mixed materials solution is a more
financially effective way to reduce mass and will not build a wholly
carbon fiber composite successor to the i3.\435\ \436\ \437\ \438\
\439\
---------------------------------------------------------------------------
\434\ ACC, at p. 5.
\435\ Brosius, Dale, ``Carbon Fiber in Automotive: At a Dead
End?'' Composites World, December 20, 2021.
\436\ Sloan, Jeff, ``AutoComposites and the Myth of $5/lb.
Carbon Fiber,'' Composites World, February 24, 2017.
\437\ Taylor, Edward and Sage, Alexandria, ``BMW Limits
Lightweight Carbon Fibre Use to Juice Profits,'' Reuters, October
2016.
\438\ Bunkley, Nick, ``BMW Limits Carbon Fiber Use to Protect
Profits,'' Autonews Gasgoo, October 31, 2016.
\439\ Schlosser, Andreas, Coskun Baban, Samith, and Siedel
Phillipp, ``After the Hype: Where is the Carbon Car?'' Arthur D.
Little, January 2019.
---------------------------------------------------------------------------
Indeed, the intrinsic anisotropic mechanical properties of
composite materials compared to the isotropic properties of metals
complicates the design process. Added testing of these novel
anisotropic structures and their associated costs will be necessary for
decades. Adding up all these contributing costs, the price tag for a
passenger car or truck monocoque would likely be multiple tens of
thousands of dollars per vehicle. This would be significantly more
expensive than transitioning to hybrid or fully electric powertrains
and potentially less effective at achieving CAFE compliance.
In addition, the CAFE Model does not currently enable direct
accounting for the stranded capital associated with a transition away
from stamped sheet metal construction to molded composite materials
construction. For decades, or in some cases half-centuries, car
manufacturers have invested billions of dollars in capital for
equipment that supports the industry's sheet metal forming paradigm. A
paradigm change to tooling and equipment developed to support molding
carbon fiber panels and monocoque chassis structures would leave that
capital stranded in equipment that would be rendered obsolete. Doing
this is possible, but the financial ramifications are not currently
reflected in the CAFE Model for MR5 and MR6 compliance pathways.
Financial matters aside, carbon fiber technology and how it is best
used to produce light-weight primary automotive structures is far from
mature. In fact, no car company knows for sure the best way to use
carbon fiber to make a passenger car's primary structure. Using this
technology in passenger cars is far more complex than using it in
racing cars where passenger egress, longevity, corrosion protection,
crash protection, etc. are lower on the list of priorities for the
design team. BMW may be the one manufacturer most able accurately opine
on the viability of carbon fiber technology for primary structure on
high-volume passenger cars, and even it decided to use a mixed
materials solution for their next generation of EVs (the iX and i4)
after the i3, thus eschewing a wholly carbon fiber monocoque structure.
Another factor limiting the application of carbon fiber technology
to mass volume passenger vehicles is indeed the availability of dry
carbon fibers. There is high global demand from a variety of industries
for a limited supply of carbon fibers. Aerospace, military/defense, and
industrial applications demand most of the carbon fiber currently
produced. Today, only roughly 10 percent of the global dry fiber supply
goes to the automotive industry, which translates to the global supply
base only being able to support approximately 80,000 cars.\440\
---------------------------------------------------------------------------
\440\ J. Sloan, ``Carbon Fiber Suppliers Gear up for Next
Generation Growth,'' compositesworld.com, February 11, 2020.
---------------------------------------------------------------------------
To account for these cost and production considerations, including
the limited global supply of dry carbon fiber, NHTSA applied phase-in
caps that limited the number of vehicles that can achieve MR5 and M6
levels of mass reduction in the CAFE Model. NHTSA applied a phase-in
cap for MR5 level technology so that 75 percent of the vehicle fleet
starting in 2020 could employ the technology, and the technology could
be applied to 100
[[Page 25830]]
percent of the fleet by MY 2022. NHTSA also applied a phase-in cap for
MR6 technology so that five percent of the vehicle fleet starting in MY
2020 could employ the technology, and the technology could be applied
to 10 percent of the fleet by MY 2025.
To develop these phase-in caps, NHTSA chose a 40,000-unit threshold
for both MR5 and MR6 technology (80,000 units total), because it
roughly reflects the number of BMW i3 cars produced per year
worldwide.\441\ As discussed above, the BMW i3 is the only high-volume
vehicle currently produced with a primary structure mostly made from
carbon fiber (except the skateboard, which is aluminum). Because mass
reduction is applied at the platform level (meaning that every car of a
given platform would receive the technology, not just special low
volume versions of that platform), only platforms representing 40,000
vehicles or less are eligible to apply MR5 and MR6 toward CAFE
compliance. Platforms representing high volume sales, like a Chevrolet
Traverse, for example, where hundreds of thousands are sold per year,
are therefore blocked from access to MR5 and MR6 technology. There are
no phase in caps for mass reduction levels MR1, MR2, MR3 or MR4.
---------------------------------------------------------------------------
\441\ However, even this number is optimistic because only a
small fraction of i3 cars are sold in the U.S. market, and combining
MR5 and MR6 allocations equates to 80k vehicles, not 40k.
Regardless, if the auto industry ever seriously committed to using
carbon fiber in mainstream high-volume vehicles, competition with
the other industries would rapidly result in a dramatic increase in
price for dry fiber. This would further stymie the deployment of
this technology in the automotive industry.
---------------------------------------------------------------------------
In addition to determining that the caps were reasonable based on
current global carbon fiber production, NHTSA determined that the MR5
phase-in cap is consistent with the NHTSA light-weighting study that
found that a 15 percent curb weight reduction for the fleet is possible
within the rulemaking timeframe.\442\
---------------------------------------------------------------------------
\442\ Singh, Harry. (2012, August). Mass Reduction for Light-
Duty Vehicles for Model Years 2017-2025. (Report No. NHTSA HS 811
666). Program Reference: NHTSA Contract DTNH22-11-C-00193. Contract
Prime: Electricore, Inc, at 356, Figure 397.
---------------------------------------------------------------------------
These phase-in caps appropriately function as a proxy for the cost
and complexity currently required (and that likely will continue to be
required until manufacturing processes evolve) to produce carbon fiber
components. Again, MR6 technology in this analysis reflects the use of
a significant share of carbon fiber content, as seen through the BMW i3
and Alfa Romeo 4c as discussed above.
Given the uncertainty and fluid nature of knowledge around higher
levels of mass reduction technology, we welcomed comments on how to
most cost effectively use carbon fiber technology in high-volume
passenger cars. We also stated that financial implementation estimates
for this technology are equally as welcome.
NHTSA received comment involving the ability of auto industry
suppliers to procure dry carbon fiber materials in quantities
consistent with supplying high-volume platforms. Commenters suggested
that the industry that produces dry carbon fiber could readily ramp-up
fiber production at a rate fast enough to accommodate the demands of
multiple high volume automotive platforms such as the Chevrolet
Traverse or Volvo XC90, all within the time frame in which this rule
applies.\443\ Commenters did not mention specific achievable production
volumes or detail a production volume trajectory as a function of time.
In addition, ACC commented that it was misleading for NHTSA to state
that only roughly 10 percent of the global dry fiber supply goes to the
automotive industry, that 10 percent would only be enough for roughly
70,000 vehicles and that producers of dry carbon fiber would not scale
their output to support high volume production automotive programs.
Based on available literature, engineering judgment and the composition
of the current fleet, we continue to believe that MR5 or MR6 will not
be achievable for large volume platforms in the rulemaking
timeframe.\444\ Sources in the literature indicate that if only three
mass volume auto makers used 8-9 kg of carbon fiber (which would not
meet MR5 or MR6 levels) in each of their vehicles, the carbon fiber
industry would need to double its output. Using only 8-9 kg of carbon
fiber per vehicle will never enable mass reduction consistent with MR5
or MR6. The amount of carbon fiber required for this would require at
least an order of magnitude more than 8-9 kg. Fiber producers cannot
double their output in the rulemaking timeframe let alone increase it
by twenty-fold within the same timeframe.\445\
---------------------------------------------------------------------------
\443\ ACC, at p. 5.
\444\ Bill, Bregar, ``Prices Keep Carbon Fiber from Mass
Adoption,'' Plastic News, August 5, 2014.
\445\ ``How to Turn Pitch into Carbon Fiber for Automotive
Applications,'' https://www.azom.com/article.aspx?ArticleID=19200
(accessed: Feb. 11, 2022).
---------------------------------------------------------------------------
In addition, since publication of the NPRM, BMW stopped producing
its i3 vehicle, the only mass-volume vehicle built with nearly full
carbon fiber construction. The i3 was replaced with a vehicle
containing only a small fraction of the amount carbon fiber composite
materials as its predecessor. BMW decided a multi-materials solution
was more cost effective.\446\ \447\ Currently, the few vehicles that
continue to use carbon fiber do so in only small fractions or they are
not mass-market vehicles.\448\ We are not currently aware of any high-
volume cars planned for the near future with nearly full carbon fiber
construction. If that remains the case, there is no incentive to
dramatically boost production of dry carbon fiber to support the auto
industry.
---------------------------------------------------------------------------
\446\ Taylor, Edward and Sage, Alexandria, ``BMW Limits
Lightweight Carbon Fibre Use to Juice Profits,'' Reuters, October
2016.
\447\ Bunkley, Nick, ``BMW Limits Carbon Fiber Use to Protect
Profits,'' Autonews Gasgoo, October 31, 2016.
\448\ See, e.g., the BMW iX and i4, and some Lamborghini
vehicles.
---------------------------------------------------------------------------
There may be some emerging methods to provide a lower cost pathway
to MR6, like selectively applying high-modulus carbon fiber tapes to
lower cost structures primarily made from fiberglass composites.\449\
Although these methods may reduce the cost of direct materials, the do
not alleviate slow production cycle times and the costs associated with
them.
---------------------------------------------------------------------------
\449\ By strategic application of carbon fiber in areas of
highest stress in a given structure, it is often possible to achieve
sufficient structural performance at a lower cost. However, this
strategy does not solve the aforementioned issues surrounding the
high costs associated with the relatively long production cycle
times of composite materials composites.
---------------------------------------------------------------------------
The analysis herein uses the 2020 fleet to evaluate the level of
mass reduction (MR0-MR6) achieved by individual vehicle platforms. In
total, a little more than 25,000 vehicles of a fleet containing roughly
16 million vehicles achieved MR5 and MR6. It is expected that achieving
MR5 will require at least some carbon fiber technology and achieving
MR6 will require nearly full carbon fiber construction. Of the 25,000
vehicles, about 5,000 vehicles have nearly full carbon fiber
construction. These vehicles are produced by BMW (the i3 and i8), the
VW Group (Bugatti and Lamborghini) and few others that are not big
enough to be included in the 2020 fleet. Noteworthy is that there are
service vans in the fleet that achieve the highest MR levels, but only
for the reason that they have large footprints (wheelbase times average
track) and do not include interior trim and luxury items. Given this
small representation of vehicles with nearly full carbon fiber
construction, and current trends in
[[Page 25831]]
automotive carbon fiber application, discussed above, we do not believe
that multiple large-volume platforms would be able to reach MR6 in the
rulemaking timeframe.
We will continue to monitor carbon fiber investments from the
automotive sector, whether for full carbon fiber construction bodies or
carbon fiber parts, and on the implications of such investments for
automotive application carbon fiber demand, capacity, and supply. Based
on these observations, however, we declined to update any of our mass
reduction adoption features for this final rule.
(d) Mass Reduction Effectiveness Modeling
As discussed in Section III.C.4, Argonne developed a database of
vehicle attributes and characteristics for each vehicle technology
class that included over 100 different attributes. Some examples from
these 100 attributes include frontal area, drag coefficient, fuel tank
weight, transmission housing weight, transmission clutch weight, hybrid
vehicle components, and weights for components that comprise engines
and electric machines, tire rolling resistance, transmission gear
ratios, and final drive ratio. Argonne used these attributes to
``build'' each vehicle that it used for the effectiveness modeling and
simulation. Important for precisely estimating the effectiveness of
different levels of mass reduction is an accurate list of initial
component weights that make up each vehicle subsystem, from which
Autonomie considered potential mass reduction opportunities.
As stated above, NHTSA uses glider weight, or the vehicle curb
weight minus the powertrain weight, to determine the potential
opportunities for weight reduction irrespective of the type of
powertrain.\450\ This is because weight reduction can vary depending on
the type of powertrain. For example, an 8-speed transmission may weigh
more than a 6-speed transmission, and a basic engine without variable
valve timing may weigh more than an advanced engine with variable valve
timing. Autonomie simulations account for the weight of the powertrain
system inherently as part of the analysis, and the powertrain mass
accounting is separate from the application and accounting for mass
reduction technology levels that are applied to the glider in the
simulations. Similarly, Autonomie also accounts for battery and motor
mass used in hybrid and electric vehicles separately. This secondary
mass reduction is discussed further below.
---------------------------------------------------------------------------
\450\ Depending on the powertrain combination, the total curb
weight of the vehicle includes glider, engine, transmission and/or
battery pack and motor(s).
---------------------------------------------------------------------------
Accordingly, in the Autonomie simulations, mass reduction
technology is simulated as a percentage of mass removed from the
specific subsystems that make up the glider, as defined for that set of
simulations (including the non-powertrain secondary mass systems such
as the brake system). For the purposes of determining a reasonable
percentage for the glider, NHTSA in consultation with Argonne examined
glider weight data available in the A2Mac1 database,\451\ in addition
to the NHTSA MY 2014 Chevrolet Silverado light-weighting study
(discussed further below). Based on these studies, NHTSA assumes that
the glider weight comprised 71 percent of the vehicle curb weight. TSD
Chapter 3.4.4 includes a detailed breakdown of the components that
NHTSA considered to arrive at the conclusion that a glider, on average,
represents 71 percent of a vehicle's curb weight.
---------------------------------------------------------------------------
\451\ A2Mac1: Automotive Benchmarking, https://a2mac1.com.
---------------------------------------------------------------------------
Any mass reduction due to powertrain improvements is accounted for
separately from glider mass reduction. Autonomie considers several
components for powertrain mass reduction, including engine downsizing,
and transmission, fuel tank, exhaust systems, and cooling system light-
weighting.
The 2015 NAS Report suggested an engine downsizing opportunity
exists when the glider mass is light-weighted by at least 10 percent.
The 2015 NAS Report also suggested that 10 percent light-weighting of
the glider mass alone would boost fuel economy by 3 percent and any
engine downsizing following the 10 percent glider mass reduction would
provide an additional 3 percent increase in fuel economy.\452\ The 2011
Honda Accord and 2014 Chevrolet Silverado light-weighting studies
applied engine downsizing (for some vehicle types but not all) when the
glider weight was reduced by 10 percent. Accordingly, this analysis
limited engine resizing to several specific incremental technology
steps as in the 2018 NPRM and 2020 final rule; important for this
discussion, engines in the analysis were only resized when mass
reduction of 10 percent or greater was applied to the glider mass, or
when one powertrain architecture was replaced with another
architecture.
---------------------------------------------------------------------------
\452\ 2015 NAS Report. National Research Council. 2015. Cost,
Effectiveness, and Deployment of Fuel Economy Technologies for
Light-Duty Vehicles. Washington, DC--The National Academies Press.
https://doi.org/10.17226/21744, (accessed: Feb. 11, 2022).
---------------------------------------------------------------------------
Specifically, we allow engine resizing upon adoption of 7.1, 10.7,
14.2, and 20 percent curb weight reduction, but not at 3.6 and 5.3
percent.\453\ Resizing is also allowed upon changes in powertrain type
or the inheritance of a powertrain from another vehicle in the same
platform. The increments of these higher levels of mass reduction, or
complete powertrain changes, more appropriately match the typical
engine displacement increments that are available in a manufacturer's
engine portfolio.
---------------------------------------------------------------------------
\453\ These curb weight reductions equate to the following
levels of mass reduction as defined in the analysis: MR3, MR4, MR5
and MR6, but not MR1 and MR2; additional discussion of engine
resizing for mass reduction can be found in Section III.C.4 and TSD
Chapter 2.4.
---------------------------------------------------------------------------
Argonne performed a regression analysis of engine peak power versus
weight for a previous analysis based on attribute data taken from the
A2Mac1 benchmarking database, to account for the difference in weight
for different engine types. For example, to account for weight of
different engine sizes like 4-cylinder versus 8-cylinder, Argonne
developed a relationship curve between peak power and engine weight
based on the A2Mac1 benchmarking data. We use this relationship to
estimate mass for all engine types regardless of technology type (e.g.,
variable valve lift and direct injection). NHTSA applies weight
associated with changes in engine technology by using this linear
relationship between engine power and engine weight from the A2Mac1
benchmarking database. When a vehicle in the analysis fleet with an 8-
cylinder engine adopts a more fuel-efficient 6-cylinder engine, the
total vehicle weight reflects the updated engine weight with two less
cylinders based on the peak power versus engine weight relationship.
When Autonomie selects a powertrain combination for a light-
weighted glider, the engine and transmission are selected such that
there is no degradation in the performance of the vehicle relative to
the baseline vehicle. The resulting curb weight is a combination of the
mass reduced glider with the resized and potentially new engine and
transmission. This methodology also helps in accurately accounting for
the cost of the glider and cost of the engine and transmission in the
CAFE Model.
Secondary mass reduction is possible from some of the components in
the glider after mass reduction is applied to the primary subsystems
(body, chassis, and interior). Similarly, engine
[[Page 25832]]
downsizing and powertrain secondary mass reduction is possible after
certain level of mass reduction is incorporated in the glider. For the
analysis, the agencies include both primary mass reduction, and when
there is sufficient primary mass reduction, additional secondary mass
reduction. The Autonomie simulations account for the aggregate of both
primary and secondary glider mass reduction, and separately for
powertrain mass.
Note that secondary mass reduction is integrated into the mass
reduction cost curves. Specifically, the NHTSA studies, upon which the
cost curves depend, first generated costs for light-weighting the
vehicle body, chassis, interior, and other primary components, and then
calculated costs for light-weighting secondary components. Accordingly,
the cost curves reflect that, for example, secondary mass reduction for
the brake system is only applied after there has been sufficient
primary mass reduction to allow the smaller brake system to provide
safe braking performance and to maintain mechanical functionality.
NHTSA enhances the accuracy of estimated engine weights by using
two curves to represent separately naturally aspirated engine designs
and turbocharged engine designs.\454\ This achieves two benefits.
First, small naturally aspirated 4-cylinder engines that adopt
turbocharging technology reflects the increased weight of associated
components like ducting, clamps, the turbocharger itself, a charged air
cooler, wiring, fasteners, and a modified exhaust manifold. Second,
larger cylinder count engines like naturally aspirated 8-cylinder and
6-cylinder engines that adopt turbocharging and downsizing technologies
have lower weight due to having fewer engine cylinders. For this
analysis, a naturally aspirated 8-cylinder engine that adopts
turbocharging technology and is downsized to a 6-cylinder turbocharged
engine appropriately reflects the added weight of the turbocharging
components, and the lower weight of fewer cylinders.
---------------------------------------------------------------------------
\454\ See Autonomie model documentation, Chapter 5.2.9, Engine
Weight Determination.
---------------------------------------------------------------------------
The range of effectiveness values for the mass reduction
technologies, for all ten vehicle technology classes are shown in
Figure III-14. In the graph, the box shows the inner quartile range
(IQR) of the effectiveness values and whiskers extend out 1.5 x IQR.
The NHTSAs outside of the whiskers show a few values outside these
ranges. As discussed earlier, Autonomie simulates all possible
combinations of technologies for fuel consumption improvements. For a
few technology combinations mass reduction has minimal impact on
effectiveness on the regulatory 2-cycle test. For example, if an engine
is operating in an efficient region of the fuel map on the 2-cycle test
further reduction of mass may have smaller improvement on the
regulatory cycles. Figure III-14 shows the range improvements based on
the full range of other technology combinations considered in the
analysis.
BILLING CODE 4910-59-P
[[Page 25833]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.087
(e) Mass Reduction Costs
The CAFE Model analysis handles mass reduction technology costs
differently than all other technology costs. Mass reduction costs are
calculated as an average cost per pound over the baseline (MR0) for a
vehicle's glider weight. While the definitions of glider may vary,
NHTSA uses the same dollar per pound of curb weight to develop costs
for different glider definitions. In translating these values, NHTSA
takes care to track units ($/kg vs. $/lb.) and the reference for
percentage improvements (glider vs. curb weight).
NHTSA calculates the cost of mass reduction on a glider weight
basis so that the weight of each powertrain configuration can be
directly and separately accounted for. This approach provides the true
cost of mass reduction without conflating the mass change and costs
associated with downsizing a powertrain or adding additional advanced
powertrain technologies. Hence, the mass reduction costs in this final
rule reflect the cost of mass reduction in the glider and do not
include the mass reduction associated with engine downsizing. The mass
reduction and costs associated with engine downsizing are accounted for
separately.
A second reason for using glider share instead of curb weight is
that it affects the absolute amount of curb weight reduction applied,
and therefore cost per pound for the mass reduction changes with the
change in the glider share. The cost for removing 20 percent of the
glider weight when the glider represents 75 percent of a vehicle's curb
weight is not the same as the cost for removing 20 percent of the
glider weight when the glider represents 50 percent of the vehicle's
curb weight. For example, the glider share of 79 percent of a 3,000-
pound curb weight vehicle is 2,370 lbs., while the glider share of 50
percent of a 3,000-pound curb weight vehicle is 1,500 lbs., and the
glider share of 71 percent of a 3,000-pound curb weight vehicle is
2,130 lbs. The mass change associated with 20 percent mass reduction is
474 lbs. for 79 percent glider share (= [3,000 lbs. x 79% x 20%]), 300
lbs. for 50 percent glider share (= [3,000 lbs. x 50% x 20%]), and 426
lbs. for 71 percent glider share (= [3,000 lbs. x 71% x 20%]). The mass
reduction cost studies that NHTSA relies on to develop mass reduction
costs for this analysis show that the cost for mass reduction varies
with the amount of mass reduction. Therefore, for a fixed glider mass
reduction percentage, different glider share assumptions will have
different costs.
NHTSA considered several sources to develop the mass reduction
technology cost curves. Several mass reduction studies have used either
a mid-size passenger car or a full-size pickup truck as an exemplar
vehicle to demonstrate the technical and cost feasibility of mass
reduction. While the findings of these studies may not apply directly
to different vehicle classes, the cost estimates derived for the mass
reduction technologies identified in these studies can be useful for
formulating general estimates of costs. As discussed further below, the
mass reduction cost curves developed for this analysis are based on two
light-weighting studies, and NHTSA also updated the curves based
[[Page 25834]]
on more recent studies to better account for the cost of carbon fiber
needed for the highest levels of mass reduction technology. The two
studies used for MR1 through MR4 costs included the teardown of a MY
2011 Honda Accord and a MY 2014 Chevrolet Silverado pickup truck, and
the carbon fiber costs required for MR5 and MR6 were updated based on
the 2021 NAS Report.\455\
---------------------------------------------------------------------------
\455\ This analysis applied the cost estimates per pound derived
from passenger cars to all passenger car segments, and the cost
estimates per pound derived from full-size pickup trucks to all
light-duty truck and SUV segments. The cost estimates per pound for
carbon fiber (MR5 and MR6) were the same for all segments.
---------------------------------------------------------------------------
Both teardown studies are structured to derive the estimated cost
for each of the mass reduction technology levels. NHTSA relies on the
results of those studies because they considered an extensive range of
material types, material gauge, and component redesign while taking
into account real world constraints such as manufacturing and assembly
methods and complexity, platform-sharing, and maintaining vehicle
utility, functionality and attributes, including safety, performance,
payload capacity, towing capacity, handling, NVH, and other
characteristics. In addition, NHTSA believes that the baseline vehicles
and mass reduction technologies assessed in the studies are still
reasonably representative of the technologies that may be applied to
vehicles in the MY 2020 analysis fleet to achieve up to MR4 level mass
reduction in the rulemaking timeframe. NHTSA adjusted the cost
estimates derived from the two studies to reflect the assumption that a
vehicle's glider weight consisted of 71 percent of the vehicle's curb
weight, and mass reduction as it pertains to achieving MR0-MR6 levels
would only come from the glider.
As discussed above, achieving the highest levels of mass reduction
often necessitates extensive use of advanced materials like higher
grades of aluminum, magnesium, or carbon fiber. We provided a survey of
information available regarding carbon fiber costs based on the Honda
Accord and Chevrolet Silverado teardown studies. In the Honda Accord
study, the estimated cost of carbon fiber was $5.37/kg, and the cost of
carbon fiber used in the Chevy Silverado study was $15.50/kg. The
$15.50 estimate closely matched the cost estimates from a BMW i3
teardown analysis,\456\ the cost figures provided by Oak Ridge National
Laboratory for a study from the IACMI Composites Institute,\457\ and
from a Ducker Worldwide presentation at the CAR Management Briefing
Seminar.\458\
---------------------------------------------------------------------------
\456\ Singh, Harry, FSV Body Structure Comparison with 2014 BMW
i3, Munro and Associates for World Auto Steel (June 3, 2015).
\457\ IACMI Baseline Cost and Energy Metrics (March 2017),
available at https://iacmi.org/wp-content/uploads/2016/10/Dale-Brosius-IACMI-1.pdf (accessed Feb. 11, 2022).
\458\ Ducker Worldwide, The Road Ahead--Automotive Materials
(2016), https://societyofautomotiveanalysts.wildapricot.org/resources/Pictures/SAA%20Sumit%20slides%20for%20Abey%20Abraham%20of%20Ducker.pdf,
(accessed: Feb. 11, 2022).
---------------------------------------------------------------------------
However, for this analysis, NHTSA relies on the cost estimates for
carbon fiber construction that NAS detailed in the 2021 Assessment of
Technologies for Improving Fuel Economy of Light-Duty Vehicles--Phase 3
recently completed by NAS.\459\ The study indicates that the sum of
direct materials costs plus manufacturing costs for carbon fiber
composite automotive components is $25.97 per pound in high volume
production. In order to use this cost in the CAFE Model it must be put
in terms of dollars per pound saved. Using an average vehicle curb
weight of 4000 lbs., a 71 percent glider share and the percent mass
savings associated with MR5 and MR6, it is possible to calculate the
number of pounds to be removed to attain MR5 and MR6. Also taken from
the NAS study is the assertion that carbon fiber substitution for steel
in an automotive component results in a 50 percent mass reduction.
Combining all this together, carbon fiber technology offers weight
savings at $24.60 per pound saved. This dollar per pound savings figure
must also be converted to a retail price equivalent (RPE) to account
for various commercial costs associated with all automotive components.
This is accomplished by multiplying $24.60 by the factor 1.5. This
brings the cost per pound saved for using carbon fiber to $36.90 per
pound saved.\460\ The analysis uses this cost for achieving MR5 and
MR6.
---------------------------------------------------------------------------
\459\ 2021 NAS Report, at p. 219.
\460\ See MR5 and MR6 CFRP Cost Increase Calculator.xlsx in the
docket for this action.
---------------------------------------------------------------------------
Table III-23 and Table III-24 show the cost values (in dollars per
pound) used in the CAFE Model with MR1-4 costs based on the cost curves
developed from the MY 2011 Honda Accord and MY 2014 Chevrolet Silverado
studies, and the updated MR5 and MR6 values that account for the
updated carbon fiber costs from the 2021 NAS Report. Both tables assume
a 71 percent higher glider share.
[GRAPHIC] [TIFF OMITTED] TR02MY22.088
[[Page 25835]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.089
There is a dramatic increase in cost going from MR4 to MR5 and MR6
for all classes of vehicles. However, while the increase in cost going
from MR4 to MR5 and MR6 is dramatic, the MY 2011 Honda Accord study,
the MY 2014 Chevrolet Silverado study, and the 2021 NAS Report all
included a steep increase to achieve the highest levels of mass
reduction technology.
Table III-25 provides an example of mass reduction costs in 2018$
over select model years for the medium car and pickup truck technology
classes as a dollar per pound value. The table shows how the $/lb.
value for each mass reduction level decreases over time because of cost
learning. For a full list of the $/lb. mass reduction costs used in the
analysis across all model years, see the Technologies file.
[GRAPHIC] [TIFF OMITTED] TR02MY22.090
BILLING CODE 4910-59-C
NHTSA received comment from the ACC regarding the costs used in the
analysis for carbon fiber technology and how new precursors will soon
be available with high potential to reduce the cost of dry carbon
fibers.\461\ These precursor materials include, lignin, mesophase pitch
and textile-grade polyacrylonitrile (TG-PAN). Commenters specifically
referenced research conducted into these precursor materials conducted
at the Carbon Fiber Technology Facility at Oak Ridge National
Laboratory.
---------------------------------------------------------------------------
\461\ ACC, at p. 5.
---------------------------------------------------------------------------
Indeed, a factor that dominates the price of dry carbon fibers is
the precursor materials from which it is made. Dry carbon fibers that
are used in the mainstream automotive industry today, like those used
by BMW,\462\ are derived from high-molecular weight PAN fibers. The
high molecular weight of these materials not only makes the material
expensive, but it makes it more expensive to convert to carbon fiber
because it takes much longer to pyrolyze the fibers. However, the
result is a consistently stiff and incredibly high-strength fiber.
Prices today for traditional 3K tow (tow refers to the width of a
strand) PAN-based carbon fiber fall within the $20/kg to $40/kg
range.463 464 These price levels are consistent with NHTSA's
understanding and with the recent 2021 NAS Report.\465\
---------------------------------------------------------------------------
\462\ J. Sloan, ``Carbon Fiber Suppliers Gear up for Next
Generation Growth,'' compositesworld.com, February 11, 2020.
\463\ Schlosser, Andreas, Coskun Baban, Samith, and Siedel
Phillipp, ``After the Hype: Where is the Carbon Car?'' Arthur D.
Little, January 2019.
\464\ 2021 NAS Report, at pp. 218, 219, 419.
\465\ Id.
---------------------------------------------------------------------------
The commenters mentioned several other advancements in carbon fiber
technologies that are under development; however, we do not believe
these materials will be available for use in the rulemaking timeframe.
Lignin, which is an organic substance found in the cells of plants, has
great potential to achieve affordable carbon fibers and could
potentially be a lower-cost alternative to PAN.466 467 While
lignin is renewable, recyclable, sustainable, and cost effective, there
are stiffness and cost issues with lignin and research into lignin-
based carbon fiber has significantly slowed.\468\ Similarly, mesophase
pitch and TG-PAN are encouraging mass reduction technologies; \469\
however, based on
[[Page 25836]]
their developmental nature we do not believe they will be available for
commercial application in this rulemaking timeframe. Therefore, we do
not believe that the lower costs cited in the ORNL studies are
representative of the costs to industry for carbon fiber technology in
the rulemaking timeframe. We will continue to closely monitor these new
fiber precursor materials and how they may enable low-cost carbon fiber
technology with competitive mechanical properties.
---------------------------------------------------------------------------
\466\ Azarova, M.T., Semakina, N.S., Konkin, A.A. Tikhomirova,
M.V. ``Carbon Fiber Based on Meso-Phase-Pitches,'' Fiber Chemistry,
1982, pp. 103-110.
\467\ Kadla, J.F, et al., ``Lignin-Based Carbon Fibers for
Composite Applications,'' Carbon, Vol. 20, 2002, pp. 2913-2920.
\468\ For example, one issue with lignin-based carbon fiber is
that the density specific stiffness of fully pyrolyzed lignin-based
carbon fiber laminated in an epoxy matrix (which is a materials
property that often dominates mass reduction potential) is barely
competitive with that of steel. Yet steel costs about $1/kg--$3/kg.
Furthermore, because the absolute stiffness of lignin-based carbon
fiber composite material is low, a component made with lignin-based
carbon fiber composite material will require more packaging space
than a steel component to achieve equivalent component level
stiffness.
\469\ Mesophase pitch is made from coal which is plentiful and
therefore low cost, and the material has a density specific
stiffness better than steel, aluminum, and magnesium. TG-PAN has a
molecular weight that is about half that of traditional PAN
materials used from making carbon fiber and consequently requires
less time to pyrolyze, thus reducing its costs. In addition, textile
grade PAN is available in much wider tows (>= 50k) than traditional
PAN which means that more material can be converted to carbon fiber
in less time.
---------------------------------------------------------------------------
Aside from precursor materials issues, how dry carbon fibers are
processed into usable carbon fiber composite components is also an
important cost driver that we do not believe is represented in the
lower cited cost estimates. As an example, the carbon fiber composite
parts used on the BMW i3 are manufactured with cycle times between five
and ten minutes,\470\ while precise and accurate metallic parts are
produced in seconds.
---------------------------------------------------------------------------
\470\ Sloan, Jeff, ``BMW Leipzig: The Epicenter of i3
Production,'' Composites World, May 31, 2014.
---------------------------------------------------------------------------
Again, we will continue to monitor composite materials processing
technology advances and make cost adjustments in future analysis to
reflect advances in this field.
Aerodynamics
The energy required to overcome aerodynamic drag accounts for a
significant portion of the energy consumed by a vehicle and can become
the dominant factor for a vehicle's energy consumption at high speeds.
Reducing aerodynamic drag can, therefore, be an effective way to reduce
fuel consumption and emissions.
Aerodynamic drag is proportional to the frontal area (A) of the
vehicle and coefficient of drag (Cd), such that aerodynamic performance
is often expressed as the product of the two values, CdA, which is also
known as the drag area of a vehicle. The coefficient of drag (Cd) is a
dimensionless value that essentially represents the aerodynamic
efficiency of the vehicle shape. The frontal area (A) is the cross-
sectional area of the vehicle as viewed from the front. It acts with
the coefficient of drag as a sort of scaling factor, representing the
relative size of the vehicle shape that the coefficient of drag
describes. The force imposed by aerodynamic drag increases with the
square of vehicle velocity, accounting for the largest contribution to
road loads at higher speeds.
Aerodynamic drag reduction can be achieved via two approaches,
either by reducing the drag coefficient or reducing vehicle frontal
area, with two different categories of technologies, passive and active
aerodynamic technologies. Passive aerodynamics refers to aerodynamic
attributes that are inherent to the shape and size of the vehicle,
including any components of a fixed nature. Active aerodynamics refers
to technologies that variably deploy in response to driving conditions.
These include technologies such as active grille shutters, active air
dams, and active ride height adjustment. It is important to note that
manufacturers may employ both passive and active aerodynamic
technologies to achieve aerodynamic drag improvements.
The greatest opportunity for improving aerodynamic performance is
during a vehicle redesign cycle when the manufacturer can make
significant changes to the shape and size of the vehicle. A
manufacturer may also make incremental improvements during mid-cycle
vehicle refresh using restyled exterior components and add-on devices.
Some examples of potential technologies that a manufacturer could apply
during mid-cycle refresh are restyled front and rear fascia, modified
front air dams and rear valances, addition of rear deck lips and
underbody panels, and low-drag exterior mirrors. While manufacturers
may nudge the frontal area of the vehicle during redesigns, large
changes in the frontal area are typically not possible without
impacting the utility and interior space of the vehicle. Similarly,
manufacturers may improve Cd by changing the frontal shape of the
vehicle or lowering the height of the vehicle, among other approaches,
but the form drag of certain body styles and airflow needs for engine
cooling often limit how much manufacturers can improve Cd.
The following sections discuss the four levels of aerodynamic
improvements that we consider in the CAFE Model, how we assign baseline
aerodynamic technology levels to vehicles in the MY 2020 fleet, the
effectiveness improvements for the addition of aerodynamic technologies
to vehicles, and the costs for adding that aerodynamic technology.
(a) Aerodynamic Technologies in the CAFE Model
We bin aerodynamic improvements into four levels--5, 10, 15, and 20
percent aerodynamic drag improvement values over a baseline computed
for each vehicle body style--which correspond to AERO5, AERO10, AERO15,
and AERO20, respectively.
The aerodynamic improvements technology pathway consists of a
linear progression, with each level superseding all previous levels, as
seen in Figure III-15.
BILLING CODE 4910-59-P
[[Page 25837]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.091
While the four levels of aerodynamic improvements are technology-
agnostic, we built a pathway to compliance for each level based on
aerodynamic data from a National Research Council (NRC) of Canada-
sponsored wind tunnel testing program. The program included an
extensive review of production vehicles utilizing these technologies,
and industry comments.471 472 Again, these technology
combinations are intended to show a potential way for a manufacturer to
achieve each aerodynamic improvement level; however, in the real world,
manufacturers may implement different combinations of aerodynamic
technologies to achieve a percentage improvement over their baseline
vehicles.
---------------------------------------------------------------------------
\471\ Larose, G., Belluz, L., Whittal, I., Belzile, M. et al.,
``Evaluation of the Aerodynamics of Drag Reduction Technologies for
Light-duty Vehicles--a Comprehensive Wind Tunnel Study,'' SAE Int.
J. Passeng. Cars--Mech. Syst. 9(2):772-784, 2016, https://doi.org/10.4271/2016-01-1613, (accessed: Feb. 11, 2022).
\472\ Larose, Guy & Belluz, Leanna & Whittal, Ian & Belzile,
Marc & Klomp, Ryan & Schmitt, Andreas. (2016). Evaluation of the
Aerodynamics of Drag Reduction Technologies for Light-duty
Vehicles--a Comprehensive Wind Tunnel Study. SAE International
Journal of Passenger Cars--Mechanical Systems. 9. 10.4271/2016-01-
1613.
---------------------------------------------------------------------------
Table III-26 and Table III-27 show the aerodynamic technologies
that could be used to achieve 5, 10, 15, and 20 percent improvements in
passenger cars, SUVs, and pickup trucks. As discussed further in
Section III.D.5.c), the model does not apply AERO20 to pickup trucks,
which is why there is no pathway to AERO20 shown in Table III-27. While
manufacturers can apply some aerodynamic improvement technologies
across vehicle classes, like active grille shutters (used in the 2015
Chevrolet Colorado),\473\ we determined that there are limitations that
make it infeasible for vehicles with some body styles to achieve a 20
percent reduction in the coefficient of drag from their baseline. This
technology path is an example of how a manufacturer could reach each
AERO level, but they would not necessarily be required to use the
technologies.
---------------------------------------------------------------------------
\473\ Chevrolet Product Information, available at https://media.chevrolet.com/content/media/us/en/chevrolet/vehicles/colorado/2015/_jcr_content/iconrow/textfile/file.res/15-PG-Chevrolet-Colorado-082218.pdf, (accessed: Feb. 11, 2022).
---------------------------------------------------------------------------
[[Page 25838]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.092
[GRAPHIC] [TIFF OMITTED] TR02MY22.093
As discussed further in Section III.D.8, this analysis assumes
manufacturers apply off-cycle technology at rates defined in the Market
Data file. While the AERO levels in the analysis are technology-
agnostic, achieving AERO20 improvements does assume the use of active
grille shutters, which is an off-cycle technology.
Auto Innovators provided two comments on aerodynamic improvements.
Auto Innovators commented that it ``does not recommend considering
additional aerodynamic improvements (such as 25 percent aerodynamic
improvements, etc.). Some additional reductions in aerodynamic forces
may be possible if side view mirrors were no longer required by NHTSA
and FMVSSs.'' \474\
---------------------------------------------------------------------------
\474\ Auto Innovators, Docket No. NHTSA-2021-0053-1492, at pp.
62, 135.
---------------------------------------------------------------------------
We agree with Auto Innovators that we should not assume additional
aerodynamics technology adoption. We do not exceed 20 percent
aerodynamic improvement for all body styles and 15
[[Page 25839]]
percent improvement for the body styles discussed below.
We also agree with Auto Innovators that side view mirrors cause
additional aerodynamic drag. Due to existing Federal motor vehicle
safety regulations, we currently do not consider aerodynamic
improvements from removing side view mirrors in the CAFE Model
analysis.\475\
---------------------------------------------------------------------------
\475\ Federal motor vehicle safety standard (FMVSS) No. 111,
``Rear Visibility,'' currently requires that vehicles be equipped
with rearview mirrors to provide drivers with a view of objects that
are to their side or to their side and rear.
---------------------------------------------------------------------------
(b) Aerodynamics Analysis Fleet Assignments
We use a relative performance approach to assign an initial level
of aerodynamic drag reduction technology to each vehicle. Each AERO
level represents a percent reduction in a vehicle's aerodynamic drag
coefficient (Cd) from a baseline value for its body style.
For a vehicle to achieve AERO5, the Cd must be at least 5
percent below the baseline for the body style; for AERO10, 10 percent
below the baseline, and so on. Baseline aerodynamic assignment is
therefore a three-step process: Each vehicle in the fleet is assigned a
body style, the average drag coefficient is calculated for each body
style, and the drag coefficient for each vehicle model is compared to
the average for the body style.
We assign every vehicle in the fleet a body style; available body
styles included convertible, coupe, sedan, hatchback, wagon, SUV,
pickup, minivan, and van. These assignments do not necessarily match
the body styles that manufacturers use for marketing purposes. Instead,
we assign them based on analyst judgement, taking into account how a
vehicle's AERO and vehicle technology class assignments are affected.
Different body styles offer different utility and have varying levels
of baseline form drag. In addition, frontal area is a major factor in
aerodynamic forces, and the frontal area varies by vehicle. This
analysis considers both frontal area and body style as utility factors
affecting aerodynamic forces; therefore, the analysis assumes all
reduction in aerodynamic drag forces come from improvement in the drag
coefficient.
We computed the average drag coefficients for each body style using
the MY 2015 drag coefficients published by manufacturers, which were
used as the baseline values in the analysis. We harmonize the Autonomie
simulation baselines with the analysis fleet assignment baselines to
the fullest extent possible.\476\
---------------------------------------------------------------------------
\476\ See TSD Chapter 2.4.2 for a table of vehicle attributes
used to build the Autonomie baseline vehicle models. That table
includes a drag coefficient for each vehicle class.
---------------------------------------------------------------------------
We source the drag coefficients for each vehicle in the analysis
fleet from manufacturer specification sheets, when possible. However,
manufacturers did not consistently publicly report drag coefficients
for MY 2020 vehicles. If we could not find a publicly reported drag
coefficient, analyst judgment was sometimes used to assign an AERO
level. If no level was manually assigned, we used the drag coefficient
obtained from manufacturers to build the MY 2016 fleet,\477\ if
available. The MY 2016 drag coefficient values may not accurately
reflect the current technology content of newer vehicles but are, in
many cases, the most recent data available.
---------------------------------------------------------------------------
\477\ See 83 FR 42986 (Aug. 24, 2018). The MY 2016 fleet was
built to support the 2018 NPRM.
---------------------------------------------------------------------------
(c) Aerodynamics Adoption Features
As already discussed, we use a relative performance approach to
assign current aerodynamic technology (AERO) level to a vehicle. For
some body styles with different utility, such as pickup trucks, SUVs
and minivans, frontal area can vary, and this can affect the overall
aerodynamic drag forces. In order to maintain vehicle utility and
functionality related to passenger space and cargo space, we assume all
technologies that improve aerodynamic drag forces do so by reducing
Cd while maintaining frontal area.
Technology pathway logic for levels of aerodynamic improvement
consists of a linear progression, with each level superseding all
previous ones. Technology paths for AERO are illustrated in Figure III-
15.
The model does not consider the highest AERO levels for certain
body styles. In these cases, this means that AERO20, and sometimes
AERO15, can neither be assigned in the baseline fleet nor adopted by
the model. For these body styles, there are no commercial examples of
drag coefficients that demonstrate the required AERO15 or AERO20
improvement over baseline levels. We also deemed the most advanced
levels of aerodynamic drag simulated as not technically practicable
given the form drag of the body style and costed technology, especially
given the need to maintain vehicle functionality and utility, such as
interior volume, cargo area, and ground clearance. In short, we
`skipped' AERO15 for minivan body styles, and `skipped' AERO20 for
convertible, minivan, pickup, and wagon body styles.
We also do not allow application of AERO15 and AERO20 technology to
vehicles with more than 780 horsepower. There are two main types of
vehicles that informed this threshold: Performance internal combustion
engine (ICE) vehicles and high-power battery electric vehicles (BEVs).
In the case of the former, we recognize that manufacturers tune
aerodynamic features on these vehicles to provide desirable downforce
at high speeds and to provide sufficient cooling for the powertrain,
rather than reducing drag, resulting in middling drag coefficients
despite advanced aerodynamic features. Therefore, manufacturers may
have limited ability to improve aerodynamic drag coefficients for high
performance vehicles with internal combustion engines without reducing
horsepower. 1,655 units of sales volume in the baseline fleet include
limited application of aerodynamic technologies because of ICE vehicle
performance.\478\
---------------------------------------------------------------------------
\478\ Market Data file.
---------------------------------------------------------------------------
In the case of high-power battery electric vehicles, the 780-
horsepower threshold is set above the highest peak system horsepower
present on a BEV in the 2020 fleet. BEVs have different aerodynamic
behavior and considerations than ICE vehicles, allowing for features
such as flat underbodies that significantly reduce drag.\479\ BEVs are
therefore more likely to achieve higher AERO levels, so the horsepower
threshold is set high enough that it does not restrict AERO15 and
AERO20 application. Note that the CAFE Model does not force high levels
of AERO adoption; rather, higher AERO levels are usually adopted
organically by BEVs because significant drag reduction allows for
smaller batteries and, by extension, cost savings. BEVs represent
252,023 units of sales volume in the baseline fleet.\480\
---------------------------------------------------------------------------
\479\ 2020 EPA Automotive Trends Report, at p. 227.
\480\ Market Data file.
---------------------------------------------------------------------------
[[Page 25840]]
(d) Aerodynamics Effectiveness Modeling
To determine aerodynamic effectiveness, the CAFE Model and
Autonomie use individually assigned road load technologies for each
vehicle to appropriately assign initial road load levels and
appropriately capture benefits of subsequent individual road load
improving technologies.
The current analysis included four levels of aerodynamic
improvements, AERO5, AERO10, AERO15, and AERO20, representing 5, 10,
15, and 20 percent reduction in drag coefficient (Cd), respectively. We
assume that aerodynamic drag reduction can only come from reduction in
Cd and not from reduction of frontal area, to maintain vehicle
functionality and utility, such as passenger space, ingress/egress
ergonomics, and cargo space.
The effectiveness values for the aerodynamic improvement levels
relative to AERO0, for all ten vehicle technology classes, are shown in
Figure III-16. Each of the effectiveness values shown is representative
of the improvements seen for upgrading only the listed aerodynamic
technology level for a given combination of other technologies. In
other words, the range of effectiveness values seen for each specific
technology (e.g., AERO 15) represents the addition of AERO15 technology
(relative to AERO0 level) for every technology combination that could
select the addition of AERO15. It must be emphasized that the change in
fuel consumption values between entire technology keys is used,\481\
and not the individual technology effectiveness values. Using the
change between whole technology keys captures the complementary or non-
complementary interactions among technologies. The box shows the inner
quartile range (IQR) of the effectiveness values and whiskers extend
out 1.5 x IQR. The dots outside the whiskers show effectiveness values
outside those thresholds.
---------------------------------------------------------------------------
\481\ Technology key is the unique collection of technologies
that constitutes a specific vehicle, see TSD Chapter 2.4.7 for more
detail.
[GRAPHIC] [TIFF OMITTED] TR02MY22.094
(e) Aerodynamics Costs
This analysis uses the AERO technology costs established in the
2020 final rule that are based on confidential business information
submitted by the automotive industry in advance of the 2018 NPRM,\483\
and on our assessment of manufacturing costs for specific aerodynamic
technologies.\484\ We received no additional comments from stakeholders
regarding the costs established in the 2018 NPRM, and
[[Page 25841]]
continued to use the established costs for the 2020 final rule and this
analysis.
---------------------------------------------------------------------------
\482\ The data used to create this figure can be found in the
FE_1 Improvements file.
\483\ See the PRIA accompanying the 2018 NPRM, Chapter
6.3.10.1.2.1.2, for a discussion of these cost estimates.
\484\ See the FRIA accompanying the 2020 final rule, Chapter
VI.C.5.e.
---------------------------------------------------------------------------
Table III-28 shows examples of costs for AERO technologies as
applied to the medium car and pickup truck vehicle classes in select
model years. The cost to achieve AERO5 is relatively low, as most of
the improvements can be made through body styling changes. The cost to
achieve AERO10 is higher than AERO5, due to the addition of several
passive aerodynamic technologies, and the cost to achieve AERO15 and
AERO20 is higher than AERO10 due to use of both passive and active
aerodynamic technologies. For a full list of all absolute aerodynamic
technology costs used in the analysis across all model years see the
Technologies file.
[GRAPHIC] [TIFF OMITTED] TR02MY22.095
Tire Rolling Resistance
Tire rolling resistance is a road load force that arises primarily
from the energy dissipated by elastic deformation of a vehicle's tires
as they roll. Tire design characteristics (for example, materials,
construction, and tread design) have a strong influence on the amount
and type of deformation and the energy the tire dissipates. Designers
can select these characteristics to minimize rolling resistance.
However, these characteristics may also influence other performance
attributes, such as durability, wet and dry traction, handling, and
ride comfort.
Lower rolling resistance tires have characteristics that reduce
frictional losses associated with the energy dissipated mainly in the
deformation of the tires under load, thereby improving fuel economy.
OEMs increasingly specify low rolling resistance tires in new vehicles,
and they are also increasingly available from aftermarket tire vendors.
They commonly include attributes such as higher inflation pressure,
material changes, tire construction optimized for lower hysteresis,
geometry changes (e.g., reduced aspect ratios), and reduced sidewall
and tread deflection. These changes are commonly accompanied by
additional changes to vehicle suspension tuning and/or suspension
design to mitigate any potential impact on other performance attributes
of the vehicle.
We continue to assess the potential impact of tire rolling
resistance changes on vehicle safety. We have been following the
industry developments and trends in application of rolling resistance
technologies to light duty vehicles. As stated in the NAP special
report on Tires and Passenger Vehicle Fuel Economy,\485\ national crash
data does not provide data about tire structural failures specifically
related to tire rolling resistance, because the rolling resistance of a
tire at a crash scene cannot be determined. However, other metrics like
brake performance compliance test data are helpful to show trends like
that stopping distance has not changed in the last ten years,\486\
during which time many manufacturers have installed low rolling
resistance tires in their fleet--meaning that manufacturers were
successful in improving rolling resistance while maintaining stopping
distances through tire design, tire materials, and/or braking system
improvements. In addition, NHTSA has addressed other tire-related
issues through rulemaking,\487\ and continues to research tire problems
such as blowouts, flat tires, tire or wheel deficiency, tire or wheel
failure, and tire degradation.\488\ However, there are currently no
data connecting low rolling resistance tires to accident or fatality
rates.
---------------------------------------------------------------------------
\485\ Tires and Passenger Vehicle Fuel Economy: Informing
Consumers, Improving Performance--Special Report 286 (2006),
available at https://www.nap.edu/read/11620/chapter/6.
\486\ See, e.g., NHTSA Office of Vehicle Safety Compliance,
Compliance Database, https://one.nhtsa.gov/cars/problems/comply/index.cfm.
\487\ 49 CFR 571.138, Tire pressure monitoring systems.
\488\ Tire-Related Factors in the Pre-Crash Phase, DOT HS 811
617 (April 2012), available at https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/811617.
---------------------------------------------------------------------------
NHTSA conducted tire rolling resistance tests and wet grip index
tests on original equipment tires installed on new vehicles. The tests
showed that there is no degradation in wet grip index values (i.e., no
degradation in traction) for tires with improved rolling resistance
technology. With better tire design, tire compound formulations and
improved tread design, tire manufacturers have tools to balance
stopping distance and reduced rolling resistance. Tire manufacturers
can use ``higher performance materials in the tread compound, more
silica as reinforcing fillers and advanced tread design features'' to
mitigate issues related to stopping distance.\489\
---------------------------------------------------------------------------
\489\ Jesse Snyder, A big fuel saver: Easy-rolling tires (but
watch braking) (July 21, 2008), https://www.autonews.com/article/20080721/OEM01/307219960/a-big-fuel-saver-easy-rolling-tires-but-watch-braking. Last visited December 3, 2019.
---------------------------------------------------------------------------
U.S. Tire Manufacturers Association (USTMA) commented on NHTSA's
conclusion that the agency did not observe any unacceptable tradeoff
between tire rolling resistance and wet grip performance, which ``NHTSA
correctly recognized is due to advanced tire design, rubber compounding
and manufacturing technologies.'' However, USTMA cautioned that ``this
inverse relationship between rolling resistance and wet grip
performance still exists, and as the tire industry continues to enhance
rolling resistance performance, new and/or enhanced countermeasures
will also need to be developed to assure
[[Page 25842]]
no unacceptable impact to wet grip performance.'' \490\
---------------------------------------------------------------------------
\490\ USTMA, Docket No. NHTSA-2021-0053-1612, at 2.
---------------------------------------------------------------------------
The following sections discuss levels of tire rolling resistance
technology considered in the CAFE Model, how the technology was
assigned in the analysis fleet, adoption features specified to maintain
performance, effectiveness, and cost.
(a) Tire Rolling Resistance in the CAFE Model
We continue to consider two levels of improvement for low rolling
resistance tires in the analysis: the first level of low rolling
resistance tires considered reduced rolling resistance 10 percent from
an industry-average baseline rolling resistance coefficient (RRC)
value, while the second level reduced rolling resistance 20 percent
from the baseline.\491\
---------------------------------------------------------------------------
\491\ To achieve ROLL10, the tire rolling resistance must be at
least 10 percent better than baseline (.0081 or better). To achieve
ROLL20, the tire rolling resistance must be at least 20 percent
better than baseline (.0072 or better).
---------------------------------------------------------------------------
We selected the industry-average RRC baseline of 0.009 based on a
CONTROLTEC study prepared for the California Air Resources Board,\492\
in addition to confidential business information submitted by
manufacturers prior to the 2018 NPRM analysis. The average RRC from the
CONTROLTEC study, which surveyed 1,358 vehicle models, was 0.009.\493\
CONTROLTEC also compared the findings of their survey with values
provided by Rubber Manufacturers Association (renamed USTMA-U.S. Tire
Manufacturers Association) for original equipment tires. The average
RRC from the data provided by RMA was 0.0092,\494\ compared to average
of 0.009 from CONTROLTEC.
---------------------------------------------------------------------------
\492\ Technical Analysis of Vehicle Load Reduction by CONTROLTEC
for California Air Resources Board (April 29, 2015).
\493\ The RRC values used in this study were a combination of
manufacturer information, estimates from coast down tests for some
vehicles, and application of tire RRC values across other vehicles
on the same platform.
\494\ Technical Analysis of Vehicle Load Reduction by CONTROLTEC
for California Air Resources Board (April 29, 2015) at page 40.
---------------------------------------------------------------------------
In past agency actions, commenters have argued that based on
available data on current vehicle models and the likely possibility
that there would be additional tire improvements over the next decade,
we should consider ROLL30 technology, or a 30 percent reduction of tire
rolling resistance over the baseline.\495\
---------------------------------------------------------------------------
\495\ Wesley Dyer, Docket No. NHTSA-2018-0067-11985, at p. 49.
---------------------------------------------------------------------------
As stated in the Joint TSD for the 2012 final rule for MY 2017-2025
and 2020 final rule, tire technologies that enable rolling resistance
improvements of 10 and 20 percent have been in existence for many
years.\496\ Achieving improvements of up to 20 percent involves
optimizing and integrating multiple technologies, with a primary
contributor being the adoption of a silica tread technology. Tire
suppliers have indicated that additional innovations are necessary to
achieve the next level of low rolling resistance technology on a
commercial basis, such as improvements in material to retain tire
pressure, and tread design to manage both stopping distance and wet
traction.\497\
---------------------------------------------------------------------------
\496\ EPA-420-R-12-901, at p. 3-210.
\497\ 2011 NAS Report, at p. 103.
---------------------------------------------------------------------------
The agency believes that the tire industry is in the process of
moving automotive manufacturers towards higher levels of rolling
resistance technology in the vehicle fleet. Importantly, as shown
below, the MY 2020 baseline fleet does include a higher percentage of
vehicles with ROLL20 technology than the MY 2017 fleet. However, we
believe that at this time, the emerging tire technologies that would
achieve 30 percent improvement in rolling resistance, like changing
tire profile, stiffening tire walls, or adopting improved tires along
with active chassis control,\498\ among other technologies, will not be
available for widespread commercial adoption in the fleet during the
rulemaking timeframe. As a result, we continue to not to incorporate 30
percent reduction in rolling resistance technology.
---------------------------------------------------------------------------
\498\ Mohammad Mehdi Davari, Rolling resistance and energy loss
in tyres (May 20, 2015), available at https://www.sveafordon.com/media/42060/SVEA-Presentation_Davari_public.pdf. Last visited
December 30, 2019.
---------------------------------------------------------------------------
USTMA agreed with this assessment, and commented that ``its members
will continue to develop advanced rolling resistance technologies for
future adoption, since vehicle manufacturers continue to prioritize
rolling resistance as one of the more cost-effective ways to achieve
advancements in vehicle fuel economy.'' \499\ Auto Innovators, in their
comments to both NHTSA and EPA, also discouraged the addition of 30
percent tire rolling resistance, stating that ``performance neutrality
for cold weather traction, hot weather performance, wet weather
traction, load handling (for addition weight of batteries, for
instance), wear and durability, and noise, vibration, and harshness can
be challenging to achieve for 20 [percent] tire rolling resistance
reduction, and the technology pathway to ROLL30 for many vehicles
remains unclear.'' \500\
---------------------------------------------------------------------------
\499\ USTMA, at 2.
\500\ Auto Innovators, Docket No. NHTSA-2021-0053-1492, at 134.
---------------------------------------------------------------------------
We will continue to monitor this issue and consider any additional
advancements in tire rolling resistance technology for future analyses.
(b) Tire Rolling Resistance Analysis Fleet Assignments
Tire rolling resistance is not a part of tire manufacturers'
publicly released specifications and thus it is difficult to assign
this technology to the analysis fleet. Manufacturers also often offer
multiple wheel and tire packages for the same nameplates, further
increasing the complexity of this assignment. We employed an approach
consistent with previous rulemaking in assigning this technology. We
relied on previously submitted rolling resistance values that were
supplied by manufacturers in the process of building older fleets and
bolstered it with agency-sponsored tire rolling resistance testing by
Smithers.\501\
---------------------------------------------------------------------------
\501\ See memo to Docket No. NHTSA-2021-0053, Evaluation of
Rolling Resistance and Wet Grip Performance of OEM Stock Tires
Obtained from NCAP Crash Tested Vehicles Phase One and Two. NHTSA
used tire rolling resistance coefficient values from this project to
assign baseline tire rolling resistance technology in the MY 2020
analysis fleet and is therefore providing the draft project
appendices for public review and comment.
---------------------------------------------------------------------------
We carried over rolling resistance assignments for nameplates where
manufacturers had submitted data on the vehicles' rolling resistance
values, even if the vehicle was redesigned. If Smithers data was
available, we replaced any older or missing values with that updated
data. Those vehicles for which no information was available from either
previous manufacturer submission or Smithers data were assigned to
ROLL0. All vehicles under the same nameplate were assigned the same
rolling resistance technology level even if manufacturers do outfit
different trim levels with different wheels and tires.
The MY 2020 analysis fleet includes the following breakdown of
rolling resistance technology: 44 percent at ROLL0, 20 percent at
ROLL10, and 36 percent at ROLL20, which shows that the majority of the
fleet has now adopted some form of improved rolling resistance
technology. The majority of the change from the MY 2017 analysis fleet
has been in implementing ROLL20 technology. There is likely more
proliferation of rolling resistance technology, but we would need
further information from manufacturers in order to account for it.
Accordingly, we made no changes to tire rolling
[[Page 25843]]
resistance assignments for this final rule.
(C) Tire Rolling Resistance Adoption Features
Rolling resistance technology can be adopted with either a vehicle
refresh or redesign. In some cases, low rolling resistance tires can
affect traction, which may adversely impact acceleration, braking, and
handling characteristics for some high-performance vehicles. Similar to
past rulemakings, the agency recognizes that to maintain performance,
braking, and handling functionality, some high-performance vehicles
would not adopt low rolling resistance tire technology. For cars and
SUVs with more than 405 horsepower (hp), the agency restricted the
application of ROLL20. For cars and SUVs with more than 500 hp, the
agency restricted the application of any additional rolling resistance
technology (ROLL10 or ROLL20). The agency developed these cutoffs based
on a review of confidential business information and the distribution
of rolling resistance values in the fleet. We received no comments on
these adoption features and made no changes for this final rule
analysis.
(d) Tire Rolling Resistance Effectiveness Modeling
As discussed above, the baseline rolling resistance value from
which rolling resistance improvements are measured is 0.009, based on a
thorough review of confidential business information submitted by
industry, and a review of other literature. To achieve ROLL10, the tire
rolling resistance must be at least 10 percent better than baseline
(.0081 or better). To achieve ROLL20, the tire rolling resistance must
be at least 20 percent better than baseline (.0072 or better).
We determined effectiveness values for rolling resistance
technology adoption using Autonomie. Figure III-17 below shows the
range of effectiveness values used for adding tire rolling resistance
technology to a vehicle in this analysis. The graph shows the change in
fuel consumption values between entire technology keys,\502\ and not
the individual technology effectiveness values. Using the change
between whole technology keys captures the complementary or non-
complementary interactions among technologies. In the graph, the box
shows the interquartile range (IQR) of the effectiveness values and
whiskers extend out 1.5 x IQR. The dots outside of the whiskers show
values for effectiveness that are outside these bounds.
---------------------------------------------------------------------------
\502\ Technology key is the unique collection of technologies
that constitutes a specific vehicle, see TSD Chapter 2.4.7 for more
information.
---------------------------------------------------------------------------
The data points with the highest effectiveness values are almost
all exclusively BEV and FCV technology combinations for medium sized
nonperformance cars. The effectiveness for these vehicles, when the low
rolling resistance technology is applied, is amplified by a
complementary effect, where the lower rolling resistance reduces road
load and allows a smaller battery pack to be used (and still meet range
requirements). The smaller battery pack reduces the overall weight of
the vehicle, further reducing road load, and improving fuel efficiency.
This complimentary effect is experienced by all the vehicle technology
classes, but the strongest effect is on the midsized vehicle non-
performance classes and is only captured in the analysis through the
use of full vehicle simulations, demonstrating the full interactions of
the technologies.
[[Page 25844]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.096
(e) Tire Rolling Resistance Costs
For this final rule analysis, we continue to use the same DMC
values for ROLL technology that were used for the 2020 final rule,
which are based on NHTSA's MY 2011 CAFE final rule and the 2006 NAS/NRC
report.\503\ Table III-29 shows the different levels of tire rolling
resistance technology cost for all vehicle classes across select model
years, which shows how the learning rate for ROLL technologies impacts
the cost. For all ROLL absolute technology costs used in the analysis
across all model years see the Technologies file.
---------------------------------------------------------------------------
\503\ ``Tires and Passenger Vehicle Fuel Economy,''
Transportation Research Board Special Report 286, National Research
Council of the National Academies, 2006, Docket No. EPA-HQ-OAR-2009-
0472-0146.
[GRAPHIC] [TIFF OMITTED] TR02MY22.097
7. Other Vehicle Technologies
We included four other vehicle technologies in the analysis--
electric power steering (EPS), improved accessory devices (IACC), low
drag brakes (LDB), and secondary axle disconnect (SAX). The CAFE Model
applied the effectiveness values for each of these technologies
directly, with unique effectiveness values for each technology and for
each technology class, rather than using Autonomie effectiveness
estimates. We used this methodology in these four cases because the
effectiveness of these technologies varies little with combinations of
other technologies. Also, applying these technologies directly in the
CAFE Model significantly reduces the required runtime of Autonomie
simulations.
(a) Electric Power Steering
Electric power steering reduces fuel consumption by reducing load
on the engine. Specifically, it reduces or
[[Page 25845]]
eliminates the parasitic losses associated with engine-driven power
steering pumps, which pump hydraulic fluid continuously through the
steering actuation system even when no steering input is present. By
selectively powering the electric assist only when steering input is
applied, the power consumption of the system is reduced in comparison
to the traditional ``always-on'' hydraulic steering system. Power
steering may be electrified on light duty vehicles with standard 12V
electrical systems and is also an enabler for vehicle electrification
because it provides power steering when the engine is off (or when no
combustion engine is present).
Power steering systems can be electrified in two ways.
Manufacturers may choose to eliminate the hydraulic portion of the
steering system and provide electric-only power steering (EPS) driven
by an independent electric motor, or they may choose to move the
hydraulic pump from a belt-driven configuration to a stand-alone
electrically driven hydraulic pump. The latter system is commonly
referred to as electro-hydraulic power steering (EHPS). As stated in
past rulemakings, manufacturers have told us that full EPS systems are
being developed for all types of light-duty vehicles, as well as large
trucks.
We described in past rulemakings that, like low drag brakes, EPS
can be difficult to observe and assign to the analysis fleet, however,
it is found more frequently in publicly available information than low
drag brakes. Based on comments received during the 2020 rulemaking, the
agency increased EPS application rate to nearly 90 percent for the 2020
final rule. The agency is maintaining this level of EPS fleet
penetration for this analysis, recognizing that some specialized,
unique vehicle types or configurations still implement hydraulically
actuated power steering systems for the baseline fleet model year.
The effectiveness of both EPS and EHPS is derived from the
decoupling of the pump from the crankshaft and is considered to be
practically the same for both. Thus, a single effectiveness value is
used for both EPS and EHPS. As indicated in the Table III-30, the
effectiveness of EPS and EHPS varies based on the vehicle technology
class it is being applied to. This variance is a direct result of
vehicle size and the amount of energy required to turn the vehicle's
two front wheels about their vertical axis. More simply put, more
energy is required for vehicles that weigh more and, typically, have
larger tire contact patches.
[GRAPHIC] [TIFF OMITTED] TR02MY22.098
(b) Improved Accessories
Engine accessories typically include the alternator, coolant pump,
cooling fan, and oil pump, and are traditionally mechanically driven
via belts, gears, or directly by other rotating engine components such
as camshafts or the crankshaft. These can be replaced with improved
accessories (IACC), which may include high efficiency alternators,
electrically driven (i.e., on-demand) coolant pumps, electric cooling
fans, variable geometry oil pumps, and a mild regeneration strategy.
Replacing lower-efficiency and/or mechanically driven components with
these improved accessories results in a reduction in fuel consumption,
as the improved accessories can conserve energy by being turned on/off
``on demand'' in some cases, driven at partial load as needed, or by
operating more efficiently.
For example, electric coolant pumps and electric powertrain cooling
fans provide better control of engine cooling. Flow from an electric
coolant pump can be varied, and the cooling fan can be shut off during
engine warm-up or cold ambient temperature conditions, reducing warm-up
time, fuel enrichment requirements, and ultimately reducing parasitic
losses.
IACC technology is difficult to observe and therefore there is
uncertainty in assigning it to the analysis fleet. As in the past, we
rely on industry-provided information and comments to assess the level
of IACC technology applied in the fleet. We believe there continues to
be opportunity for further implementation of IACC. The analysis has an
IACC fleet penetration of approximately eight percent compared to the
six percent value in the MY 2017 analysis fleet used for the 2020 final
rule analysis.
The agency believes improved accessories may be incorporated in
coordination with powertrain related changes occurring at either a
vehicle refresh or vehicle redesign. This coordination with powertrain
changes enables related design and tooling changes to be implemented
and systems development, functionality and durability testing to be
conducted in a single product change program to efficiently manage
resources and costs.
This analysis carries forward work on the effectiveness of IACC
systems conducted in the Draft TAR and EPA Proposed Determination that
is originally founded in the 2002 NAS Report \504\ and confidential
manufacturer data. This work involved gathering information by
monitoring press reports, holding meetings with suppliers and OEMs, and
attending industry technical conferences. The
[[Page 25846]]
resulting effectiveness estimates we use are shown in Table III-31. As
indicated in this table, the effectiveness values of IACC varies based
on the vehicle technology class it is being applied to. This variance,
like EPS, is a direct result of vehicle size as well as the amount of
energy generated by the alternator, the size of the coolant pump to the
cool the necessary systems, the size of the cooling fan required, among
other characteristics and it directed related to a vehicle size and
mass.
---------------------------------------------------------------------------
\504\ National Research Council 2002. Effectiveness and Impact
of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC:
The National Academies Press. https://doi.org/10.17226/10172.
[GRAPHIC] [TIFF OMITTED] TR02MY22.099
(c) Low Drag Brakes
We have defined low drag brakes (LDB) as brakes that reduce the
sliding friction of disc brake pads on rotors when the brakes are not
engaged because the brake pads are pulled away from the rotating disc
either by mechanical or electric methods since 2009 for the MY 2011
CAFE rule.\505\ At that time, we estimated the effectiveness of LDB
technology to be a range from 0.5-1.0 percent, based on CBI data. We
applied a learning curve to the estimated cost for LDB, but noted that
the technology was considered high volume, mature, and stable.
Confidential manufacturer comments in response to the NPRM for MY 2011
(73 FR 24352, May 2, 2008) indicated that most passenger cars have
already adopted LDB technology, but ladder frame trucks have not.
---------------------------------------------------------------------------
\505\ Final Regulatory Impact Analysis, Corporate Average Fuel
Economy for MY 2011 Passenger Cars and Light Trucks (March 2009), at
V-135.
---------------------------------------------------------------------------
We and EPA used the same definition for LDB in the MY 2012-2016
joint rule, with an estimated effectiveness of up to 1 percent based on
CBI data.\506\ We only allowed LDB technology to be applied to large
car, minivan, medium and large truck, and SUV classes because the
agency determined the technology was already largely utilized in most
other subclasses. The 2011 NAS committee also utilized our definition
for LDB and added that most new vehicles have low-drag brakes.\507\ The
committee confirmed that the impact over conventional brakes may be
about a 1 percent reduction of fuel consumption.
---------------------------------------------------------------------------
\506\ Final Regulatory Impact Analysis, Corporate Average Fuel
Economy for MY 2012-MY 2016 Passenger Cars and Light Trucks (March
2010), at 249.
\507\ 2011 NAS Report, at 103-104.
---------------------------------------------------------------------------
For the 2012 final rule for MY 2017-2025, however, we and EPA
updated the effectiveness estimate for LDB to 0.8 percent based on a
2011 Ricardo study and updated lumped-parameter model.\508\ The
agencies considered LDB technology to be off the learning curve (i.e.,
the DMC does not change year-over-year). The 2015 NAS Report continued
to use the agencies' definition for LDB and commented that the 0.8
percent effectiveness estimate is a reasonable estimate.\509\ The 2015
NAS committee did not opine on the application of LDB technology in the
fleet. The agencies used the same definition, cost, and effectiveness
estimates for LDB in the Draft TAR, but also noted the existence of
zero drag brake systems which use electrical actuators that allow brake
pads to move farther away from the rotor.\510\ However, the agencies
did not include zero drag brake technology in either compliance
simulation. EPA continued with this approach in its first 2017 Proposed
Determination that the standards through 2025 were appropriate.\511\
---------------------------------------------------------------------------
\508\ Joint Technical Support Document: Final Rulemaking for
2017-2025 Light-Duty Vehicle Greenhouse Gas Emission Standards and
Corporate Average Fuel Economy Standards (August 2012), at 3-211.
\509\ 2015 NAS Report, at 231.
\510\ Draft TAR, at 5-207.
\511\ EPA Proposed Determination TSD, at 2-422.
---------------------------------------------------------------------------
In the 2020 final rule, the agencies applied LDB sparingly in the
MY 2017 analysis fleet using the same cost and effectiveness estimates
from the 2011 Ricardo study, with approximately less than 15 percent of
vehicles being assigned the technology. In addition, we noted the
existence of zero drag brakes in production for some BEVs, similar to
the summary in the Draft TAR, but did not opine on the existence of
zero drag brakes in the fleet. Some stakeholders commented to the 2020
rule that other vehicle technologies, including LDB, were actually
overapplied in the analysis fleet.
For this analysis, we considered the conflicting statements that
LDB were both universally applied in new vehicles and that the new
vehicle fleet still had space to improve LDB technology. We determined
that LDB technology as previously defined going back to the MY 2011
rule (73 FR 24352, May 2, 2008) was universally applied in the MY 2020
fleet. However, we determined that zero drag brakes, the next level of
brake technology, was sparingly applied in the MY 2020 analysis fleet.
Currently, we do not believe that zero drag brake systems will be
available for wide scale application in the rulemaking timeframe and we
did not include it as a technology for this analysis. We sought comment
on the issue, including any data on the use
[[Page 25847]]
advanced LDB systems on current and forthcoming production vehicles,
but did not receive any comments. We will consider how to define a new
level of low drag brake technology that either encompasses the
definition of zero drag brakes or similar technology in future
rulemakings.
(d) Secondary Axle Disconnect
AWD and 4WD vehicles provide improved traction by delivering torque
to the front and rear axles, rather than just one axle. When a second
axle is rotating, it tends to consume more energy because of additional
losses related to lubricant churning, seal friction, bearing friction,
and gear train inefficiencies.\512\ Some of these losses may be reduced
by providing a secondary axle disconnect function that disconnects one
of the axles when driving conditions do not call for torque to be
delivered to both.
---------------------------------------------------------------------------
\512\ Pilot Systems, ``AWD Component Analysis,'' Project Report,
performed for Transport Canada, Contract T8080-150132, May 31, 2016.
---------------------------------------------------------------------------
The terms AWD and 4WD are often used interchangeably, although they
have also developed a colloquial distinction, and are two separate
systems. The term AWD has come to be associated with light-duty
passenger vehicles providing variable operation of one or both axles on
ordinary roads. The term 4WD is often associated with larger truck-
based vehicle platforms providing a locked driveline configuration and/
or a low range gearing meant primarily for off-road use.
Many 4WD vehicles provide for a single-axle (or two-wheel) drive
mode that may be manually selected by the user. In this mode, a primary
axle (usually the rear axle) will be powered, while the other axle
(known as the secondary axle) is not. However, even though the
secondary axle and associated driveline components are not receiving
engine power, they are still connected to the non-driven wheels and
will rotate when the vehicle is in motion. This unnecessary rotation
consumes energy,\513\ and leads to increased fuel consumption that
could be avoided if the secondary axle components were completely
disconnected and not rotating.
---------------------------------------------------------------------------
\513\ Any time a drivetrain component spins it consumes some
energy, primarily to overcome frictional forces.
---------------------------------------------------------------------------
Light-duty AWD systems are often designed to divide variably torque
between the front and rear axles in normal driving to optimize traction
and handling in response to driving conditions. However, even when the
secondary axle is not necessary for enhanced traction or handling, in
traditional AWD systems it typically remains engaged with the driveline
and continues to generate losses that could be avoided if the axle was
instead disconnected. The SAX technology observed in the marketplace
disengages one axle (typically the rear axle) for 2WD operation but
detects changes in driving conditions and automatically engages AWD
mode when it is necessary. The operation in 2WD can result in reduced
fuel consumption. For example, Chrysler has estimated the secondary
axle disconnect feature in the Jeep Cherokee reduces friction and drag
attributable to the secondary axle by 80 percent when in disconnect
mode.\514\
---------------------------------------------------------------------------
\514\ Brooke, L. ``Systems Engineering a new 4x4 benchmark'',
SAE Automotive Engineering, June 2, 2014.
---------------------------------------------------------------------------
Observing SAX technology on actual vehicles is very difficult.
Manufacturers do not typically identify the technology on technical
specifications or other widely available information. We employed an
approach consistent with previous rulemaking in assigning this
technology. Specifically, we assigned SAX technology based on a
combination of publicly available information and previously submitted
confidential information. In the analysis fleet, 38 percent of the
vehicles that had AWD or 4WD are determined to have SAX technology. All
vehicles in the analysis fleet with FWD or RWD have SAX skipped since
SAX technology is a way to emulate FWD or RWD in AWD and 4WD vehicles,
respectively. We did not allow for the application of SAX technology to
FWD or RWD vehicles because they do not have a secondary driven axle to
disconnect.
SAX technology can be adopted by any vehicle in the analysis fleet,
including those with a HEV or BEV powertrain,\515\ which was identified
as having AWD or 4WD. It does not supersede any technology or result in
any other technology being excluded for future implementation for that
vehicle. SAX technology can be applied during any refresh or redesign.
---------------------------------------------------------------------------
\515\ The inefficiencies addressed on ICEs by SAX technology may
not be similar enough, or even present, in HEVs or BEVs.
---------------------------------------------------------------------------
This analysis carries forward work on the effectiveness of SAX
systems conducted in the Draft TAR and EPA Proposed Determination.\516\
This work involved gathering information by monitoring press reports,
holding meetings with suppliers and OEMs, and attending industry
technical conferences. We did not simulate SAX effectiveness in the
Autonomie modeling because, similar to LDB, IACC, and EFR, the fuel
economy benefits from the technology are not fully captured on the two-
cycle test. The secondary axle disconnect effectiveness values, for the
most part, have been accepted as plausible based on the rulemaking
record and absence of contrary comments. As such, the agency has
prioritized its extensive Autonomie vehicle simulation work toward
other technologies that are emerging or considered more critical for
total system effectiveness. Table III-32 shows the resulting
effectiveness estimates we used in this analysis.
---------------------------------------------------------------------------
\516\ Draft TAR, at 5-412; Proposed Determination TSD, at 2-422.
---------------------------------------------------------------------------
[[Page 25848]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.100
[[Page 25849]]
(e) Other Vehicle Technology Costs
The cost estimates for EPS, IACC, SAX, and LDB \517\ rely on
previous work published as part of past rulemakings with learning
applied to those cost values which is founded in the 2002 NAS
Report.\518\ The cost values are the same values that were used for the
Draft TAR and 2020 final rule, updated to 2018 dollars. Table III-33
shows examples of costs for these technologies across select model
years. Note that these costs are the same for all vehicle technology
classes. For all absolute EPS, IACC, LDB, and SAX technology costs
across all model years, see the Technologies file.
---------------------------------------------------------------------------
\517\ Note that because LDB technology is applied universally as
a baseline technology in the MY 2020 fleet, there is functionally
zero costs for this technology associated with this rulemaking.
\518\ National Research Council 2002. Effectiveness and Impact
of Corporate Average Fuel Economy (CAFE) Standards. Washington, DC:
The National Academies Press. https://doi.org/10.17226/10172.
[GRAPHIC] [TIFF OMITTED] TR02MY22.101
8. Simulating Air Conditioning Efficiency and Off-Cycle Technologies
Off-cycle and air conditioning (AC) efficiency technologies can
provide fuel economy benefits in real-world vehicle operation, but
those benefits cannot be fully captured by the traditional 2-cycle test
procedures used to measure fuel economy.\519\ Off-cycle technologies
include technologies like high efficiency alternators and high
efficiency exterior lighting.\520\ AC efficiency technologies are
technologies that reduce the operation of or the loads on the
compressor, which pressurizes AC refrigerant. The less the compressor
operates or the more efficiently it operates, the less parasitic load
the compressor places on the engine, resulting in better fuel
efficiency.
---------------------------------------------------------------------------
\519\ See 49 U.S.C. 32904(c) (``The Administrator shall measure
fuel economy for each model and calculate average fuel economy for a
manufacturer under testing and calculation procedures prescribed by
the Administrator . . . . the Administrator shall use the same
procedures for passenger automobiles the Administrator used for
model year 1975 (weighted 55 percent urban cycle and 45 percent
highway cycle), or procedures that give comparable results.'').
\520\ 40 CFR 86.1869-12(b)--Credit available for certain off-
cycle technologies.
---------------------------------------------------------------------------
Vehicle manufacturers have the option to generate credits for off-
cycle technologies and improved AC systems under the EPA's CO2 program
and receive an FCIV equal to the value of the benefit not captured on
the 2-cycle test under NHTSA's CAFE program. The FCIV is not a
``credit'' in the NHTSA CAFE program,\521\ but the FCIVs increase the
reported fuel economy of a manufacturer's fleet, which is used to
determine compliance. EPA applies FCIVs during determination of a
fleet's final average fuel economy reported to NHTSA.\522\ In the CAFE
Model, we only calculate and apply FCIVs at a fleet level for a
manufacturer based on the volume of the manufacturer's fleet that
contain qualifying technologies.\523\
---------------------------------------------------------------------------
\521\ Unlike, for example, the statutory overcompliance credits
prescribed in 49 U.S.C. 32903.
\522\ 49 U.S.C. 32904(c)-(e). EPCA granted EPA authority to
establish fuel economy testing and calculation procedures. See
Section VII for more information.
\523\ 40 CFR 600.510-12(c).
---------------------------------------------------------------------------
There are three pathways that manufacturers can use to determine
the value of AC efficiency and off-cycle adjustments. First,
manufacturers can use a predetermined list or ``menu'' of g/mi values
that EPA established for specific off-cycle technologies.\524\ Second,
manufacturers can use 5-cycle testing to demonstrate off-cycle CO2
benefit; \525\ the additional tests allow emissions benefits to be
demonstrated over some elements of real-world driving not captured by
the 2-cycle compliance tests, including high speeds, rapid
accelerations, hot temperatures, and cold temperatures. Third,
manufacturers can seek EPA approval, through a notice and comment
process, to use an alternative methodology other than the menu or 5-
cycle methodology for determining the off-cycle technology improvement
values.\526\ For further discussion of the AC and off-cycle compliance
and application process, see Section VII.
---------------------------------------------------------------------------
\524\ See 40 CFR 86.1869-12(b). The TSD for the 2012 final rule
for MYs 2017 and beyond provides technology examples and guidance
with respect to the potential pathways to achieve the desired
physical impact of a specific off-cycle technology from the menu and
provides the foundation for the analysis justifying the credits
provided by the menu. The expectation is that manufacturers will use
the information in the TSD to design and implement off-cycle
technologies that meet or exceed those expectations in order to
achieve the real-world benefits of off-cycle technologies from the
menu.
\525\ See 40 CFR 86.1869-12(c). EPA proposed a correction for
the 5-cycle pathway in a separate technical amendments rulemaking.
See 83 FR 49344 (Oct. 1, 2019). EPA is not approving credits based
on the 5-cycle pathway pending the finalization of the technical
amendments rule.
\526\ See 40 CFR 86.1869-12(d).
---------------------------------------------------------------------------
We and EPA have been collecting data on the application of these
technologies since implementing the AC and off-cycle programs.\527\
\528\ Most manufacturers are applying AC efficiency and off-cycle
technologies; in MY 2020, 17 manufacturers employed AC efficiency
technologies and 20 manufacturers employed off-cycle technologies,
though the level of deployment varies by manufacturer.\529\
---------------------------------------------------------------------------
\527\ See 77 FR 62832, 62839 (Oct. 15, 2012). EPA introduced AC
and off-cycle technology credits for the CO2 program in
the MYs 2012-2016 rule (75 FR 25324, May 7, 2010) and revised the
program in the MY 2017-2025 rule (77 FR 62624, Oct. 15, 2012) and
NHTSA adopted equivalent provisions for MYs 2017 and later in the MY
2017-2025 rule.
\528\ Vehicle and Engine Certification. Compliance Information
for Light-Duty Gas (GHG) Standards. Compliance Information for
Light-Duty Greenhouse Gas (GHG) Standards [bond] Certification and
Compliance for Vehicles and Engines [bond] U.S. EPA. Last accessed
December 22, 2021.
\529\ See 2021 EPA Automotive Trends Report, at 90 and 92.
---------------------------------------------------------------------------
Manufacturers have only recently begun including detailed
information on off-cycle and AC efficiency technologies equipped on
vehicles in compliance reporting data. For this analysis, though, such
information was not sufficiently complete to support a detailed
representation of the application of off-cycle technology to specific
vehicle
[[Page 25850]]
model/configurations in the MY 2020 fleet. To account for the AC and
off-cycle technologies equipped on vehicles and the potential that
manufacturers will apply additional AC and off-cycle technologies in
the rulemaking timeframe, we specify CAFE Model inputs for AC
efficiency and off-cycle FCIVs in grams/mile for each manufacturer's
fleet in each model year. We estimate future potential AC efficiency
and off-cycle technology application in the CAFE analyses based on an
expectation that manufacturers already relying heavily on these
adjustments would continue do so, and that other manufacturers would,
over time, also approach the limits on adjustments allowed for such
improvements.
The next sections discuss how the CAFE Model simulates the
effectiveness and cost for AC efficiency and off-cycle technology
adjustments.
(a) AC and Off-Cycle Effectiveness Modeling in the CAFE Model
In this analysis, the CAFE Model applies AC and off-cycle
flexibilities to manufacturer's CAFE regulatory fleet performance in a
similar way to the regulation.\530\ As the CAFE Model simulates the
addition of technology to vehicles in a given model year fleet, the
model first applies conventional technologies to vehicles in an attempt
to meet a given standard, and then applies AC efficiency and off-cycle
FCIVs to each regulatory fleet. In other words, first the CAFE Model
applies conventional technologies to each manufacturers' vehicles in
each model year to assess the 2-cycle sales weighted harmonic average
CAFE rating. Then, the CAFE Model assesses the CAFE rating to use for a
manufacturer's compliance value after applying the AC efficiency and
off-cycle FCIVs designated in the Market Data file. The CAFE Model does
this on a year-by-year basis. The CAFE Model attempts to apply
technologies and FCIVs in a way that both minimizes cost and allows the
manufacturer to meet their standards without over or under complying.
---------------------------------------------------------------------------
\530\ 49 CFR 531.6 and 49 CFR 533.6 Measurement and Calculation
procedures.
---------------------------------------------------------------------------
To determine how manufacturers might adopt AC efficiency and off-
cycle technologies in the rulemaking timeframe, we use data from EPA's
2021 Trends Report for MY 2020 and CBI compliance material from
manufacturers.\531\ \532\ We use manufacturer's MY 2020 AC efficiency
and off-cycle FCIVs as a starting point, and then extrapolate values in
each model year until MY 2026, for light trucks to the proposed
regulatory cap, for each manufacturer's fleets by regulatory class.
---------------------------------------------------------------------------
\531\ Vehicle and Engine Certification. Compliance Information
for Light-Duty Gas (GHG) Standards. Compliance Information for
Light-Duty Greenhouse Gas (GHG) Standards [verbar] Certification and
Compliance for Vehicles and Engines [verbar] U.S. EPA. Last accessed
May 24, 2021.
\532\ 49 U.S.C. 32907.
---------------------------------------------------------------------------
To determine the rate at which to extrapolate the addition of AC
and off-cycle technology adoption for each manufacturer, we use
historic AC and off-cycle technology applications, each manufacturer's
fleet composition (i.e., breakdown between passenger cars (PCs) and
light trucks (LTs)), availability of AC and off-cycle technologies that
manufacturers could still use, and CBI compliance data. Different
manufacturers show different levels of historical AC efficiency and
off-cycle technology adoption; therefore, different manufacturers hit
the proposed regulatory caps for AC efficiency technology for both
their PC and LT fleets, and different manufacturers hit caps for off-
cycle technologies in the LT regulatory class. We do not extrapolate
off-cycle technology adoption for PCs to the proposed regulatory cap
for a few reasons. First, past EPA Trends Reports showed that many
manufacturers did not adopt off-cycle technology to their passenger car
fleets. Next, manufacturers limited PC offerings in MY 2020 as compared
to historical trends. Last, available CBI compliance data indicated
that PCs adopt a lower level of menu item off-cycle technologies than
LTs. We accordingly limit the application of off-cycle FCIVs to 10 g/mi
for PCs but allow LTs to apply 15 g/mi of off-cycle FCIVs starting in
MY 2023 for the final rule analysis. This decision also aligns with
EPA's treatment of off-cycle adjustments in its final rule. The inputs
for AC efficiency technologies are set to 5 g/mi and 7.2 g/mi for PCs
and LTs, respectively. We allow AC efficiency technologies to reach the
regulatory caps by MY 2024, which is the first year of standards
assessed in this analysis.
We apply FCIVs in this way because the AC and off-cycle
technologies are generally more cost-effective than other technologies.
The details of this assessment (and the calculation) are further
discussed in the CAFE Model Documentation.\533\ The AC efficiency and
off-cycle adjustment schedules used in this analysis are shown in TSD
Chapter 3.8 and in the Market Data file's Credits and Adjustments
worksheet. Like the NPRM, for this final rule analysis we did not allow
some manufacturers to reach the AC efficiency and off-cycle caps to
avoid over compliance in the rulemaking time frame. Table III-34 and
Table III-35 show the average FCIVs applied to the regulatory fleets
for the final rule analysis.
---------------------------------------------------------------------------
\533\ CAFE Model Documentation, S5.
[GRAPHIC] [TIFF OMITTED] TR02MY22.102
[[Page 25851]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.103
We received limited comments on how we model off-cycle and AC
efficiency for this rulemaking analysis. Auto Innovators stated that
``due to the static nature of the forecasts and input structure, the
NHTSA forecasts on the quantity of off-cycle credits do not vary by
scenario, and this creates material distortions in the model outputs.
For instance, the projected Central case adoption of off-cycle
technologies may contribute to over-compliance with some scenarios,
especially low stringency scenarios.'' \534\ On the other hand, UCS
stated that ``NHTSA has not acknowledge that its [CAFE Model] does not
consider increased adoption of off-cycle technology to yield any real-
world benefit . . . there is supportive evidence of their real-world
benefits, and at any rate NHTSA must state explicitly its rationale for
excluding these technologies from the benefits of the rule, as the
credits associated with these technologies represent a substantial
share of the credits accrued for compliance by manufacturers.'' UCS
also stated that ``NHTSA should correct the [CAFE Model] to ensure it
adjusts a vehicle's fuel economy to account for reductions in emissions
and fuel use from off-cycle technologies, which will yield a more
accurate accounting of the benefits from the CAFE program.'' \535\
---------------------------------------------------------------------------
\534\ Auto Innovators, Docket No. NHTSA-2021-0053-0021 Appendix
VII, at 125-126.
\535\ UCS, Docket No. NHTSA-2021-0053-1567, at 31.
---------------------------------------------------------------------------
In response to comments from Auto Innovators, we agree that, in
theory, the way the CAFE Model is set up to apply off-cycle benefits
statically could create overcompliance for some manufacturers. However,
as discussed earlier and in TSD Chapter 3.8, we apply off-cycle and
other flexibilities differently for each manufacturer rather than apply
adjustments consistently to the cap for each manufacturer. For example,
if a manufacturer is on a trajectory to reach the off-cycle regulatory
cap, then we allow the model to reach that cap regardless of
alternatives. On the other hand, if a manufacturer has historically
lagged in the adoption of off-cycle technology, we use this historic
rate of application through the rulemaking time frame. As shown in
Table III-34 and Table III-35, on average, the fleet does not reach the
regulatory caps based on our extrapolation.
We understand UCS's concern, that because the CAFE Model accounts
for off-cycle technology at the fleet level, the benefits do not
directly appear in the vehicle-level benefits analysis. Although
further refinement may be possible for future analyses, at this time
there are only limited vehicle-level data available. We agree that some
manufacturers have relied on these flexibilities more so than others,
but as indicated by the 2021 EPA Trends Report many are still lagging
in adopting these technologies.\536\ This is one reason why we declined
to apply off-cycle benefits up to the cap for each vehicle to have
those benefits automatically count in the benefits calculations. Based
on the ratio of benefits that manufacturers can expect from on-cycle
versus off-cycle technology, we believe that the small off-cycle
technology benefit that is not accounted for in the benefits
calculations does not make a material difference to the analysis.
---------------------------------------------------------------------------
\536\ 2021 EPA Trends Report at 104-106.
---------------------------------------------------------------------------
For the final rule analysis, we updated the baseline fleet off-
cycle data to reflect the 2021 EPA Trends Report, using the same
modeling methodology as the NPRM. We believe that this approach is
appropriate to capture the costs and benefits of off-cycle
technologies.
(b) AC and Off-Cycle Costs
For this analysis, AC and off-cycle technologies are applied
independently of the decision trees using the extrapolated values shown
above, so it is necessary to account for the costs of those
technologies independently. Table III-36 shows the costs used for AC
and off-cycle FCIVs in this analysis. The costs are shown in dollars
per gram of CO2 per mile ($ per g/mile). The AC efficiency and off-
cycle technology costs are the same costs used in the EPA Proposed
Determination and described in the EPA Proposed Determination TSD.\537\
---------------------------------------------------------------------------
\537\ EPA PD TSD. EPA-420-R-16-021. November 2016. At 2-423-2-
245. https://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=P100Q3L4.pdf. Last
accessed May 24, 2021.
---------------------------------------------------------------------------
To develop the off-cycle technology costs, we selected the second
generic 3 g/mile package estimated to cost $170 (in 2015$) to apply in
this analysis in $ per g/mile. We updated the costs used in the
Proposed Determination TSD from 2015$ to 2018$, adjusted the costs for
RPE, and applied a relatively flat learning rate.
Similar to off-cycle technology costs, we used the cost estimates
from EPA Proposed Determination TSD for AC efficiency technologies that
relied on the 2012 rulemaking TSD.\538\ We updated these costs to 2018$
and adjusted for RPE for this analysis and applied the same mature
learning rate that we had applied for off-cycle technologies.
---------------------------------------------------------------------------
\538\ Joint NHTSA and EPA 2012 TSD, see Section 5.1.
---------------------------------------------------------------------------
[[Page 25852]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.104
In the NPRM we sought comment on whether our costs were appropriate
or if other costs should be used. Overall, comments from UCS, Consumer
Reports, and ICCT stated that our costs for off-cycle technologies were
high.\539\ Consumer Reports indicated that they did not investigate the
NHTSA approach to AC and off-cycle adjustments and costs. However
Consumer Reports did find ``that under the EPA proposal the use of
similar costs for off-cycle technologies resulted in compliance costs
for those technologies that were more than three times the average
compliance costs of all the technology applied to achieve the Preferred
Alternative.'' \540\ ICCT stated that ``the agencies use an arbitrarily
and unrealistically high estimate of off-cycle credit cost in their
compliance modeling.'' \541\ UCS conducted an analysis of off-cycle
costs using the 2020 final rule's CAFE Model and data from the 2021 NAS
Report to show that the average costs could be different if the
agencies used different inputs.\542\ This approach is similar to the
one used by EPA in the final rule for MYs 2023-2026 in determining the
costs of off-cycle.
---------------------------------------------------------------------------
\539\ Consumer Reports, Docket No. NHTSA-2021-0053-1576, at 22;
UCS, at 30; ICCT, Docket No. NHTSA-2021-0053-1581, at 8.
\540\ Consumer Reports, at 22-23.
\541\ ICCT, at 8.
\542\ UCS, at 30.
---------------------------------------------------------------------------
As we discussed in the NPRM and explained again above, the CAFE
Model was updated from the 2020 final rule model to better account for
costs of AC and Off-Cycle technologies.543 544 This update
fixed many of the issues highlighted by the commenters by baking in the
costs per vehicle of the off-cycle technology in the baseline vehicle
and excluding the costs from affecting the new vehicle model output
costs. The CAFE Model used by EPA in their rulemaking analysis for MYs
2023-2026 did not have this feature, and they were required to re-
evaluate the costs as described in the EPA Regulatory Impacts
Analysis.\545\
---------------------------------------------------------------------------
\543\ 86 FR 49605 (Sept. 3, 2021).
\544\ ``More accurate accounting for off-cycle incremental costs
relative to MY 2020 baseline fleet.''
\545\ EPA Final Rule for MYs 2023-2026 RIA, Chapter 4.1.1.1,
Off-Cycle Credit Cost and changes since the Proposed Rule, at p. 4-
6.
---------------------------------------------------------------------------
Separately, none of these commenters provided alternative AC and
off-cycle technology costs in response to our request that commenters
provide any data or information on which any alternative costs are
based on. General statements that costs should be lower, without
specific data and analysis to support those statements, are not enough
to justify a change from the NPRM values. As one example, the 2021 NAS
Report observed an AC efficiency technology similar to one used by
Toyota, and they estimated the cost of that technology to be $170 in
2025.546 547 However, that was not enough information for us
to update our gram per mile cost for all technologies. We will continue
to research this issue for future analyses.
---------------------------------------------------------------------------
\546\ 2021 NAS Report, at 68.
\547\ EPA Decision Document. ``Off-Cycle Credits for Toyota
Motor North America.'' EPA-420-R-21-024. October 2021. https://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=P1013CFF.pdf. (Accessed: March
15, 2022)
---------------------------------------------------------------------------
E. Consumer Responses to Manufacturer Compliance Strategies
The previous subsections in Section III have so far discussed how
manufacturers might respond to changes to the standards. While the
technology analysis is informative of the different compliance
strategies available to manufactures, the tangible costs and benefits
that accrue because of CAFE standards also depend on how consumers
respond to the decisions made by manufacturers. Many of the benefits
and costs resulting from changes to CAFE standards are private benefits
that accrue to the buyers of new cars and trucks produced in the model
years subject to this rulemaking. These benefits and costs largely flow
from the changes to vehicle purchases, ownership, and operating costs
that result from improved fuel economy, as well as from the costs of
the technology required to achieve those improvements. In addition,
buyers' and owners' decisions about the use of their vehicles can
impose costs or create benefits that fall on others, which the agency
refers to as ``external'' costs or benefits. The following subsections
describe how NHTSA's analyzes consumer responses to changing vehicles
and prices.
1. Assumptions About Macroeconomic Conditions and Consumer Behavior
This final rule includes a comprehensive economic analysis of the
impacts of establishing more stringent CAFE standards, and most of the
effects it measures are influenced by future macroeconomic conditions
that are beyond the agency's influence. For example, domestic fuel
prices are mainly determined by global petroleum supply and demand as
well as refining costs, yet they determine how much technology
manufacturers will employ to improve the fuel economy of cars and light
trucks produced for the U.S. market, how much consumers are willing to
pay for new vehicles offering different levels of fuel economy, how
much new and used cars and light trucks will be driven, and the value
of each gallon saved through higher CAFE standards. Similarly,
projecting sales of new cars and light trucks produced during the model
years subject to the standards this final rule establishes requires
robust projections of demographic and macroeconomic variables that span
the entire timeframe of the analysis, including U.S. population, Gross
Domestic Product (GDP), consumer confidence about future economic
conditions, and disposable personal income.
To ensure internal consistency within the agency's analysis,
projections of most of the economic variables used in our analysis are
obtained from the same source. The analysis presented here relies on
forecasts of fuel prices issued by the U.S. Energy Information
Administration (EIA), an agency within the DOE that collects, analyzes,
and disseminates independent and impartial energy data and forecasts to
promote sound policymaking, efficient markets, and public understanding
of energy and its interaction with the economy and the environment. EIA
uses its National Energy Model System (NEMS) to produce its Annual
Energy Outlook
[[Page 25853]]
(AEO), which includes forecasts of future U.S. macroeconomic growth and
fuel prices among many other energy-related variables. NHTSA's main
analysis uses forecasts of fuel prices, from the AEO 2021 Reference
Case. The agency also uses forecasts of the U.S. population, the number
of U.S. households, the Nation's Gross Domestic product (GDP),
disposable personal income, and consumer confidence to develop its
projections of new car and light truck sales as well as of total light-
duty vehicle travel. For the current analysis, NHTSA obtained forecasts
of these variables from the IHS Markit Global Insight October 2021
Macroeconomic Outlook base case, which represents the most likely
scenario from that organization's most current forecast. EIA also
relies on the IHS Markit Global Insight Macroeconomic Outlook to
develop the macroeconomic and energy price forecasts included as part
of its Annual Energy Outlook. However, the forecasts EIA presents in
its Annual Energy Outlook 2021 are based on the IHS Markit Global
Insight March 2021 Macroeconomic Outlook, rather than the more recent
October 2021 Outlook the agency relies on in this analysis. Because the
forecasts of population, GDP, disposable income, and other variables in
the March 2021 and October 2021 Macroeconomic Outlooks are very
similar, the forecasts the agency relies on in this analysis are
generally consistent with those reported in EIA's AEO 2021. TSD Chapter
4.1 includes a more complete discussion of the macroeconomic
assumptions made for the analysis.
While these macroeconomic assumptions are some of the most critical
inputs to the analysis, they are also subject to the most uncertainty--
particularly over the lifetimes of the vehicles subject to this final
rule, which can extend as far as forty years into the future. The
agency also uses low and high economic growth and global oil price
forecasts issued by EIA as part of its Annual Energy Outlook as
alternative cases in its sensitivity analyses. The purpose of these
sensitivity analyses, which are discussed in greater detail in FRIA
Chapters 6 and 7, is not to posit a more credible future state of the
world than the central case, which the agency assumes represents the
most likely future state of the world. Instead, the sensitivity
analyses are intended to illustrate the degree to which important
future outcomes resulting from this final rule might change under
different assumptions about fuel prices, economic growth, and other
factors.
The agency received several comments about the macroeconomic
assumptions used in the analysis. Auto Innovators correctly noted that
fuel prices will influence the adoption of advanced technologies and
the cost and benefits realized under the new standards, and commented
that EIA's projections may overestimate fuel prices. In support of its
claim, Auto Innovators notes that EIA's projections have historically
overestimated fuel prices and speculates that the current forecasts
could overestimate domestic demand if the ``EIA Central Case gasoline
forecast assumes fewer than 50 [percent] plug-in vehicles by 2030.''
\548\ In that event, Auto Innovators recommended that NHTSA instead
rely on the IHS Markit Global Insight forecast of fuel prices
throughout its main analysis, which as its comment showed falls
considerably below the AEO 2021 Reference Case forecast after about the
year 2030. Auto Innovators recognized that NHTSA does use the Global
Insight forecast it recommended for the purpose of sensitivity analysis
but encouraged the agency to feature it more prominently.
---------------------------------------------------------------------------
\548\ Auto Innovators, Docket No. NHTSA-2021-0053-0021, at 58-
59. The AEO 2021 Reference Case forecasts that less than 2 percent
of new car and light truck sales during 2030 will be plug-in hybrid
models and including projected sales of conventional hybrid models
increases that figure to somewhat more than 6 percent.
---------------------------------------------------------------------------
In contrast, Consumer Reports asserted that the AEO 2021
projections underestimated how quickly fuel prices would rebound from
the diminished demand caused by onset of COVID-19. Consumer Reports
suggested that the agency use the AEO 2020 reference case instead of
that from AEO 2021 to avoid the potential for fuel prices from calendar
year 2020 to unduly influence the rest of the analysis.\549\ As
discussed earlier, projections are inherently uncertain and actual
prices are likely to deviate from those forecast for any given future
year, and the accuracy of a multi-year forecast should not be judged by
its ability to predict the value realized in a single period. In any
case, the agency determined that the AEO 2021 projections of fuel
prices were more appropriate for this analysis, because they
incorporate the potential long-term impacts of the COVID-19 pandemic
and its effects on travel activity, gasoline demand, and future fuel
prices.\550\
---------------------------------------------------------------------------
\549\ Consumer Reports, Comment Body, Docket No. NHTSA-2021-
0053-1576, at 23.
\550\ EIA reports that actual retail gasoline prices during 2021
averaged $3.10 per gallon, considerably higher than the $2.36
average price projected for 2021 as part of AEO 2021. While part of
this discrepancy probably owes to an overly cautious view of how
rapidly global demand for petroleum products would return to its
pre-pandemic level, other unforeseen factors apparently contributed
as well. This is evidenced by the fact that actual gasoline prices
during 2021 were well above their levels during the pre-pandemic
years of 2018 and 2019, when they averaged $2.81 and $2.69 per
gallon.
---------------------------------------------------------------------------
Commenters also raised concerns about the included electricity
price forecast. Auto Innovators, for example, proposed electricity rate
inputs are too low in the face of anticipated increases in renewable
electricity generation and may therefore overestimate benefits of the
regulatory action.\551\ The commenters pointed to research from the
National Renewable Energy Laboratory that suggests price increases are
possible and noted EPA's fuel price inputs increase to $0.133 per kWh
in 2040 (compared to $0.120 in the NHTSA's NPRM). Auto Innovators did
not suggest alternative price series and NHTSA is wary of varying fuel
prices without simultaneously varying assumptions about electricity
grid mix. Further, the CAFE Model is unable to simulate regional
differences in electricity generation and fuel prices and cannot
capture regional differences in electricity prices, which may arise
from heterogeneity in grid mix. The agency did include a sensitivity
case that varied projections about electricity supply and included a
case with high levels of renewable energy generation from EIA. These
results are included in FRIA Chapter 7.
---------------------------------------------------------------------------
\551\ Auto Innovators, A1, at 85.
---------------------------------------------------------------------------
Another key assumption that has important ramifications throughout
the agency's analysis is how much consumers are willing to pay for
improved fuel economy. If buyers fully value the savings in fuel costs
that result from driving (and potentially re-selling) vehicles with
higher fuel economy and manufacturers supply all improvements in fuel
economy that buyers demand, market-determined levels of fuel economy
would reflect both the cost of improving it and the private benefits
from doing so.\552\ In that case, regulations on fuel economy would
only be necessary to reflect environmental or other benefits other than
to buyers themselves. But if consumers instead undervalue future fuel
savings or are otherwise unable to purchase their optimal levels of
fuel economy due to market failures, they will underinvest in fuel
economy and manufacturers would spend too little on fuel-saving
technology (or deploy its energy-saving benefits to improve vehicles'
other
[[Page 25854]]
attributes). In that case, more stringent fuel economy standards could
lead manufacturers to adopt improvements in fuel economy that not only
reduce external costs from producing and consuming fuel to appropriate
levels but also improve consumer welfare.
---------------------------------------------------------------------------
\552\ Besides fuel savings, the private benefits from increased
fuel economy may also include increased driving range, decreased
costs per mile driven, and refueling benefits such as the experience
of not having to stop as often to refuel.
---------------------------------------------------------------------------
Increased fuel efficiency offers vehicle owners significant
potential savings; in fact, our analysis shows that the value of
prospective fuel savings exceeds manufacturers' technology costs to
comply with even the most stringent standards considered for this final
rule when both are discounted at a either a 3 percent or 7 percent
rate. It would seem reasonable to assume that well-informed vehicle
shoppers, if without time constraints or other barriers to rational
decision-making, will recognize the full value of fuel savings from
purchasing a model that offers higher fuel economy, since they would
enjoy an equivalent increase in their disposable income and the other
consumption opportunities it affords them. If consumers did value the
full amount of fuel savings, more fuel-efficient vehicles would
functionally be less costly for consumers to own when considering both
their initial purchase prices and subsequent operating costs, thus
making the models that manufacturers are likely to offer under stricter
alternatives more attractive than those available under the No-Action
Alternative.
Recent econometric research is divided between studies concluding
that consumers value most or all of the potential savings in fuel costs
from driving higher-mpg vehicles, and those concluding that consumers
significantly undervalue expected fuel savings. Based on a detailed
analysis of changes in recent sale values of cars and light trucks in
response to variation in fuel prices, Busse et al. (2013) estimated
that buyers value 54 to 117 percent of fuel savings from purchasing
higher-mpg models, with the exact value depending on the discount rate
they apply to future savings; their estimates for new car buyers ranged
from 75 to 133 percent of future fuel savings, Using similar methods
and an extremely large sample of used vehicle sales, Allcott and Wozny
(2014) estimated a corresponding range of 55 to 76 percent depending on
their assumptions about buyers' discount rates and expectations for
future fuel prices, with a figure of 93 percent for buyers of the
newest (1-3 year old) cars in their sample. Again using similar
methods, Sallee et al. (2016) estimated that car and light truck buyers
are willing to pay from 60 percent to perhaps as much as 142 percent of
the value of future fuel savings to purchase models offering higher
fuel economy. Most recently, Leard and Zhou's (2021) analysis puts the
most likely value for this figure at slightly above half (54 percent),
and Gillingham et al. (2021) find that ``consumers systematically
undervalue fuel economy in vehicle purchases to a larger degree than
reported by much of the recent literature.'' 553 554
---------------------------------------------------------------------------
\553\ Busse, M., C. Knittel, and F. Zettelmeyer. 2013. ``Are
Consumers Myopic? Evidence from New and Used Car Purchases.''
American Economic Review 103(1): 220-56; Allcott, H., and N. Wozny.
2014. ``Gasoline Prices, Fuel Economy, and the Energy Paradox.'' The
Review of Economics and Statistics 96(5): 779-95; Sallee, J., S.
West, and W. Fan. 2016. ``Do Consumers Recognize the Value of Fuel
Economy? Evidence from Used Car Prices and Gasoline Price
Fluctuations.'' Journal of Public Economics 135: 61-73; Leard, B.,
J. Linn, and Y. Zhou. 2021. ``How Much Do Consumers Value Fuel
Economy and Performance? Evidence from Technology Adoption.'' The
Review of Economics and Statistics: 1-45 (forthcoming); Gillingham,
K.T., S. Houde, and A. van Bentham, 2021. ``Consumer Myopia in
Vehicle Purchases: Evidence from a Natural Experiment.'' American
Economic Journal: Economic Policy 13(3): 207-238.
\554\ Other research asks the more fundamental questions of
whether consumers are adequately informed about and attentive to
potential fuel savings from buying higher-mpg models when they shop
for new cars, and again arrives at mixed conclusions. This includes
Allcott, H. and C. Knittel, 2019. ``Are Consumers Poorly Informed
about Fuel Economy? Evidence from Two Experiments'', AEJ: Economic
Policy, 11(1): 1-37, and D. Duncan, A. Ku, A. Julian, S. Carley, S.
Siddiki, N. Zirogiannis and J. Graham, 2019. ``Most Consumers Don't
Buy Hybrids: Is Rational Choice a Sufficient Explanation?'', J. of
Benefit-Cost Analysis, 10(1): 1-38. The former analysis concludes
that consumers appear to be relatively well-informed about the value
of higher fuel economy when they shop for new vehicles, while the
latter concludes that some buyers appear inattentive to savings
available from buying higher-MPG hybrid versions of certain vehicle
models.
---------------------------------------------------------------------------
More circumstantial evidence appears to show that consumers do not
fully value the expected lifetime fuel savings from purchasing higher-
mpg models. Although the average fuel economy of new vehicles reached
an all-time high of 25.7 MPG in MY 2020, this is still significantly
below the fuel economy of the fleet's most efficient vehicles that are
readily available for consumers to purchase.555 556
Manufacturers have repeatedly informed the agency that consumers value
only 2 to 3 years of the future fuel savings that higher-mpg cars and
light trucks offer when choosing among available models.
---------------------------------------------------------------------------
\555\ See EPA 2020 Automotive Trends Report at 6 and 9,
available at https://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=P1010U68.pdf. (Accessed: March 15, 2022)
\556\ Of course, this could simply suggest that the future
savings in fuel costs those models offer--given potential buyers'
expectations about future fuel prices--do not justify manufacturers'
costs for providing them, since those are presumably reflected in
their higher purchase prices.
---------------------------------------------------------------------------
The potential for car buyers to voluntarily forgo improvements in
fuel economy that appear to offer future savings exceeding their
initial costs is one example of what is often termed the ``energy-
efficiency gap.'' The appearance of a gap between the level of energy
efficiency that would minimize consumers' overall expenses and what
they actually purchase is typically based on engineering calculations
that compare the initial cost for providing higher energy efficiency to
the discounted present value of the resulting savings in future energy
costs. There has long been an active debate about why such a gap might
arise and whether it exists. Economic theory predicts that economically
rational individuals will purchase more energy-efficient products only
if the savings in future energy costs they offer promise to offset
their higher initial costs. On the other hand, various market failures,
including information asymmetries between consumers, dealerships, and
manufacturers; market power; first-mover disadvantages for both
consumers and manufacturers; split incentives between vehicle
purchasers and vehicle drivers; and other failures may prevent
consumers from purchasing the optimal level of fuel economy in an
unregulated market. Furthermore, behavioral economics has documented
numerous situations in which the decision-making of consumers differs
in important ways from the predictions of the model of the fully
optimizing consumer (e.g., Dellavigna, 2009).\557\
---------------------------------------------------------------------------
\557\ Dellavigna, S., 2009. ``Psychology and economics: Evidence
from the field,'' Journal of Economic Literature, 47(2), 315-372.
Available at https://pubs.aeaweb.org/doi/pdfplus/10.1257/jel.47.2.315. (Accessed: Mar. 24, 2022).
---------------------------------------------------------------------------
One explanation for such `undervaluation' of the savings from
purchasing higher-mpg models is myopia or present bias, where consumers
focus unduly on short-term costs while giving insufficient attention to
long-term benefits.\558\ This situation could arise because buyers are
unsure whether they will actually realize the fuel savings indicated by
test data posted on cars' fuel economy labels under the conditions
where they drive, what future fuel prices will be, how long they will
own a new vehicle, or whether they will drive it enough to realize the
promised savings. As a consequence, they may view choosing
[[Page 25855]]
to purchase a more fuel-efficient vehicle as a risky ``bet,'' and
experimental research has shown that when faced with a risky choice,
some consumers appear to weigh the potential loss from an adverse
outcome approximately twice as heavily as the potential gain from
``winning'' the bet, leading them to significantly undervalue that
choice relative to its probabilistic ``expected'' value (e.g., Kahneman
and Tversky, 1979; \559\ Kahneman, 2011).\560\ Viewed in the context of
a choice to pay more for a higher-mpg car, loss aversion has been shown
to have the potential to cause undervaluation of future fuel savings
like that reported by manufacturers (Greene, 2011; \561\ Greene et al.,
2013).\562\
---------------------------------------------------------------------------
\558\ Gillingham et al., 2021, which is an AEJ: Economic Policy
paper, just published on consumer myopia in vehicle purchases; a
standard reference on present bias generally is O'Donoghue, Ted, and
Matthew Rabin. 2015. ``Present Bias: Lessons Learned and To Be
Learned.'' American Economic Review: Papers & Proceedings 105(5):
273-79. Available at https://pubs.aeaweb.org/doi/pdfplus/10.1257/aer.p20151085. (Accessed: Mar. 30, 2022).
\559\ Kahneman, D. and A. Tversky, 1979. ``Prospect theory: An
analysis of decision making under risk,'' Econometrica, 47, 263-291.
\560\ Kahneman, D., 2011. Thinking Fast and Slow. Farrar, Straus
and Giroux, New York.
\561\ Greene, D.L., 2011. ``Uncertainty, Loss Aversion and
Markets for Energy Efficiency,'' Energy Economics, 33, 608-616.
\562\ Greene, D.L., D.H. Evans, and J. Hiestand, 2013. ``Survey
evidence on the willingness of U.S. consumers to pay for automotive
fuel economy,'' Energy Policy, 61, 1539-1550. Application of
investment under uncertainty will yield similar results as costs may
be more certain and up front while the fuel savings or benefits of
the investment may be perceived as more uncertain and farther into
future, thereby reducing investments in fuel saving technologies.
---------------------------------------------------------------------------
The ``behavioral'' model of consumer choice also holds that
consumers' decisions are affected by the context of choices and its
effect on how consumers ``frame'' decisions. From this perspective, it
is possible that consumers respond to changes in the fuel economy new
vehicles offer required by government regulations such as CAFE
standards differently than they respond to manufacturers voluntarily
offering buyers the option to purchase models featuring the same fuel
economy levels those regulations would require.\563\ The intuition
behind this possibility is that if a consumer is shopping for a new car
in an unregulated market and considering two models--one that offers
higher fuel economy but is more expensive and another that does not but
is cheaper--she may buy the less fuel efficient version even if
choosing the more expensive model could save money in the long run. If
instead the consumer faced the decision to buy a new car or keep an
older one, and all new car models were required to meet fuel economy
standards, she may view the decision differently and elect to purchase
a new model offering the same price and fuel economy that she
previously declined to purchase. Further, if fuel economy standards
increased gradually over a period of years, this would allow time for
consumers to consult other information sources and verify that
potential fuel savings are likely to prove real and of substantial
value.
---------------------------------------------------------------------------
\563\ See NASEM (2021), Ch. 11.3.3, We explain this potential
differential response more thoroughly in TSD Chapter 4.2.1.1.
---------------------------------------------------------------------------
Another alternative explanation for consumers' reluctance to
purchase more costly models whose lower fuel costs would ultimately
repay their higher purchase prices is that consumers view those higher
prices in the context of tradeoffs they make among their purchasing
decisions. Households must choose how to spread their limited incomes
over many competing goods and services, including deciding how much to
spend on a new vehicle, or even whether to opt for another form of
transportation instead. While a consumer may correctly recognize the
cumulative long-term value of fuel savings, they may also prefer to
spend the extra cost of buying a car that offers those savings on other
items, whether other vehicle attributes--more interior space and
comfort, for example, or a more luxurious trim package--or on other
unrelated goods and services. Some of the same technologies that
manufacturers have available to increase fuel economy can also enable
increased vehicle size, power, or weight while maintaining fuel
economy.\564\ While increased fuel efficiency will free up disposable
income throughout the lifetime of the vehicle (and may ultimately
exceed the additional upfront costs to purchase a more expensive but
more fuel-efficient vehicle), the value of owning a different good
sooner may provide consumers with even more benefit.\565\
---------------------------------------------------------------------------
\564\ Other technologies may simultaneously increase both fuel
economy and certain performance attributes.
\565\ While households have budgets, both individual vehicle
purchasers and the purchasers of large fleets of vehicles may have
access to financing for vehicle purchases. Given sufficient
financing, a rational consumer could both purchase fuel economy
improvements that will pay for themselves over time as well as other
desired goods. Failure to do so would seem to indicate either a lack
of efficient access to financing or some market failure.
---------------------------------------------------------------------------
NHTSA's NPRM included an extensive theoretical discussion of
consumer valuation of fuel economy, including a detailed theoretical
analysis of consumer choices between vehicle performance and fuel
economy when buyers are constrained by limited budgets and
manufacturers by fuel economy standards. That analysis showed that when
fuel economy standards are binding, consumers might prefer that
manufacturers employ newly available technologies that could be used to
improve performance or increase fuel economy to improve performance,
and that manufacturers would be likely to do so. NHTSA's analysis also
suggested that if fuel economy standards no longer constrained
consumers' choices, due either to shifting preferences for fuel economy
(for example, in response to changes in the price of gasoline) or to
changes in buyers' income levels, manufacturers would be likely to use
new technologies to improve both performance and fuel economy. NHTSA
then presented trends in new vehicle fuel economy and performance over
time and suggested that its theoretical analysis was consistent with
the historical record, which shows the fuel economy of the new vehicle
fleet increases when the price of gasoline increases.\566\ NHTSA
solicited comments on its theoretical analysis and the potential
implications for its FRIA, and also sought potential approaches for
valuing the tradeoff between performance and fuel economy when NHTSA's
standards constrain consumers to choose more fuel-efficient options.
---------------------------------------------------------------------------
\566\ For additional details, see 86 FR 49723-31 (Sept. 3,
2021).
---------------------------------------------------------------------------
NHTSA noted in the NPRM that the substantial literature on the
topic of consumer valuation of fuel economy is approximately evenly
divided between studies that suggest consumer undervalue fuel economy
and studies that support valuation at the full discounted present value
(no undervaluation). This potential undervaluation, frequently referred
to as the ``energy paradox'' or ``fuel efficiency gap,'' has prompted
an extensive exploration of potential behavioral explanations why
consumers might undervalue fuel economy. NHTSA explored the possibility
that the context and framing around consumer decisions may influence
consumer choices--and that consumers may value fuel-saving technology
differently when their choices are constrained to more fuel-efficient
options. NHTSA also discussed how the value consumers place on fuel
economy may change over time, and that they may come to value the
future stream of fuel savings more once they begin to experience those
savings when the rule is in place. NHTSA noted that if fuel economy
standards lead consumers to value fuel economy more once they
experience a savings, the new higher valuation of fuel economy may
offset some or all of the negative impact
[[Page 25856]]
on sales due to the higher prices of fuel-efficient vehicles.
As explained in more detail in TSD Chapter 4.2.1.1, the agency's
analyses of the extent to which manufacturers will voluntarily improve
fuel economy and of the response of new car and light truck sales to
higher sales prices assume that potential buyers of new cars and light
trucks value only the undiscounted savings in fuel costs they would
expect to realize over the first 30 months they own a newly purchased
vehicle. Depending on the discount rate buyers are assumed to apply,
this amounts to 25-30 percent of the expected savings in fuel costs
they (and any subsequent owners) would ultimately realize over the
vehicle's entire expected lifetime. However, NHTSA establishes CAFE
standards by comparing vehicles' lifetime savings in fuel costs and
other economic benefits from reducing fuel consumption to
manufacturers' costs to improve fuel economy, which leads the agency to
set standards that require much higher levels of fuel economy than it
assumes buyers are willing to pay for. Thus, the agency's analysis does
assume that new car shoppers are somewhat myopic--and that an ``energy
paradox'' exists in the case of fuel economy--but only at the time they
are consider purchasing a new car or light truck, and that they
ultimately value the lifetime fuel savings that purchasing a higher-mpg
model provides.\567\ The agency also assumes that manufacturers'
compliance costs will ultimately be borne by vehicle buyers in the form
of higher purchase prices for new cars and light trucks. This means
that the fraction of savings in future fuel costs buyers are assumed to
take into account at the time of purchase (again, 25-30 percent) when
choosing among models would offset only that same fraction of the
expected increase in new car and light truck prices.
---------------------------------------------------------------------------
\567\ In addition to myopia, other market failures may also
cause consumers to undervalue fuel savings at the time of purchase
but still fully value the lifetime fuel savings they actually
experience, including information asymmetries, split incentives,
first-mover effects, and others. Moreover, it is appropriate in a
social cost-benefit analysis to fully value the resource savings
that will result from the purchase of vehicles with greater fuel
economy.
---------------------------------------------------------------------------
NHTSA sought comment on the length of time that should be used for
this ``payback period'' assumption, and asked commenters to specify the
length of time they believed it should span, provide an explanation of
why that period is preferable to the agency's assumption, include
reference to any data or information on which an alternative payback
period is based, and discuss how changing this assumption would
interact with other elements in the analysis. In response, NHTSA
received a handful of comments on this apparent ``energy efficiency
gap'' and the agency's assumption about consumers' willingness to pay.
NADA and Auto Innovators agreed with the agency's assumption of a 30-
month payback period, while stressing the need to account for the
utility of other vehicle attributes that might be improved in the
absence of mandates to provide higher fuel economy.\568\ NADA commented
that consumers are not myopic, and any appearance that they are
actually reflects their wide range of preferences for other vehicle
attributes, which also explains their willingness to forgo some fuel
savings in favor of improvements to vehicles' other features. NADA
asserted that potential buyers of new cars and light trucks focus on
the total lifetime cost of vehicle ownership, and by doing so consider
the cost and value of purchasing models that offer higher levels of not
just fuel economy, but other desirable features as well. To support its
claim, NADA cited to data from the 2021 Strategic Vision New Vehicle
Efficiency Survey that found fuel economy ranked as the 12th most
important attribute to consumers. NADA argued that NHTSA needed to
examine ``actual sales and lease data or studies assessing how new
light-duty vehicle consumers value fuel economy technology when making
purchasing decisions,'' and implored the agency to account for the
``temporal shifting of consumer preferences.'' Auto Innovators
supported analyzing sensitivity cases with payback periods ranging from
1 to 4 years.\569\
---------------------------------------------------------------------------
\568\ NADA, Docket No. NHTSA-2021-0053-1471, at 8-9.
\569\ Auto Innovators, at 83-84.
---------------------------------------------------------------------------
EDF commented that the agency should assume a longer repayment
period and cited as support a Consumer Reports study showing that 64
percent of consumers rank fuel economy as extremely or very important,
and view fuel economy as ``the number one attribute that has room for
improvement.'' \570\ NHTSA notes that the same Consumer Report study
also polled consumers about how quickly fuel savings would have to
offset higher vehicle purchase prices for them to be willing to pay for
increased fuel efficiency. Responses to this question showed that the
average consumer is willing to pay for only 2-3 years of fuel savings,
which aligns well with the agency's estimate of 30 months, and that
only 39 percent of consumers are willing to pay for fuel economy
improvements with a payback period longer than 3 years.\571\
---------------------------------------------------------------------------
\570\ Environmental Defense Fund, Docket No. NHTSA-2021-0053-
1617, at 5.
\571\ Consumer Reports, Consumer Attitudes Towards Fuel
Economy'' 2020 Survey Results (Feb. 2021), page 5, https://advocacy.consumerreports.org/wp-content/uploads/2021/02/National-Fuel-Economy-Survey-Report-Feb-2021-FINAL.pdf. (Accessed: March 15,
2022).
---------------------------------------------------------------------------
CBD et al. commented that the agency is underestimating consumers'
willingness to pay by assuming that they require a 30-month payback
period, but did not explain why it believes this is the case or suggest
an alternative estimate.\572\
---------------------------------------------------------------------------
\572\ Center for Biological Diversity, Chesapeake Bay
Foundation, Conservation Law Foundation, Earthjustice, Environmental
Law & Policy Center, Natural Resources Defense Council, Public
Citizen, Inc., Sierra Club, and Union of Concerned Scientists
(NHTSA-2021-0053-1572) (CBD et al.), Joint Summary Comments, Docket
No. NHTSA-2021-0053-1572, at 6.
---------------------------------------------------------------------------
Institute for Policy Integrity at New York School of Law (IPI)
urged the agency consider using different payback assumptions at
different points throughout its analysis. Specifically, IPI commented
that NHTSA should use a lower willingness to pay under the baseline
scenario to determine how much manufacturers would voluntarily improve
fuel economy in the absence of stricter standards, but should assume a
higher willingness to pay when analyzing how the standards will affect
sales of new vehicles and the turnover of the used vehicle fleet.\573\
IPI endorsed the possibility the agency raised in its proposal that
CAFE regulations can ameliorate myopia among potential buyers or
information asymmetries between vehicle manufacturers and buyers, and
by doing so lead potential buyers to value a larger fraction of future
fuel savings from choosing a higher-mpg model. IPI also listed other
potential market failures that CAFE regulations could potentially
mitigate.\574\
---------------------------------------------------------------------------
\573\ IPI, Docket No. NHTSA-2021-0053-1579-A1, at 16-17.
\574\ See generally, id., at 9-14.
---------------------------------------------------------------------------
Specifically, IPI suggested that the agency use a 1.7-year payback
period to identify the technologies manufacturers would adopt and to
estimate the resulting increase in fuel economy under the baseline, but
assume that actual buyers of new cars and light trucks would value fuel
savings over the first 7 years of their lifetimes when evaluating
whether to scrap a vehicle. scrappage rates. However, IPI did not offer
NHTSA a framework for implementing differing payback periods, or
explain whether the difference in payback periods was intended to
reflect manufacturers
[[Page 25857]]
underestimation of buyers' valuation of fuel economy and if so, why
manufacturers would do so only under the No-Action Alternative. Nor did
IPI specify how long after new standards were adopted would be required
for consumers to begin to value additional fuel economy, or why they
would revert to their original lower valuation once new standards took
effect and became the baseline for evaluating further increases. IPI
also commented that if the agency opted not to use differing payback
assumptions, then the agency should use a shorter payback period (1.7
years) throughout the analysis to avoid overestimating overcompliance
in the baseline,\575\ and suggested that the agency conduct expert
elicitation to derive a better estimate.\576\
---------------------------------------------------------------------------
\575\ Id.
\576\ Id. at 15.
---------------------------------------------------------------------------
IPI also commented that NHTSA's theoretical analysis of constrained
consumer choice lacked an empirical test of its validity and that other
explanations for the historical pattern of increases in fuel economy
and changes to vehicles' other attributes may be more plausible than
that offered by the agency. IPI also argued that consumers' choices
involving higher-mpg models cannot be constrained by their budgets
because fuel savings compensate consumers for paying the higher upfront
costs (thus enabling buyers to finance those additional costs). IPI
argued further that failures in the market for auto financing that make
consumers unable to obtain favorable financing to purchase more fuel-
efficient vehicles may constrain consumers' choices more than any
budgetary limits. IPI continued that NHTSA's prior estimates of the
opportunity cost of other vehicle attributes lacked an empirical basis
and ignored potential countervailing effects such as reduced compliance
costs.
In contrast, NADA commented that a consumer's willingness to
purchase fuel-economy technology must be viewed in the context of
losses in other vehicle attributes like power or safety, and argued
that consumers are not myopic. In support of its position, NADA cited
Leard et al.'s (2021) finding that consumers undervalue fuel economy
but place high values on performance and other attributes,\577\ as well
as Klier and Linn's (2016) finding that tighter vehicle standards
reduce horsepower and torque relative to their levels where standards
remain unchanged.\578\ Finally, IPI cited the conclusion of EPA's
Scientific Advisory Board that it found little ``useful consensus'' on
the subject of the opportunity cost of other vehicle attributes \579\
and Greene (2018), who found extensive variation in willingness-to-pay
estimates across the literature.
---------------------------------------------------------------------------
\577\ Leard, B., J. Linn, and Y. Zhou. 2021. ``How Much Do
Consumers Value Fuel Economy and Performance? Evidence from
Technology Adoption.'' The Review of Economics and Statistics: 1-45
(forthcoming). Adoption, The Review of Economics and Statistics 2021
(Leard, et al.).
\578\ Klier, Thomas, and Joshua Linn. 2016. ``The Effect of
Vehicle Fuel Economy Standards on Technology Adoption.'' Journal of
Public Economics 133, pp. 41-63).
\579\ EPA Sci. Advisory Bd., Consideration of the Scientific and
Technical Basis for the EPA's Proposed Rule Titled The Safer
Affordable Fuel-Efficient (SAFE) Vehicles Rule for Model Years 2021-
2026 Passenger Cars and Light Trucks, at 2 (Feb. 27, 2020),
available at https://sab.epa.gov/ords/sab/f?p=100:18:6529621058907:::RP,18:P18_ID:2550 (``We concur with the
agencies that it is not yet feasible to quantify the impact on new
vehicle sales of additional vehicle characteristics (beyond fuel
economy) that are desired by consumers but restrained by federal
standards.''). David Greene et al., Consumer Willingness to Pay for
Vehicle Attributes: What Do We Know?, 118 TRANSP. RES. PART A: POL'Y
& PRAC. 258, 264, 273 (2018); see also id. at 274 (finding that,
even after trimming outliers, ``one standard deviation exceeds the
mean of the [willingness to pay] estimates for most of the
attributes'' and that ``the interquartile range also exceeds the
median'').
---------------------------------------------------------------------------
NHTSA agrees with IPI that the theoretical discussion of
constrained consumer choice under binding fuel economy standards has
not been tested empirically, and for this reason has not incorporated
an estimate of the opportunity cost of sacrifices in other vehicle
attributes in its FRIA. NHTSA notes that the alternative explanations
posited by IPI to explain the fuel efficiency gap also lack an
empirical basis--instead, both the agency's and IPI's explanations are
consistent with consumers' apparent willingness to forgo some fuel
savings in favor of improvements to vehicles' other features. However,
NHTSA notes that, because--as acknowledged later in its comment--IPI's
comment overlooks the theoretical possibility that automakers could at
some point run out of technologies that could improve performance such
that the use of a technology to improve fuel economy rather than
performance would necessarily mean a lack of availability of
performance enhancements. Even if all available technologies were
deployed to improve fuel economy rather than performance, and those
technologies fully paid for themselves with discounted future fuel
savings, then manufacturers would have no remaining technologies
available to meet buyers' demands for improved performance. However, no
such absolute technological constraint has been observed. Furthermore,
the agency notes that IPI's comment lacks any consideration of how much
households can afford to spend on vehicle loan payments, instead
assuming that households will assume as much debt as necessary to
purchase a vehicle with their preferred bundle of attributes. NADA
commented that most households already cannot afford to purchase new
vehicles, and noted that financing does not take into consideration
potential fuel savings but instead relies on a borrower's income,
finance amount, and credit worthiness.\580\
---------------------------------------------------------------------------
\580\ NADA, at 6-7. We note that EPA disagrees and has found
that some lenders give discounts for loans to purchase more fuel-
efficient vehicles. See EPA, Revised 2023 and Later Model Year
Light-Duty Vehicle GHG Emissions Standards: Regulatory Impact
Analysis at 8-27 and n.87 (2021).
---------------------------------------------------------------------------
NHTSA acknowledges that the opportunity cost of regulations on
other vehicle attributes is still an under-researched topic and relies
heavily on economic theory, and for this reason, we are excluding
estimates of this particular theoretical opportunity cost in its
primary analysis. NADA provided some literature that it believes may
assist the agency in developing an estimate of the opportunity cost of
other vehicle attributes in the future, but NHTSA agrees with the EPA's
Scientific Advisory board that there is little consensus on this issue.
For illustrative purposes, NHTSA has included a sensitivity analysis
estimating the theoretical opportunity cost of other vehicle attributes
in the FRIA, although as discussed elsewhere, NHTSA is not confident
that the assumptions used to generate this estimate are sound. NHTSA
notes that the sensitivity analysis of opportunity costs is a rough,
speculative proxy with multiple limitations that does not reflect many
other effects that may largely offset such opportunity costs. The
sensitivity estimate should be considered as an overestimate of the
potential effects, and is not sufficiently robust to include in the
main analysis. Opportunity cost from other vehicle attributes, to the
extent it exists, may be small. NHTSA notes that consideration of such
sensitivity analysis does not change NHTSA's conclusion that
Alternative 2.5 is the maximum feasible and most appropriate standard
under its statutory factors.
NADA also comments that the agency's assumption that potential
buyers consider their expected future fuel savings over some assumed
``payback period'' when deciding whether to purchase models offering
higher fuel economy oversimplified buyer's choices, even if other
attributes
[[Page 25858]]
of models they are comparing are closely comparable.\581\ Specifically,
NADA argues that both the importance vehicle shoppers attach to higher
fuel economy and the time horizon over which they evaluate savings in
fuel costs from buying higher-MPG models vary in response to the
direction and speed of recent movements in fuel prices, and that
potential buyers appear to make the calculations the agency assumes
only when fuel prices are increasing rapidly. When fuel prices are more
stable, NADA argues that consumers appear to focus on vehicles' other
attributes, and at current fuel prices NADA asserts that buyers are
unlikely to demand more fuel-efficient cars and light trucks,
particularly as their preferences continue to evolve toward SUV and CUV
models.
---------------------------------------------------------------------------
\581\ NADA, at 9.
---------------------------------------------------------------------------
On these points, NADA does not offer specific recommendations about
how the agency could represent its interpretation of buyers' choice
process, and the agency's interpretation is that doing so would require
it to vary the assumed duration of buyers' payback period in response
to both the direction and pace of recent changes in fuel prices,
lengthening it when fuel prices are rising rapidly and shortening it
when prices are stable or declining. While the agency does not believe
that this approach is reasonable or practical, it has included
sensitivity cases in the accompanying FRIA that consider both shorter
and longer payback periods than the 2.5 years assumed in the central
analysis, and believes their results should shed useful light on the
potential effects of NADA's recommended approach.
For several reasons, we decided to retain our 30-month payback
assumption for evaluating the alternatives we considered for the final
rule. First, there was no consensus among commenters about a more
appropriate payback period; approximately equal numbers of commenters
urged the agency to lengthen, maintain, and shorten the duration of its
assumed payback period. Second, none of the commenters who urged the
agency to change the duration of its assumed payback period provided
any additional evidence to support doing so, and thus NHTSA continues
to believe that the information on which the payback decision is based
is reasonable and appropriate. Finally, none provided plausible
explanations for why adopting fuel economy standards should change
vehicle buyers' time perspectives on future fuel savings, why their
longer-term perspectives would revert to their original shorter terms
once those standards took effect, or why repeat buyers' values would
once again adopt a longer-term perspective when valuing future fuel
savings when standards were once again raised.
While we will continue to explore whether payback periods should
differ between the baseline and regulatory alternatives that would
establish higher standards, the agency still lacks a clear basis for
identifying whether, how much, or how quickly future changes in CAFE
standards could alter consumer perceptions of fuel economy and its
value. In addition, neither the agency nor commenters has identified a
satisfactory explanation for why once having adapted to the presence of
higher fuel economy standards by lengthening the time horizon over
which they value fuel savings, consumers would revert to their former
lower values once those new standards became the reference point for
evaluating further increases in required fuel economy. The agency will
also re-examine whether a 30-month payback period is appropriate to use
in analyzing future increases in standards, and will consider whether
an expert elicitation is appropriate.
2. Fleet Composition
The composition of the on-road fleet--and how it changes in
response to CAFE standards--determines many of the costs and benefits
of the final standards. For example, how much fuel the light-duty fleet
consumes is dependent on the number of new vehicles sold, how many
older (and less efficient) vehicles are retired, and how much vehicles
are driven.
Until recently, all previous CAFE rulemaking analyses used static
fleet forecasts that were based on a combination of manufacturer
compliance data, public data sources, and proprietary forecasts (or
product plans submitted by manufacturers). When simulating compliance
with regulatory alternatives, those analyses projected identical sales
and retirements across the alternatives, for each manufacturer down to
the make/model level--where the exact same number of each model variant
was assumed to be sold in a given model year under both the least
stringent alternative (typically the baseline) and the most stringent
alternative considered (intended to represent ``maximum technology''
scenarios in some cases). To the extent that an alternative matched the
assumptions made in the production of the proprietary forecast, using a
static fleet based upon those assumptions may have been warranted.
However, a fleet forecast is unlikely to be representative of a
broad set of regulatory alternatives with significant variation in the
cost of new vehicles. Several commenters on previous regulatory actions
and peer reviewers of the CAFE Model encouraged consideration of the
potential impact of fuel efficiency standards on new vehicle prices and
sales, the changes to compliance strategies that those shifts could
necessitate, and the downstream impact on vehicle retirement rates. In
particular, the continued growth of the utility vehicle segment causes
changes within some manufacturers' fleets as sales volumes shift from
one region of the footprint curve to another, or as mass is added to
increase the ride height of a vehicle on a sedan platform to create a
crossover utility vehicle, which exists on the same place of the
footprint curve as the sedan upon which it might be based.
The analysis now dynamically simulates changes in the vehicle
fleet's size, composition, and usage as manufacturers and consumers
respond to regulatory alternatives, fuel prices, and macroeconomic
conditions. The analysis of fleet composition comprises two forces, how
new vehicle sales--the flow of new vehicles into the registered
population--change in response to regulatory alternatives, and the
influence of economic and regulatory factors on vehicle retirement
(otherwise known as scrappage).
While commenters raised specific objections to several of the
assumptions within the sales and scrappage modules--which are described
below--commenters generally were supportive of the agency's approach to
modeling fleet turnover. We did receive one comment from IPI suggesting
that we should consider returning to a static fleet model if we were
unable to correct what they perceived as modeling flaws. We disagree
with IPI's assessment, because it is widely acknowledged that CAFE
standards and other regulations on new vehicles can influence
consumers' decisions about both purchasing new vehicles and retiring
used ones, so to assume that the composition of the vehicle fleet is
unaffected by regulations would ignore these well documented impacts.
The agency feels that it is important to provide policymakers with the
most comprehensive and complete analysis of the regulations, which
includes understanding how CAFE standards will affect fleet turnover.
Below are brief descriptions that of how the agency models sales
and scrappage. For a full explanation, refer to TSD Chapter 4.2.
[[Page 25859]]
(a) Sales
For the purposes of regulatory evaluation, the relevant sales
metric is the difference in sales between alternatives rather than the
absolute number of sales in any of the alternatives. As such, the sales
response model currently contains three parts: A nominal forecast that
provides the level of sales in the baseline (based upon macroeconomic
inputs, exclusively), a price elasticity that creates sales differences
relative to that baseline in each year, and a fleet share model that
produces differences in the passenger car and light truck market share
in each alternative. The nominal forecast does not include price and is
merely a (continuous) function of several macroeconomic variables that
are provided to the model as inputs. The price elasticity is also
specified as an input. In the proposal, the agency assumed a price
elasticity of sales of -1.0 and sought comment on this assumption.
Many commenters argued that NHTSA's unit elastic response
assumption of -1.0 is inaccurate. The California Attorney General et
al., IPI, ICCT, UCS, CBD et al., CARB and Dr. Kenneth Gillingham, all
commented that -1.0 is too large and unsupported by the evidence.\582\
CBD et al. and the California Attorney General noted that recent
literature suggests a much lower figure, with California's Attorney
General suggesting using the estimate from Leard (2021) of -0.34 and
the CBD et al. suggesting between -0.2 or -0.4 (or lower). IPI
suggested reducing the figure to at least -0.4, the figure used by EPA
in a recent sensitivity analysis. ICCT suggested that NHTSA use -0.5,
and further recommended that NHTSA consider using different elasticity
estimates for different vehicle classes.
---------------------------------------------------------------------------
\582\ California Attorney General et al., Docket No. NHTSA-2021-
0053-1499, Appendix A, at 32; IPI, A1, at 26-28; ICCT, Docket No.
NHTSA-2021-0053-1581, at 3, 14, 19; UCS Docket No. NHTSA-2021-0053-
1567, at 29; CBD et al., Joint Summary Comments, at 3-4, 6; CARB,
Docket No. NHTSA-2021-0053-1542, Attachment 2, at 3.
---------------------------------------------------------------------------
IPI and CBD et al. supported their suggested estimates by arguing
that NHTSA should utilize a long-run elasticity estimate, not a short-
run elasticity estimate.\583\ IPI explained that long-run price
elasticity of demand for vehicles tends to be much lower than short run
elasticity, because, due to the limited substitution options for
personal vehicles, consumers will delay purchases when prices increase
but are likely to still purchase a vehicle down the road. CBD et al.
noted that that a long-run estimate is more appropriate because
consumers replace vehicles in the long run as they age and because it
more closely matches the timeline of this agency action in which fuel
economy standards apply years into the future. They also argued that a
``more reasonable'' price elasticity estimate would likely lead to
greater projected increases in employment than already estimated in the
proposed rule.
---------------------------------------------------------------------------
\583\ IPI, at 26; CBD et al., Joint Summary Comments, at 6.
---------------------------------------------------------------------------
Dr. Mark Jacobsen commented that the demand elasticity that the
agency used in the proposal is the improper measurement. Dr. Jacobsen
argued that NHTSA should instead employ a ``policy elasticity'' since
CAFE regulations will influence not only new vehicles prices but also
used vehicle prices, since the two are substitutes.\584\ Because used
vehicle prices are anticipated to increase, the change in sales in
response to increasing CAFE standards will be less than what would be
anticipated if only new vehicle prices were affected. Dr. Jacobsen
suggested the policy elasticity ranges from -0.5 in the short-run to -
0.28 in the long-run.
---------------------------------------------------------------------------
\584\ Dr. Mark Jacobsen, Docket No. NHTSA-2021-0053-1586, at 2.
---------------------------------------------------------------------------
In contrast, NADA expressed support for a sales elasticity of -
1.0.\585\
---------------------------------------------------------------------------
\585\ NADA, at 11.
---------------------------------------------------------------------------
While evaluating the concerns raised by commenters, NHTSA
identified an error in the CARs report that the agency relied upon as a
key source for selecting -1.0. The CARs report erroneously reported the
own-price elasticity of cars (-0.79) and trucks (-0.85) instead of the
long-run elasticity of all light-duty vehicles (-0.39) for Fischer
(2007). When considering the actual long-run elasticity in Fischer
(2007), the totality of the evidence presented in the CARs report no
longer supports an elasticity of -1.0. In addition, after the
publication of NHTSA's proposed rule, EPA issued a new report exploring
the effects of changes in vehicle prices that arise from due to fuel
efficiency regulations on vehicle sales. Since that report was authored
by Dr. Jacobsen, it unsurprisingly echoed his comments summarized
above, and recommended that the agency reduce the magnitude of the
sales price elasticity it uses in its analysis to the range suggested
above.\586\
---------------------------------------------------------------------------
\586\ Chapter 4.3.2 of the FRIA accompanying this final rule
includes a detailed discussion of the interactions between new and
used vehicle markets identified in Dr. Jacobsen's report to EPA and
their implications for the sensitivity of new vehicle sales and
retirement of used vehicles to higher sales prices.
---------------------------------------------------------------------------
For these reasons, NHTSA has elected to use a price elasticity of
sales equal to -0.4--meaning that a ten percent increase in the average
price of a new vehicle produces a four percent decrease in total
sales--for the final rule. The price change to which this elasticity is
applied is calculated as the per-vehicle average of manufacturers'
estimated costs to meet higher CAFE standards, net of the fraction of
vehicles expected lifetime fuel savings that new vehicle buyers are
assumed to value (2.5 years or 25-30 percent of lifetime savings, as
discussed in Section III.E.1. above). NADA commented that it believed
the agency's sales model was not appropriately applying the sales
elasticity to the assumed price increase and thus underestimated the
likely decline in sales.\587\ However, the agency notes that NADA's
rough sales estimates excluded any value of future fuel savings, and
that this omission was likely to have caused the divergence between
NADA's and NHTSA's estimates of changes in sales.
---------------------------------------------------------------------------
\587\ NADA, at 12.
---------------------------------------------------------------------------
The current baseline sales module reflects the idea that total new
vehicle sales are primarily driven by conditions in the economy that
are exogenous to the automobile industry. Over time, new vehicle sales
have followed macroeconomic cycles closely, rising when prevailing
economic conditions are positive (periods of growth) and falling during
periods of economic contraction. While the kinds of changes to vehicle
offerings that occur because of manufacturers' compliance actions exert
some influence on the total volume of new vehicle sales, their effects
on new vehicle sales are secondary to those of overall economic
conditions. Instead, they drive the kinds of marginal differences
between regulatory alternatives that the current sales module is
designed to simulate--making vehicles more expensive generally reduces
total sales, although only modestly.
The first component of the sales response model is a nominal
forecast, which is a statistical model (using a small set of inputs)
that projects the size of the new vehicle market in each calendar year
in the analysis period under the baseline (No-Action Alternative). Past
reviewers expressed concerns about the possibility of econometrically
estimating an industry average price elasticity in a way that isolates
the causal effect of new vehicle prices on new vehicle sales (and
properly addresses the issue of endogeneity between sales and price).
However, the agency's current nominal forecast model does not include
prices and is not intended for statistical inference around the
question of price response in the new vehicle market;
[[Page 25860]]
instead, it is intended to simulate the general trajectory of the
market for light duty vehicles. As discussed in more detail in Section
III below, the current economic climate and the economy's performance
during the continuing pandemic has created unusually extreme
uncertainty about this year-to-year forecast. Particularly in the near-
term, there is significant uncertainty about the pace at which the
market for automobiles will recover--and the scale and timing of the
recovery's peak--before the market returns to its long-term trend.
The second component of the sales response model captures how price
changes affect the number of vehicles sold, by applying an assumed
price elasticity to the percentage change in average price (in each
future year) to determine the percent change in sales from its
projected baseline value. This price change does not represent an
increase/decrease over the last observed year, but rather the
percentage difference under each regulatory alternative relative to the
estimated baseline price during that year. In the baseline, the average
price is defined as the observed new vehicle price in 2019 (the last
historical year before the simulation begins) plus the average
regulatory cost associated with the No-Action Alternative.\588\ The
central analysis in this final rule simulates multiple programs
simultaneously (CAFE final standards, EPA final greenhouse gas
standards, ZEV, and the California Framework Agreements), and the
regulatory cost includes both technology costs and civil penalties paid
for non-compliance (with CAFE standards) in a model year. Because the
elasticity assumes no perceived change in the quality of the product,
and the vehicles produced under different regulatory scenarios have
inherently different operating costs, the price metric must account for
this difference. The price to which the elasticity is applied in this
analysis represents the residual price change between scenarios after
accounting for 2.5 years' worth of fuel savings to the new vehicle
buyer.
---------------------------------------------------------------------------
\588\ The CAFE Model currently operates as if all costs incurred
by the manufacturer as a consequence of meeting regulatory
requirements, whether those are the cost of additional technology
applied to vehicles in order to improve fleetwide fuel economy or
civil penalties paid when fleets fail to achieve their standard, are
``passed through'' to buyers of new vehicles in the form of price
increases.
---------------------------------------------------------------------------
The third and final component of the sales model is the dynamic
fleet share module (DFS). Some commenters to previous rules noted that
the market share of SUVs continues to grow, while conventional
passenger car body-styles continue to lose market share. For instance,
in the 2012 final rule, the agencies projected fleet shares based on
the continuation of the baseline standards (MYs 2012-2016) and a fuel
price forecast that was much higher than the realized prices since that
time. As a result, that analysis assumed passenger car body-styles
would comprise about 70 percent of the new vehicle market by 2025,
which was internally consistent. The reality, however, has been quite
different: In MY 2020, light truck models accounted for 57 percent of
new light-duty vehicle sales.\589\ The CAFE Model includes the DFS
model in an attempt to address these market realities. The DFS
distributes the total industry sales across two different body-types:
``cars'' and ``light trucks.'' While there are specific definitions of
``passenger cars'' and ``light trucks'' that determine a vehicle's
regulatory class, the distinction used in this phase of the analysis is
more simplistic. All body-styles that are obviously cars--sedans,
coupes, convertibles, hatchbacks, and station wagons--are defined as
``cars'' for the purpose of determining fleet share. Everything else--
SUVs, smaller SUVs (crossovers), vans, and pickup trucks--are defined
as ``light trucks''--even though they may not be treated as such for
compliance purposes. The DFS uses two functions from the National
Energy Modeling System (NEMS) used in the 2017 AEO to independently
estimate the share of passenger cars and light trucks, respectively,
given average new market attributes (fuel economy, horsepower, and curb
weight) for each group and current fuel prices, as well as the prior
year's market share and prior year's attributes. The two independently
estimated shares are then normalized to ensure that they sum to one.
These shares are applied to the total industry sales derived in the
first stage of the sales response. This produces total industry volumes
of car and light truck body styles. Individual model sales are then
determined from there based on the following sequence: (1) Individual
manufacturer shares of each body style (either car or light truck)
times the total industry sales of that body style, then (2) each
vehicle within a manufacturer's volume of that body-style is given the
same percentage of sales as appear in the 2020 fleet. This implicitly
assumes that consumer preferences for particular styles of vehicles are
determined in the aggregate (at the industry level), but that
manufacturers' sales shares of those body styles are consistent with MY
2020 sales. Within a given body style, a manufacturer's sales shares of
individual models are also assumed to be constant over time. This
approach implicitly assumes that manufacturers are currently pricing
individual vehicle models within market segments in a way that
maximizes their profit. Without more information about each OEM's true
cost of production and operation, fixed and variables costs, and both
desired and achievable profit margins on individual vehicle models,
there is no basis to assume that strategic shifts within a
manufacturer's portfolio will occur in response to standards.
---------------------------------------------------------------------------
\589\ Calculated from summary data tables accompanying EPA
Automotive Trends Report, 2021 edition, https://www.epa.gov/automotive-trends/explore-automotive-trends-data#SummaryData.
(Accessed: March 15, 2022).
---------------------------------------------------------------------------
The DFS model shows passenger car styles gaining share with higher
fuel prices and losing them when prices are decline. Similarly, as fuel
economy increases in light truck models, which offer consumers other
desirable attributes beyond fuel economy (ride height or interior
volume, for example) their relative share increases. However, this
approach does not suggest that consumers dislike fuel economy in
passenger cars, but merely recognizes the fact that fuel economy has
diminishing returns in terms of fuel savings. As the fuel economy of
light trucks increases, the tradeoff between passenger car and light
truck purchases increasingly involves a consideration of other
attributes. The coefficients also show a relatively stronger preference
for power improvements in cars than light trucks because that is an
attribute where trucks have typically outperformed cars, just as cars
have outperformed trucks for fuel economy.
NHTSA received a several comments about the dynamic fleet share
model. ICCT commented that the coefficient for horsepower for passenger
cars was negative, implying that passenger cars with lower fuel economy
and less power are more attractive to consumers.\590\ Both ICCT and IPI
also noted the counterintuitive sign for fuel economy, and suggested
that the model was inadequate because it estimates the share of cars
and trucks independently and fails to consider other vehicle attributes
such as sales prices.\591\ Neither IPI nor ICCT suggested revisions to
the current DFS model structure that would address these concerns.
Alternative approaches such as the simplified discrete choice model of
market share suggested by ICCT or
[[Page 25861]]
assuming that fleet shares remain constant could be readily
implemented, although both have potentially important drawbacks.
---------------------------------------------------------------------------
\590\ ICCT, Appendix: Additional Comments, at 14.
\591\ ICCT, Appendix: Additional Comments, at 14, 20; IPI, at
29.
---------------------------------------------------------------------------
The agency agrees with ICCT that a discrete choice model calibrated
to aggregate market share data may avoid some of the challenges of
discrete choice modeling using data on individual buyers' choices but
notes that other impediments to using it would undoubtedly still
arise--for example, accounting for future changes in the classification
of some individual vehicle models, or for shifts in buyers' preferences
toward car or truck-based designs. The agency also believes that
assuming fixed fleet shares is clearly an unsatisfactory approach in
light of both gradual longer-term changes in buyers' apparent
preferences and the very rapid recent shifts in market shares for cars
and light trucks.
NHTSA agrees that a dynamic fleet share model that includes the
attributes identified by commenters, such as IPI, would be preferable.
In fact, the agency developed a number of simplified market share
models for potential use in this analysis, each of which estimated the
shares of cars and light trucks jointly using different combinations of
attributes buyers are likely to consider when choosing among competing
models. We also attempted to incorporate vehicle prices and develop
specifications that would produce logically consistent coefficients for
each variable they included. The agency was unable to produce a model
that met all three criteria--including vehicle prices proved
particularly troublesome--and these alternative models each suffered
from their own limitations.\592\ For two main reasons, the agency
ultimately decided to retain the DFS used in the proposal instead of
employing one of the newer models it developed: First, the alternative
models did not clearly meet the criteria we established to be
considered a better model. Second, the agency feels that the DFS used
in the proposal produced logically consistent results among the
alternatives it considered in this analysis. As noted elsewhere in this
rule, isolating the impact of alternatives is more an art of internal
precision within the model than an exercise in ``external validity'' or
accuracy. The agency will continue to explore alternative DFS models
for future rulemakings.\593\
---------------------------------------------------------------------------
\592\ See ``Exploration of alternate fleet share module'' in
Docket No. NHTSA-2021-0053.
\593\ As with all aspects of this analysis, uncertainty abounds.
If NHTSA's current approach to modeling fleet share inaccurately
overestimates the future fleet's proportion of light trucks, then
NHTSA may have underestimated fuel savings and overestimated
emissions of the regulatory alternatives included in this analysis.
---------------------------------------------------------------------------
Over the course of past rulemakings, many commenters have
encouraged the agency to consider vehicle attributes beyond price and
fuel economy when estimating a sales response to fuel economy
standards. Some have suggested that a more detailed representation of
the new vehicle market would enable the agency to incorporate the
effect of additional vehicle attributes on buyers' choices among
competing models, reflect consumers' differing preferences for specific
vehicle attributes, and provide the capability to simulate responses
such as strategic pricing strategies by manufacturers intended to alter
the mix of models they sell and enable them to comply with new CAFE
standards. For these purposes, nearly all of those commenters have
suggested that the agency develop a disaggregate model of buyers'
vehicle choices.\594\
---------------------------------------------------------------------------
\594\ Comments to this effect on the proposed rule were
infrequent, and the only example generally cited much more detailed
applications or advantages of discrete choice models; see Auto
Innovators, Docket No. NHTSA-2021-0053-1492, at 56.
---------------------------------------------------------------------------
A correctly specified choice model with parameters estimated from
characteristics of individual shoppers (or households) and their
choices among vehicle models--including decisions by some not to
purchase new vehicles--offers the potential to produce consistent
forecasts of total sales of new vehicles and the shares represented by
cars and light trucks (as well as specific body styles and potentially
even individual models). Developing such a model would also provide
estimates of the value buyers attach to improved fuel economy and other
vehicle attributes that were consistent with and reflected in its
forecasts of total sales and market shares for individual vehicle
types. For these reasons, the agency has invested considerable
resources in developing such a discrete choice model of the new
automobile market, although those investments have not yet produced a
satisfactory and operational model.
The agency's experience partly reflects the fact that discrete
choice models are highly sensitive to their data inputs and estimation
procedures, and even versions that fit well when calibrated to data
from a single period--usually a cross-section of vehicles and shoppers
or actual buyers--often produce unreliable forecasts for future
periods, which the agency's regulatory analyses invariably require.
This occurs because they are often unresponsive to relevant shifts in
economic conditions or consumer preferences, and also because it is
difficult to incorporate factors such as the introduction of new model
offerings--particularly those utilizing advances in technology or
vehicle design--or shifts in manufacturers' pricing strategies into
their representations of choices and forecasts of future sales or
market shares. For these reasons, most vehicle choice models have been
better suited for analysis of the determinants of historical variation
in sales patterns than to forecasting future sales volumes and market
shares of particular categories.
Although these challenges have so far precluded the agency from
employing a discrete choice model in its regulatory analyses, we
believe they are not insurmountable and recognize the considerable
advantages such a model could offer.\595\ Thus, the agency intends to
continue its attempts to develop some suitable variant of such a model
for use in future fuel economy rulemakings.
---------------------------------------------------------------------------
\595\ For an additional overview of the challenges of employing
a discrete choice model, see TSD Section 4.2.1.
---------------------------------------------------------------------------
(b) Scrappage
New and used vehicles are substitutes. When the price of a good's
substitute increases (decreases), the demand curve for that good shifts
upwards (downwards) and the equilibrium price and quantity supplied
also increases (decreases). Thus, increasing the quality-adjusted price
of new vehicles will result in an increase in equilibrium price and
quantity of used vehicles. Since, by definition, used vehicles are not
being ``produced'' but rather ``supplied'' from the existing fleet, the
increase in quantity must come via a reduction in their scrappage
rates. Practically, when new vehicles become more expensive, demand for
used vehicles increases (and they become more expensive). Because used
vehicles are more valuable in such circumstances, they are scrapped at
a lower rate, and just as rising new vehicle prices push marginal
prospective buyers into the used vehicle market, rising used vehicle
prices force marginal prospective buyers of used vehicles to acquire
older vehicles or vehicles with fewer desired attributes. The effect of
fuel economy standards on scrappage is partially dependent on how
consumers value future fuel savings and our assumption that consumers
value only the first 30 months of fuel savings.
Many competing factors influence the decision to scrap a vehicle,
including the cost to maintain and operate it, the household's demand
for VMT, the cost of alternative means of transportation,
[[Page 25862]]
and the value that can be attained through reselling or scrapping the
vehicle for parts. A car owner will decide to scrap a vehicle when the
value of the vehicle is less than the value of the vehicle as scrap
metal, plus the cost to maintain or repair the vehicle. In other words,
the owner gets more value from scrapping the vehicle than continuing to
drive it, or from selling it. Typically, the owner that scraps the
vehicle is not the first owner.
While scrappage decisions are made at the household level, the
agency is unaware of sufficient household data to capture scrappage at
that level. Instead, the agency uses aggregate data measures that
capture broader market trends. Additionally, the aggregate results are
consistent with the rest of the CAFE Model as the model does not
attempt to model how manufacturers will price new vehicles; the model
instead assumes that all regulatory costs to make a particular vehicle
compliant are passed onto the purchaser who buys the vehicle. It is
more likely that manufacturers will defray a portion of the increased
regulatory cost across its vehicles or to other manufacturers' buyers
through the sale of credits.
The most predictive element of vehicle scrappage is ``engineering
scrappage.'' This source of scrappage is largely determined by the age
of a vehicle and the durability of a specific model year vintage. The
agency uses proprietary vehicle registration data from IHS/Polk to
compute vehicle age and durability for each model year or vintage.
Other factors affecting scrappage include fuel economy and new vehicle
prices. For historical data on new vehicle transaction prices, the
agency uses National Automobile Dealers Association (NADA) data.\596\
These data consist of the average transaction price of all light-duty
vehicles; since the transaction prices are not broken-down by body
style, the model may miss unique trends within a particular vehicle
body style. The transaction prices are the amount consumers paid for
new vehicles and exclude any trade-in value credited towards the
purchase. This may be particularly relevant for pickup trucks, which
have experienced considerable changes in average price as luxury and
high-end options entered the market over the past decade. Future models
will further consider incorporating price series that represent the
price trends for cars, SUVs and vans, and pickups separately. Vehicle
scrappage is also influenced by cyclical market trends, which the model
captures using forecasts of GDP and fuel prices.
---------------------------------------------------------------------------
\596\ The data can be obtained from NADA. For reference, the
data for MY 2020 may be found at https://www.nada.org/nadadata/.
---------------------------------------------------------------------------
Vehicle scrappage follows a roughly ``S-shaped'' pattern with
increasing age--that is, when a model year (or ``vintage'') is
relatively new few vehicles of its age are scrapped; progressively more
are retired as they age and accumulate use, but after some age
retirements again slow. Although fewer and fewer of the vehicles
originally produced during a model year remain on the road as they age,
the annual rate at which they are retired typically reaches a peak
sometime around age 20 and declines gradually after that.\597\ The
agency's model employs a logistic function to capture this relationship
of vehicle scrappage rates to age.
---------------------------------------------------------------------------
\597\ The retirement rate is usually measured by the number of
vehicles originally produced during a model year that are retired
during a subsequent (calendar) year, expressed as a fraction of the
number that remained in use at its outset.
---------------------------------------------------------------------------
Historical registration data show that vehicles produced during
more recent model years generally last longer than those from earlier
vintages, indicating that the durability of successive model years has
improved over time, although there are occasional exceptions to this
broader pattern. Annual scrappage rates for vehicles produced during
more recent model years are also observed to be lower than those of
earlier vintages up to a certain age, but are necessarily higher after
that age to account for the fact that the share of original vehicles
remaining in use ultimately converges toward the minimal share (zero,
in the extreme) observed for earlier vintages.\598\
---------------------------------------------------------------------------
\598\ Examples of why durability may have changed are new
automakers entering the market or general changes to manufacturing
practices like switching some models from a car chassis to a truck
chassis. The agency caps model years' lifetimes at 40 years in its
accounting; by that age a slightly larger share of each successive
model year tends to remain in use, although this share so far
remains below 2 percent of those originally produced.
---------------------------------------------------------------------------
The agency includes indicator variables for each model year in its
scrappage model to capture these historical improvements in vehicles'
durability over successive model years. Additionally, to ensure that
vehicles approaching the end of their assumed 40-year service life are
retired, the agency applies a decay function to the number remaining in
use after they reach age 30. Retirement rates for individual model
years are modeled primarily as a polynomial function of age to capture
the non-linear shape described above. The effective change in new
vehicle prices projected in the model (defined as technology costs
minus 30 months of fuel savings, as discussed previously) is also
included in the model, which produces differing scrappage rates across
regulatory alternatives since each one includes different estimates of
technology costs and fuel savings. Finally, the model also includes
year-to-year differences in U.S. GDP (to capture the effects of
macroeconomic cycles on owners' decisions to keep older vehicles in
use), fuel prices, and fuel costs for used vehicles of each age, as
well as the share of vehicles originally produced during each model
year remaining in use.
In addition to the variables included in the scrappage model, the
agency considered several other variables that may influence scrappage
in the real world including, maintenance and repair costs, the value of
scrapped metal, vehicle characteristics, the quantity of new vehicles
purchased, higher interest rates, and unemployment. These variables
were excluded from the model either because of a lack of underlying
data or modeling constraints. Their exclusion from the model is not
intended to reflect their unimportance, but rather highlights the
practical constraints of modeling intricate decisions like scrappage.
The agency received some comments on modeling approaches that could
explicitly represent interactions between the new and used vehicle
markets, such as the influence of prices for new models on demand for
used vehicles (and the reverse), and the relationship between scrappage
rates and consumers' decisions about replacing retired vehicles (e.g.,
Jacobsen as discussed in Section III.E.2.a) and FRIA Chapter 4.3.2). On
scrappage rates specifically, the American Fuel & Petrochemical
Manufacturers (AFPM) cautioned the agency against overestimating
scrappage rates, highlighting the effect of current macroeconomic
conditions on new and used car prices and thus on owners' decision to
retire used vehicles.\599\ While we agree with the assertion of AFPM
that scrappage rates are important in accurately representing fleet
turnover and the resulting composition of the light duty vehicle fleet,
the agency found it difficult to quantitatively isolate the effect of
economic conditions on short-term scrappage decisions from longer term
trends in vehicle durability and other factors affecting retirement
rates when developing its scrappage model. For this reason, NHTSA has
elected to maintain the existing treatment of scrappage for this rule,
but will continue to monitor
[[Page 25863]]
research related to both short- and long-term scrappage patterns in the
vehicle fleet.
---------------------------------------------------------------------------
\599\ AFPM, Docket No. NHTSA-2021-0053-1530, at 18.
---------------------------------------------------------------------------
Changes in Vehicle Miles Traveled (VMT)
The anticipated level of future vehicle use, usually measured by
the number of vehicle-miles driven annually (VMT), directly influences
most of the effects of raising fuel economy standards that decision-
makers consider in determining what standards to establish. Most
important, the amount and value of fuel saved by requiring new cars and
light trucks to achieve higher fuel economy both depend on the number
of miles they are driven each year over their lifetimes, as well as of
course on how much raising CAFE standards improves their fuel economy
and on future fuel prices. Similarly, critical indirect impacts from
raising fuel economy standards such as changes in emissions of criteria
air pollutants and greenhouse gases, potential increases in fatalities
and injuries, and congestion levels also depend directly on the
consequences of higher standards for vehicle use.
NHTSA's CAFE Model estimates total yearly VMT as the product of
average annual usage per vehicle and the number of vehicles making up
each future year's fleet, which itself depends on new vehicle sales
during the current and previous years and owners' decisions about when
to retire used vehicles. Since cars and light trucks of different model
years (or ``vintages'') and body styles will experience different cost
increases and varying increases in their fuel economy when CAFE
standards are raised--particularly when standards increase over a
succession of model years--the costs necessary to achieve their
required fuel economy levels as well as the resulting fuel savings and
indirect benefits will differ. Vehicles originally produced during a
model year are gradually retired and the usage of those remaining in
service tends to decline as they age (at least on average), so fuel
savings and other benefits from requiring them to achieve higher fuel
economy also decline gradually over their lifetimes. In any future
calendar year, the contributions of progressively older model years to
total benefits will also decline gradually, since fewer will remain in
use and those that do will be driven less, although this pattern will
also be affected by the increases in fuel economy required for earlier
model years.\600\
---------------------------------------------------------------------------
\600\ A vehicle's age during a future calendar year is equal to
that calendar year minus the model year in which it was originally
produced (and assumed to be sold); for example, model year 2020 cars
and light trucks will be 10 or 11 years old during calendar year
2030, depending on whether they were considered to be 0 or 1 year
old during 2020. (The agency's analysis uses the former convention,
so as an illustration, model year 2010 vehicles are considered to be
11 years old during 2020.)
---------------------------------------------------------------------------
Thus, accounting properly for the effects of vehicle use on the
costs and benefits from establishing higher CAFE standards requires
estimates of VMT in each future calendar year accounted for by vehicles
of different types and original model years (which determines their
current age during that year). The agency estimates VMT by vehicles of
different types and ages during future calendar years as the product of
the number of vehicles of each type and age in service during that year
and their average annual use. Because vehicles' annual use throughout
their lifetimes is influenced by their fuel economy--through its effect
on the cost of driving each mile--the VMT accounted for by vehicles of
each body type and model year will vary among regulatory alternatives
that require larger increases in fuel economy from its baseline level.
To develop estimates of average vehicle use by body type and model
year for future calendar years, the agency used odometer readings
collected at different dates for a very large sample of vehicles to
estimate average annual use at each age for cars and light trucks of
different body types (automobiles, SUVs/vans, and pickups). These
initial ``mileage accumulation schedules'' summarize how much vehicles
of each body type and age were driven during 2016, and provide a basis
to estimate how much vehicles produced during future model years will
be driven at each age throughout their lifetimes. As described in
detail in TSD Chapter 4.3, these initial schedules are adjusted to
incorporate the effects of both differences in fuel prices between 2016
and future calendar years, and differences in the fuel economy of
vehicles of each age during 2016 and those that will be of that same
during each future calendar years.
The agency's CAFE Model uses the estimates of future sales of new
cars and light trucks and annual retirement rates for used vehicles of
different ages constructed as described previously to project the
number of vehicles of each type and age that will be in use during each
future calendar year it analyzes. It combines these with the estimates
of average vehicle use at each age for different vehicle types to
calculate their total VMT and uses the shares operating on different
fuels (gasoline, diesel, and electricity) and their on-road fuel
efficiency to estimate total consumption of each fuel. Finally, the
model applies per-mile and per-gallon emission rates to estimate total
emissions accounted for vehicles of each type and age during future
calendar years. For more aggregate reporting of costs and benefits, the
agency sums these estimates to obtain total vehicle use, fuel
consumption, emissions, and other measures by vehicle type in each
calendar year, as well as lifetime travel, fuel use, emissions, etc.
for vehicles of each type and model year.
NHTSA's perspective is that total demand for car and light truck
travel should not vary significantly among the regulatory alternatives
it considers, since the basic travel demands of a typical household are
unlikely to be influenced much by the differences in vehicle prices or
driving costs likely to be associated with different CAFE standards.
However, the method the CAFE Model uses to calculate total VMT
described above (and in more detail in TSD Chapter 4.3), can create
modest differences in total VMT across the range of regulatory
alternatives, even without considering the potential effect of fuel
economy differences among those alternatives no vehicle use. These
arise from the effects of differences in new vehicle sales and
retirement rates for used vehicles among alternatives on the
composition of the vehicle fleet--its makeup by vehicle type and age or
original model year--during future years. Although small, these
differences in the representation of vehicle types and model years in
the future fleet can have significant impacts on the incremental costs
and benefits of different regulatory alternatives when those are
measured against the baseline.
To prevent the estimated effects of our standards from having
unrealistic implications for household vehicle ownership or travel
demand, the agency sought in this analysis to ensure that the fuel
consumption, emissions, safety, and other impacts it reports for
different regulatory alternatives reflect only differences in total
vehicle use that are specifically attributable to their differing fuel
economy requirements, and do not incorporate differences in the number
of cars and light trucks in use under each alternative. To do this the
CAFE Model constrains the level of future vehicle use under each
regulatory alternative before applying the fuel economy rebound effect
to match values projected using the Federal Highway Administration's
VMT forecasting model. In future years where this total ``pre-rebound
effect'' VMT calculated internally by the CAFE Model differs from the
FHWA forecast, each model year cohort's average VMT
[[Page 25864]]
is adjusted up or down so that the two estimates match. This process
ensures that any differences in total VMT among regulatory alternatives
is attributable to the fuel economy rebound effect. It also ensures
that the forecasts of total VMT for future years constructed using the
``bottom up'' process of estimating VMT separately for each vehicle
type and age and summing the results, as described immediately above,
are consistent with forecasts of aggregate VMT that are based on an
underlying theory of household travel demand and independent forecasts
of its demographic and economic determinants.
The agency's analysis of this final rule begins with the year 2020
and relies on actual data rather than forecasts for that year wherever
possible. The elements of the analysis that rely most heavily on
macroeconomic inputs--aggregate demand for VMT, new vehicle sales, and
used vehicle retirement rates--all reflect the economy's unexpectedly
rapid return to pre-pandemic levels of activity and expected future
growth, and these conditions prevail under each of the regulatory
alternatives considered. The Federal Highway Administration (FHWA)
publishes annual estimates of VMT for the light-duty vehicle fleet;
while FHWA's definition of light-duty vehicles differs slightly from
those subject to CAFE standards, over the period from 2016 through 2019
FHWA's estimates of VMT have agreed closely with those generated
internally by NHTSA's CAFE Model.601 602 In 2020, however,
the effects of the COVID pandemic--including sharply reduced demand for
travel and mandated travel restrictions--reduced light-duty VMT
significantly from its 2019 level, and this decline persisted through
much of 2021.
---------------------------------------------------------------------------
\601\ See Highway Statistics 2017, Table VM-1, available at
https://www.fhwa.dot.gov/policyinformation/statistics/2017/vm1.cfm.
(Accessed: March 15, 2022) FHWA's estimates of VMT include travel by
light-duty trucks up to 10,000 lbs. GVW, while the CAFE program
excludes trucks with GVWs exceeding 8,500 lbs. FHWA reported light-
duty VMT of 2.86 trillion for calendar year 2016, while NHTSA's
model generated an internal estimate of 2.85 trillion VMT by
vehicles subject to CAFE standards. The two estimates did not
compare as closely for subsequent years, but never differed by more
than 2 percent.
\602\ NHTSA's estimates of total VMT rely on estimates of
average annual mileage for light-duty vehicles at each age,
calibrated to 2016 data, together with the number of registered
light-duty vehicles at each age. Chapter 4 of the TSD accompanying
this rulemaking describes these data and the process NHTSA uses to
estimate total VMT in detail.
---------------------------------------------------------------------------
Although this downturn in travel activity was accurately reflected
in FHWA's published estimates of light-duty vehicle travel for the year
2020 and monthly travel volumes during 2021, it was not captured in the
VMT estimates produced internally by NHTSA's CAFE Model because those
rely on vehicle use and registration estimates that could not readily
be adjusted to account for sharply reduced commuting, shopping, and
recreational travel or for restrictions on vehicle use that were
imposed in some locations. To avoid the problems that relying on the
models' internally generated forecasts for 2020 and 2021 would have
caused, the agency's analysis for this final rule relied on FHWA's
published estimate of light-duty VMT for 2020 and extrapolated the
volumes reported in that agency's monthly travel updates through
October of 2021 to develop an estimate of annual VMT for 2021.
The fuel economy rebound effect--a specific example of the well-
documented energy efficiency rebound effect for energy-consuming
capital goods--refers to the tendency of motor vehicles' use to
increase when their fuel economy is improved and the fuel cost to drive
each mile declines as a result. A regulatory alternative that
establishes more stringent CAFE standards than those assumed to prevail
under the baseline scenario will increase the fuel economy of new cars
and light trucks, thereby reducing their pre-mile fuel consumption and
fuel costs and increasing the number of miles they are driven annually
over their lifetimes. The assumed magnitude of this fuel economy
rebound effect influences the overall costs and benefits associated
with each regulatory alternative considered, as well as the estimates
of its effects on fatalities and other safety measures. Thus, its
value--together with fuel prices, technology costs, and other
analytical inputs--is part of the body of information that agency
decision-makers have considered in selecting the CAFE standards this
final rule establishes. By magnifying the effect of higher fuel economy
on vehicle use, larger values of the fuel economy rebound effect also
reduce the economic and environmental benefits associated with
increased fuel efficiency.
The agency received a number of comments on the value of the
rebound effect. Most commenters argued that the agency rebound
selection of 15 percent was too high and suggested that the literature
supported a rebound magnitude ranging from 5 to 10 percent; most
commenters supported using a rebound of 10 percent.\603\ A few
commenters argued that an even lower value such as 5 percent should be
used instead.\604\ While Auto Innovators did not comment directly on
the agency's choice of 15 percent, it argued that the agency's estimate
of rebound did not take into consideration of ``attribute
substitution,'' whereby a household will buy a less fuel efficient
vehicle as their second vehicle and will make a decision on which
vehicle to use depending on the purpose for any particular trip.\605\
The agency notes that Auto Innovators did not provide any guidance on
the likely direction of this ``attribute substitution'' effect--which
is not clear a priori--in its comment, nor provide any suggestions for
how to account for it in the analysis.
---------------------------------------------------------------------------
\603\ See California Attorney General et al., Docket No. NHTSA-
2021-0053-1526-A1, at 2; UCS, Docket No. NHTSA-2021-0053-1567-A1, at
32; CBD et al., Joint Summary Comments, at 2-3; ICCT, A1, at 14;
Lucid, Docket No. NHTSA-2021-0053-1584-A1, at 6; IPI, at 35-37; and
CARB, Docket No NHTSA-2021-0053-1521-A2, at 2-3.
\604\ See e.g., CFA, Docket No. NHTSA-2021-0053-1535, at 4-5.
\605\ Auto Innovators, at 93-94.
---------------------------------------------------------------------------
ICCT commented in general support of the methodology used to
construct the vehicle mileage accumulation schedules, but suggested
that the agency could further improve them by considering how increased
durability of successive models could cause newer vehicles to be driven
more as they age than their older counterparts.\606\ The agency notes
that ICCT is correct that increased durability can increase VMT. NHTSA
captures this possibility in the scrappage model, where more recent
model years tend to be retained in service longer, and also in its
application of the fuel economy rebound effect, where vehicles
featuring higher fuel economy are assumed to be used more intensively
throughout their lifetimes. The agency notes that the data and methods
it used to develop the mileage accumulation schedules capture the
increasing durability of recent model year to some extent, because as
described in detail in TSD Chapter 4.3 those data include a range of
model years observed over several decades, and increased durability is
not a recent phenomenon. Treating model years as a ``panel'' when
estimating the pattern of vehicle use with age explicitly accounts for
both increases in the fraction of vehicles produced during successive
model years that remain in use at each age and any accompanying
increase in the average use of vehicles of different ages.
---------------------------------------------------------------------------
\606\ ICCT, at 22-23.
---------------------------------------------------------------------------
Several of the commenters also seemed to suggest that we should not
consider the impacts of rebound driving at all since they are freely
chosen.\607\ We note that rebound driving is an expected
[[Page 25865]]
result of this final rule, and that understanding how increased fuel
efficiency will affect additional mobility deserves consideration even
if there is an offsetting mobility benefit. In addition, the question
of whether and how to consider the rebound effect and its consequences
is an aspect of the agency's determination of what standard represents
the ``maximum feasible,'' which is a separate question from the more
technical issue of what the appropriate value for the rebound effect
should be in the analysis.
---------------------------------------------------------------------------
\607\ See, e.g., CBD et al., at 17.
---------------------------------------------------------------------------
As described in detail in TSD Chapter 4.3.5, the agency conducted a
thorough and detailed review of recent research on the fuel economy
rebound effect, which includes several new estimates it had not
previously considered and also incorporates statistical uncertainty
surrounding different estimates. The agency's updated review shows that
research measuring the response of vehicle use to fuel economy itself
suggests a rebound effect ranging from 5 to 15 percent, while studies
examining the association of vehicle use to fuel costs of driving
suggest that the rebound effect is most likely to lie in the range from
10 to 20 percent.
Based on this updated analysis, the agency selected a rebound
effect of 10 percent for this analysis, because it was well-supported
by the totality of the evidence and aligned closely with the response
of total vehicle use to fuel costs incorporated in FHWA's forecasting
model (approximately 14 percent). This value is also consistent with
the value used in EPA's recent final rule. To recognizing the wide
range of uncertainty surrounding the true value of the fuel economy
rebound effect, we also examine the sensitivity of estimated impacts to
values ranging from 5 to 20 percent.
To calculate levels of total light-duty that incorporate the fuel
economy rebound effect, the CAFE Model interprets the assumed magnitude
of the rebound effect as an elasticity of average vehicle use with
respect to fuel cost per mile, and applies this to changes in fuel
costs resulting from the higher fuel economy levels each regulatory
alternative requires. It then adds the resulting proportional increases
in average vehicle use to their values under the No-Action Alternative,
as previously adjusted to reconcile the CAFE Model's estimate of total
VMT with that produced by FHWA's travel forecasting model. TSD Chapter
4.3 provides an extensive discussion of how the agency calculates
changes in VMT to account for the rebound effect.
Jacobsen and Liao commented on the agency's procedures for
estimating VMT and incorporating the rebound effect, noting that while
still in progress, their recent research shows that by raising prices
for new cars and light trucks, higher CAFE standards increase the
depreciation cost their owners incur in driving each mile.\608\ They
assert that the response of vehicle use to higher per-mile
depreciations costs outweighs its response to the reduction in fuel
costs from required increases in their fuel economy, although they do
not report empirical results demonstrating this effect. These
commenters also argue that the reduction in sales of new vehicles in
response to higher new car and light truck prices will reinforce this
effect, because households owning fewer vehicles will drive less in
total as complementarity between the number of vehicles households own
and their trip-making frequency operates in reverse. They argue that as
these two effects interact with the usual fuel economy rebound effect,
higher CAFE standards will reduce average vehicle use on balance rather
than increasing it as the agency estimates.\609\
---------------------------------------------------------------------------
\608\ Jacobsen and Liao, NHTSA-2021-0053-0065, at 1.
\609\ Jacobsen and Liao, at 2.
---------------------------------------------------------------------------
The agency agrees that higher per-mile depreciation costs are
likely by themselves to reduce vehicle use but notes that only some
fraction of vehicles' total depreciation costs owes to their usage,
with the remainder attributable to the passage of time and
technological progress in new vehicle designs and utility. Empirical
estimates of this breakdown are scarce, so it is difficult to assess
how large the increase in per-mile depreciation costs associated with a
given increase in new vehicles' prices might be. We also note that
increasing durability of new cars and light trucks over time tends to
reduce the depreciation costs associated with their use, simply because
their lifetime use-related depreciation is distributed over a larger
number of miles. The agency notes further that the increases in new car
and light truck prices it estimates will occur as consequences of the
alternatives it considered for this analysis are quite modest,
particularly after they are adjusted to reflect their buyers' assumed
valuation of the higher fuel economy they provide. Combined with their
increased durability and the fact that only a fraction of their higher
prices is reflected in increased use-related depreciation, the implied
increases in their per-mile depreciation costs are likely to be
extremely small. Finally, we also note that empirical estimates of the
fuel economy rebound effect generally do not control for potential
increases in vehicles' purchase prices and accompanying depreciation
costs. As a consequence, the association between higher fuel economy
(or lower per-mile fuel costs) and higher per-mile depreciation is
likely to be incorporated to some extent in estimates the rebound
effect, in which case they can be interpreted as the combined or net
effect of these countervailing changes on vehicle use.
4. Changes to Fuel Consumption
The agency combines modeled fuel economy levels with age and body-
style VMT estimates to determine changes in fuel consumption over time
and across alternatives. The agency computes the amount of fuel
consumed by dividing expected total travel by predicted MPG at the
vehicle level and then aggregates to produce estimates of total fuel
consumed in each alternative.\610\
---------------------------------------------------------------------------
\610\ Total value of fuel consumed is computed across all fuel
types and draws fuel price values (e.g., retail prices for gasoline
and electricity) from the set of model inputs.
---------------------------------------------------------------------------
F. Simulating Environmental Impacts of Regulatory Alternatives
In estimating the environmental impacts of each regulatory
alternative we considered, the agency accounted for the projected
application of many fuel-saving technologies to vehicles that could
continue to use only gasoline or diesel fuel (including hybrid electric
vehicles that do not require external charging), as well as the
projected increased application of plug-in hybrid electric vehicles
and, with some analytical constraints, battery electric vehicles.\611\
By reducing overall energy consumption and the production and use of
petroleum-based fuels, the alternatives the agency considered would
thus have important consequences for the environment and public health.
These occur because each alternative would reduce tailpipe emissions of
both GHGs and criteria air pollutants during vehicle operation, as well
as ``upstream'' emissions that occur during petroleum extraction,
transportation, and refining to produce fuel, as well as during the
transportation, storage, and distribution of refined fuel. In turn,
reduced emissions of GHGs and air pollutants would improve
environmental quality, reduce the health consequences of
[[Page 25866]]
exposure to air pollution (whether climate-exacerbated or not), and
mitigate economic damages attributable to changes in the global climate
and air pollution levels.
---------------------------------------------------------------------------
\611\ This document and FRIA do not consider the potential for
manufacturers to respond to new standards for MYs 2024-2026 by
introducing new BEV models in MYs 2024-2026. However, the
accompanying Supplemental Environmental Impact Analysis (SEIS) does
account for such potential introductions of new BEV models in these
model years.
---------------------------------------------------------------------------
This section provides an overview of how we develop the assumptions
and parameters used to estimate emissions of criteria air pollutants,
greenhouse gases, and air toxics. It also describes how we develop and
apply estimates of the air quality and climate-related impacts of these
emissions and their consequences for human health, focusing
particularly on the rule's effects on emissions of criteria air
pollutants that cause poor air quality and can damage human health. The
agency's analysis utilizes the ``emissions inventory'' approach to
estimate these impacts. Vehicle-related emissions inventories are often
described as three-legged stools, since they depend on measures of
vehicle activity (i.e., miles traveled, hours operated, or gallons of
fuel burned), the number of vehicles in use, and emission factors per
unit of vehicle activity.
An emissions factor is a rate that measures the quantity of a
pollutant released to the atmosphere per unit of vehicle activity.\612\
This analysis relies on vehicle-miles traveled (VMT) as its measure of
vehicle activity, and emission rates are measured by emissions (in mass
units) per vehicle-mile; the vehicle-related or ``tailpipe'' emission
inventory for most pollutants is the product of their per-mile
emissions factor and the appropriate estimate of the number of miles
traveled. Exceptions include tailpipe emissions of sulfur oxides
(SOX) and carbon dioxide (CO2), which are
estimated by applying emissions factors per gallon of fuel consumed
derived from the chemical properties of different fuels to the
appropriate values of fuel consumption in gallons. Vehicle activity
levels--both the number of miles traveled and the number of gallons of
fuel consumed--are generated by the CAFE Model (as described in
Sections III.E.3. and 4. above), while the per-mile and per-gallon
emission factors have been extracted from other models developed by
other Federal agencies. In this rulemaking, vehicle-related emissions
also include those that occur throughout the process of supplying fuel
and other forms of vehicle energy (such as electric power), and these
are termed upstream emissions. The agency estimates these upstream
emissions from the volume or energy content of fuel supplied and
consumed by cars and light trucks, together with factors that express
emissions of air pollutants and GHGs in mass per unit of fuel volume
(usually grams per gallon) or fuel energy (e.g., grams per million Btu)
supplied. Total upstream emissions of each pollutant are estimated as
the product of the number of gallons of fuel supplied and the relevant
per-gallon emission factor, or as the product of total energy supplied
and emissions per unit of energy produced and delivered.
---------------------------------------------------------------------------
\612\ U.S. EPA, Basics Information of Air Emissions Factors and
Quantification, https://www.epa.gov/air-emissions-factors-and-quantification/basic-information-air-emissions-factors-and-quantification. (Accessed: March 15, 2022)
---------------------------------------------------------------------------
For this rule, vehicle tailpipe (sometimes called ``downstream'')
and upstream emission factors as well as estimates of total emissions
from both sources were developed independently using separate data
sources. Tailpipe emission factors are estimated from the highway
emissions model developed for use in regulatory analysis by the U.S.
Environmental Protection Agency's (EPA) National Vehicle and Fuel
Emissions Laboratory, known as the Motor Vehicle Emission Simulator
(MOVES). Upstream emission factors are estimated from a lifecycle
emissions model developed by the U.S. Department of Energy's (DOE)
Argonne National Laboratory, the Greenhouse Gases, Regulated Emissions,
and Energy Use in Transportation (GREET) Model.\613\ For this final
rule, we updated the CAFE Model to utilize data from the most current
versions of each model, MOVES 3 and GREET 2021.
---------------------------------------------------------------------------
\613\ U.S. Department of Energy, Argonne National Laboratory,
Greenhouse gases, Regulated Emissions, and Energy use in
Transportation (GREET) Model, Last Update: 11 Oct. 2021, https://greet.es.anl.gov/. (Accessed: March 15, 2022) Upstream emission
factors for criteria air pollutants may be undercounted, but are
nonetheless important.
---------------------------------------------------------------------------
Adverse human health outcomes caused by exposure to harmful
accumulations of criteria air pollutants, such as asthma episodes and
respiratory or cardiovascular distress requiring hospitalization, are
generally reported as incidences per ton of emissions of each pollutant
(or its chemical precursors). The incidence per ton values used to
estimate changes in health impacts were developed using several EPA
studies and recently updated to better account for the specific sources
of emissions estimated by the CAFE Model. Finally, EPA also applies
estimates of the affected population's willingness to pay to avoid each
incidence of these adverse health impacts and sums the results to
obtain estimates of the economic cost of air pollutant emissions in
dollars per ton, which can be interpreted as estimates of the economic
benefit from reducing each ton of emissions of different pollutants.
Chapter 5 of the TSD accompanying this final rule includes a detailed
discussion of the procedures we used to simulate the environmental
impacts of the different regulatory alternatives that were considered,
and the implementation of these procedures within the CAFE Model is
discussed in detail in the supporting Model Documentation. Further
discussion of how the health impacts of upstream and tailpipe emissions
of criteria air pollutants have been monetized and the resulting values
used in this analysis can be found in Section III.G.2.b)(2). The Final
SEIS accompanying this analysis also includes a detailed discussion of
both criteria pollutant and GHG emissions and their impacts on human
health as well as on the natural environment.
1. Activity Levels Used To Calculate Emissions Impacts
The CAFE Model estimates the annual number of miles driven (VMT)
for each individual car and light truck model produced in every future
model year at each age over their lifetimes, which extend for a maximum
of 40 years. Since a vehicle's age is equal to the current calendar
year minus the model year in which it was originally produced, the age
span of each vehicle model's lifetime corresponds to a sequence of 40
calendar years beginning in the calendar year corresponding to the
model year it was produced.\614\ These estimates reflect the gradual
decline in the fraction of each car and light truck model's original
model year production volume that is expected to remain in service
during each year of its lifetime, as well as the well-documented
decline in their typical use as they age. Using this relationship, the
CAFE Model calculates total VMT for cars and light trucks in service
during each calendar year spanned in this analysis.
---------------------------------------------------------------------------
\614\ In practice, many vehicle models bearing a given model
year designation become available for sale in the preceding calendar
year, and their sales can extend through the calendar year following
their designated model year as well. However, the CAFE Model does
not attempt to distinguish between model years and calendar years;
vehicles bearing a model year designation are assumed to be produced
and sold in that same calendar year.
---------------------------------------------------------------------------
Based on these estimates, the model also calculates quantities of
each type of fuel or energy, including gasoline, diesel, and
electricity, consumed in each calendar year. By combining these with
estimates of each model's fuel or energy efficiency, the model also
estimates the quantity and energy content of each type of fuel consumed
(including gasoline, diesel, and electricity) by cars and light trucks
at
[[Page 25867]]
each age, or viewed another way, during each calendar year of their
lifetimes. As with the accounting of VMT, these estimates of annual
fuel or energy consumption for each vehicle model and model year
combination are combined to calculate the total volume of each type of
fuel or energy consumed during each calendar year, as well as its
aggregate energy content.
The procedures the CAFE Model uses to estimate annual VMT for
individual car and light truck models produced during each model year
over their lifetimes and to combine these into estimates of annual
fleet-wide travel during each future calendar year, together with the
sources of its estimates of their survival rates and average use at
each age, are described in detail in Section III.E.2. The data and
procedures it employs to convert these estimates of VMT to fuel and
energy consumption by individual model, and to aggregate the results to
calculate total consumption and energy content of each fuel type during
future calendar years, are also described in detail in that same
section.
The model documentation accompanying this final rule also describes
these procedures in detail.\615\ The quantities of travel and fuel
consumption estimated for the cross section of model years and calendar
years constitutes a set of ``activity levels'' based on which the model
calculates emissions. The model does so by multiplying activity levels
by emission factors. As indicated in the previous section, the
resulting estimates of vehicle use (VMT), fuel consumption, and fuel
energy content are combined with emission factors drawn from various
sources to estimate emissions of GHGs, criteria air pollutants, and
airborne toxic compounds that occur throughout the fuel supply and
distribution process, as well as during vehicle operation, storage, and
refueling. Emission factors measure the mass of each GHG, or criteria
pollutant emitted per vehicle-mile of travel, gallon of fuel consumed,
or unit of fuel energy content. The following sections identifies the
sources of these emission factors and explains in detail how the CAFE
Model applies them to its estimates of vehicle travel, fuel use, and
fuel energy consumption to estimate total annual emissions of each GHG,
criteria pollutant, and airborne toxic.
---------------------------------------------------------------------------
\615\ CAFE Model documentation is available at https://www.nhtsa.gov/corporate-average-fuel-economy/compliance-and-effects-modeling-system.
---------------------------------------------------------------------------
2. Simulating Upstream Emissions Impacts
Building on the methodology for simulating upstream emissions
impacts used in prior CAFE rules, this final rule analysis uses
emissions factors developed with the U.S. Department of Energy's
Greenhouse gases, Regulated Emissions, and Energy use in Transportation
(GREET) Model, specifically GREET 2021.\616\ The analysis includes
emissions impacts estimates for regulated criteria pollutants,\617\
greenhouse gases,\618\ and air toxics.\619\
---------------------------------------------------------------------------
\616\ U.S. Department of Energy, Argonne National Laboratory,
Greenhouse gases, Regulated Emissions, and Energy use in
Transportation (GREET) Model, Last Update: 11 Oct. 2021, https://greet.es.anl.gov/.
\617\ Carbon monoxide (CO), volatile organic compounds (VOCs),
nitrogen oxides (NOX), sulfur oxides (SOX),
and particulate matter with 2.5-micron ([micro]m) diameters or less
(PM2.5).
\618\ Carbon dioxide (CO2), methane (CH4),
and nitrous oxide (N2O).
\619\ Acetaldehyde, acrolein, benzene, butadiene, formaldehyde,
diesel particulate matter with 10-micron ([micro]m) diameters or
less (PM10).
---------------------------------------------------------------------------
The upstream emissions factors included in the CAFE Model input
files include parameters for 2020 through 2050 in five-year intervals
(e.g., 2020, 2025, 2030, and so on). For gasoline and diesel fuels,
each analysis year includes upstream emissions factors for the four
following upstream emissions processes: Petroleum extraction, petroleum
transportation, petroleum refining, and fuel transportation, storage,
and distribution (TS&D). In contrast, the upstream electricity
emissions factor is only a single value per analysis year. We briefly
discuss the components included in each upstream emissions factor here,
and a more detailed discussion is included in Chapter 5 of the TSD
accompanying this rule and the CAFE Model Documentation.
The first step in the process for calculating upstream emissions
includes any emissions related to the extraction, recovery, and
production of petroleum-based feedstocks, namely conventional crude
oil, oil sands, and shale oils. Then, the petroleum transportation
process accounts for the transport processes of crude feedstocks sent
for domestic refining. The petroleum refining calculations are based on
the aggregation of fuel blendstock processes rather than the crude
feedstock processes, like the petroleum extraction and petroleum
transportation calculations. The final upstream process after refining
is the transportation, storage, and distribution (TS&D) of the finished
fuel product.
The upstream gasoline and diesel emissions factors are aggregated
in the CAFE Model based on the share of fuel savings leading to reduced
domestic oil fuel refining and the share of reduced domestic refining
from domestic crude oil.\620\ The CAFE Model applies a fuel savings
adjustment factor to the petroleum refining process and a combined fuel
savings and reduced domestic refining adjustment to both the petroleum
extraction and petroleum transportation processes for both gasoline and
diesel fuels and for each pollutant. These adjustments are consistent
across fuel types, analysis years, and pollutants, and are unchanged
from the previous CAFE analyses. Additional discussion of the
methodology for estimating the share of fuel savings leading to reduced
domestic oil refining is located in Chapter 6.2.4.4 of the TSD.
---------------------------------------------------------------------------
\620\ Upstream emissions are underestimated to the extent that
they do not account for any toxic pollutants (like mercury) and
criteria pollutants (i.e., from refining/production in Mexico/
Canada, as such pollutants can cross boundaries), as well as certain
greenhouse gas emissions, that originate outside the borders of the
United States and are attributable to changes in gasoline
consumption as a result of these standards.
---------------------------------------------------------------------------
Upstream electricity emissions factors are also calculated using
GREET 2021. GREET 2021 projects a national default electricity
generation mix for transportation use from the latest Annual Energy
Outlook (AEO) data.\621\ As discussed above, the CAFE Model uses a
single upstream electricity factor for each analysis year.
---------------------------------------------------------------------------
\621\ For this CAFE analysis, this was AEO 2021, released
February 3, 2021, https://www.eia.gov/outlooks/archive/aeo21.
---------------------------------------------------------------------------
The Environmental Defense Fund (EDF) submitted comments to the
Draft SEIS docket stating that NHTSA's estimates of reductions in
global GHG emissions associated with lower domestic consumption of
gasoline and diesel and its consequences for U.S. imports of crude
petroleum should incorporate empirical estimates of the specific
sources of U.S. imports that would be reduced and the rates of GHG
emissions associated with producing crude petroleum at each of those
sources and transporting it to the U.S. for refining.\622\
---------------------------------------------------------------------------
\622\ EDF, NHTSA-2021-0054-0016, at pp. 4-5.
---------------------------------------------------------------------------
We do not have the detailed production and supply modeling
capability that would be necessary to estimate reductions in U.S.
imports of crude petroleum from specific sources, and the global nature
of the market for crude petroleum suggests that those reductions are
unlikely to be proportional to the volumes currently imported from
different sources, as EDF
[[Page 25868]]
appears to assume. The global nature of the market for crude petroleum
also means that reductions in U.S. purchases from specific sources
would not necessarily be met by corresponding reductions in petroleum
production and associated GHG emissions at those locations, since those
producers' reduced exports to the U.S. might simply be redirected to
supply other purchasers.
In light of this situation, we believe the most reasonable
assumption to use for estimating reductions in global GHG emissions
associated with lower U.S. petroleum imports and global production is
to apply the emission factors associated with crude petroleum
production at different global locations and with current
transportation patterns, weighted by each location's projected
contribution to future global production. This is in fact the
assumption implicitly reflected in the agency's reliance on GHG
emission factors for crude petroleum transportation and distribution
derived using GREET. Even this assumption is likely to lead to an
overestimate of the reduction in global GHG emissions, since it implies
that the estimated decline in U.S. imports will be fully reflected in
an overall reduction in global petroleum production, rather than being
partly or fully absorbed by other oil-consuming nations. We have
therefore elected to retain this assumption and its current procedure
for estimating reduced GHG emissions from petroleum production. These
assumptions are discussed in further detail in Section 0.
EDF also commented that that NHTSA's estimates of reductions in
domestic emissions of criteria air pollutants resulting from lower U.S.
production and consumption of transportation fuels and its assumed
effect on U.S. petroleum imports should include reductions in emissions
that occur during the transportation of imported petroleum ``. . . on
U.S. soil or within established distances from our borders where
emissions still affect U.S. ambient air quality.'' This would include
emissions by tanker ships operating within U.S. Emission Control Areas
(ECAs, which can extend as far as 200 miles from U.S. shores),
including those to which petroleum is transferred when large oceangoing
tankers cannot enter some U.S. ports, as well as emissions by
petroleum-carrying barges, rail tank cars, and pipelines operating
within U.S. borders.
In fact, our analysis does include emissions that occur during
transportation of crude petroleum as domestic emissions associated with
petroleum imports. In effect, it assumes that transportation modes and
shipment distances for moving crude petroleum from U.S. coastal ports
to domestic refineries are similar to those for moving domestically
extracted crude petroleum from oilfields or other domestic petroleum
production facilities to U.S. refineries. Thus, some reductions in
emissions that occur during transportation of imported crude petroleum
within U.S. coastal and interior areas are included in the agency's
estimates of total reductions in domestic emissions of criteria
pollutants attributable to reduced U.S. petroleum imports. The agency
believes this approach provides a satisfactory substitute for detailed
estimation of movement distances and shipment modes for carrying
imported crude petroleum from ports to refineries. This is discussed
further in TSD Chapter 5.2 and TSD Chapter 6.2.4.2.
3. Simulating Tailpipe Emissions Impacts
Tailpipe emission factors are generated using a regulatory model
for on-road emission inventories from the U.S. Environmental Protection
Agency, the Motor Vehicle Emission Simulator (MOVES3), November 2020
release. MOVES3 is a state-of-the-science, mobile-source emissions
inventory model for regulatory applications.\623\ MOVES3 tailpipe
emission factors have been incorporated into the CAFE parameters, and
these updates supersede tailpipe data previously provided by EPA from
MOVES2014 for past CAFE analyses. MOVES3 accounts for a variety of
processes related to emissions impacts from vehicle use, examples
include exhaust and evaporative processes, among others.\624\
---------------------------------------------------------------------------
\623\ U.S. Environmental Protection Agency, Office of
Transportation and Air Quality, Motor Vehicle Emission Simulator
(MOVES), Last Updated: September 2021, https://www.epa.gov/moves/latest-version-motor-vehicle-emission-simulator-moves. For the CAFE
analysis, MOVES 3.0.1 was used to generate the emission factors.
\624\ For CAFE modeling, the post-processing of emission factors
for PM2.5 included exhaust processes (running, start,
crankcase running, and crankcase start) and excluded brake and tire
wear.
---------------------------------------------------------------------------
The CAFE Model uses tailpipe emissions factors for all model years
from 2020 to 2060 for criteria pollutants and air toxics. To maintain
continuity in the historical inventories, only emission factors for MYs
2020 and after were updated; all emission factors prior to MY 2020 were
unchanged from previous CAFE rulemakings. In addition, the updated
tailpipe data in the current CAFE reference case no longer account for
any fuel economy improvements or changes in vehicle miles traveled from
the 2020 final rule. In order to avoid double-counting effects from the
previous rulemaking in the current rulemaking, the tailpipe baseline
backs out 1.5 percent year-over-year stringency increases in fuel
economy, and 0.3 percent VMT increases assumed each year (20 percent
rebound on the 1.5 percent improvements in stringency). Note that the
MOVES3 data do not cover all the model years and ages required by the
CAFE Model; MOVES only generates emissions data for vehicles made in
the last 30 model years for each calendar year being run. This means
emissions data for some calendar year and vehicle age combinations are
missing. To remedy this, we take the last vehicle age that has
emissions data and forward fill those data for the following vehicle
ages. Due to incomplete available data for years prior to MY 2020,
tailpipe emission factors for MY 2019 and earlier have not been
modified and continue to utilize MOVES2014 data.
For tailpipe CO2 emissions, these factors are defined
based on the fraction of each fuel type's mass that represents carbon
(the carbon content) along with the mass density per unit of the
specific type of fuel. To obtain the emission factors associated with
each fuel, the carbon content is then multiplied by the mass density of
a particular fuel as well as by the ratio of the molecular weight of
carbon dioxide to that of elemental carbon. This ratio, a constant
value of 44/12, measures the mass of carbon dioxide that is produced by
complete combustion of mass of carbon contained in each unit of fuel.
The resulting value defines the emission factor attributed to
CO2 as the amount of grams of CO2 emitted during
vehicle operation from each type of fuel. This calculation is repeated
for gasoline, E85, diesel, and compressed natural gas (CNG) fuel types.
In the case of CNG, the mass density and the calculated CO2
emission factor are denoted as grams per standard cubic feet (scf),
while for the remainder of fuels, these are defined as grams per gallon
of the given fuel source. Since electricity and hydrogen fuel types do
not cause CO2 emissions to be emitted during vehicle
operation, the carbon content, and the CO2 emission factors
for these two fuel types are assumed to be zero. The mass density,
carbon content, and CO2 emission factors for each fuel type
are defined in the Parameters file.
The CAFE Model calculates CO2 tailpipe emissions
associated with vehicle operation of the surviving on-road fleet by
multiplying the number of gallons (or scf for CNG) of a specific fuel
consumed by the CO2 emissions factor
[[Page 25869]]
for the associated fuel type. More specifically, the amount of gallons
or scf of a particular fuel are multiplied by the carbon content and
the mass density per unit of that fuel type, and then the model applies
the ratio of carbon dioxide emissions generated per unit of carbon
consumed during the combustion process.\625\
---------------------------------------------------------------------------
\625\ Chapter 3, Section 4 of the CAFE Model Documentation
provides additional description for calculation of CO2
tailpipe emissions with the model.
---------------------------------------------------------------------------
4. Estimating Health Impacts From Changes in Criteria Pollutant
Emissions
The CAFE Model computes select health impacts resulting from three
criteria pollutants: NOX, SOX,\626\ and
PM2.5. Out of the six criteria pollutants currently
regulated, NOX, SOX, and PM2.5 are
known to be emitted regularly from mobile sources and have the most
adverse effects to human health. These health impacts include several
different morbidity measures, as well as a mortality estimate, and are
measured by the number of instances predicted to occur per ton of
emitted pollutant.\627\ The model reports total health impacts by
multiplying the estimated tons of each criteria pollutant by the
corresponding health incidence per ton value. The inputs that inform
the calculation of the total tons of emissions resulting from criteria
pollutants are discussed above. This section discusses how the health
incidence per ton values were obtained. See Section III.G.2.b)(2) and
Chapter 6.2.2 of the TSD accompanying this notice for information
regarding the monetized damages arising from these health impacts.
---------------------------------------------------------------------------
\626\ Any reference to SOX in this section refers to
the sum of sulfur dioxide (SO2) and sulfate particulate
matter (pSO4) emissions, following the methodology of the EPA papers
cited.
\627\ The complete list of morbidity impacts estimated in the
CAFE Model is as follows: acute bronchitis, asthma exacerbation,
cardiovascular hospital admissions, lower respiratory symptoms,
minor restricted activity days, non-fatal heart attacks, respiratory
emergency hospital admissions, respiratory emergency room visits,
upper respiratory symptoms, and work loss days.
---------------------------------------------------------------------------
The Final SEIS associated with this document also includes a
detailed discussion of the criteria pollutants and air toxics analyzed
and their potential health effects. Consistent with past analyses, we
have performed full-scale photochemical air quality modeling and
presented those results in the Final SEIS. That analysis provides
additional assessment of the human health impacts from changes in
PM2.5 and ozone associated with this rule. We note that
compliance with CAFE standards is based on the average performance of
manufacturers' production for sale throughout the U.S., and that the
FRIA involves sensitivity analysis spanning a range of model inputs,
many of which impact estimates of future emissions from passenger cars
and light trucks. Chapter 6 of the FRIA includes a discussion of
overall changes in health impacts associated with criteria pollutant
changes across the different rulemaking scenarios.
In previous rulemakings, health impacts were split into two
categories based on whether they arose from upstream emissions or
tailpipe emissions. In the current analysis, these health incidence per
ton values have been updated to reflect the differences in health
impacts arising from each emission source sector, according to the
latest publicly available EPA reports that appropriately correspond to
these sectors. Five different upstream emission source sectors
(petroleum extraction, petroleum transportation, refineries, fuel
transportation, storage and distribution, and electricity generation)
are now represented. The tailpipe source sector is now disaggregated
based on fuel and vehicle type. As the health incidences for the
different source sectors are all based on the emission of one ton of
the same pollutants, NOX, SOX, and
PM2.5, the differences in the incidence per ton values arise
from differences in the geographic distribution of the pollutants, a
factor which affects the number of people impacted by the
pollutants.\628\
---------------------------------------------------------------------------
\628\ See Environmental Protection Agency (EPA). 2018.
Estimating the Benefit per Ton of Reducing PM2.5
Precursors from 17 Sectors. https://www.epa.gov/sites/production/files/2018-02/documents/sourceapportionmentbpttsd_2018.pdf.
---------------------------------------------------------------------------
The CAFE Model health impacts inputs are based partially on the
structure of EPA's 2018 TSD, Estimating the Benefit per Ton of Reducing
PM2.5 Precursors from 17 Sectors (referred to here as the
2018 EPA source apportionment TSD),\629\ which reported benefit per ton
values for the years 2016, 2020, 2025, and 2030.\630\ For the years in
between the source years used in the input structure, the CAFE Model
applies values from the closest source year. For instance, 2020 values
are applied for 2020-2022, and 2025 values are applied for 2023-2027.
For further details, see the CAFE Model documentation, which contains a
description of the model's computation of health impacts from criteria
pollutant emissions.
---------------------------------------------------------------------------
\629\ Environmental Protection Agency (EPA). 2018. Estimating
the Benefit per Ton of Reducing PM2.5 Precursors from 17
Sectors. https://www.epa.gov/sites/production/files/2018-02/documents/sourceapportionmentbpttsd_2018.pdf.
\630\ As the year 2016 is not included in this analysis, the
2016 values were not used.
---------------------------------------------------------------------------
Despite efforts to be as consistent as possible between the
upstream emissions sectors utilized in the CAFE Model with the 2018 EPA
source apportionment TSD, the need to use up-to-date sources based on
newer air quality modeling updates led to the use of multiple papers.
In addition to the 2018 EPA source apportionment TSD used in the 2020
final rule, we used additional EPA sources and conversations with EPA
staff to appropriately map health incidence per ton values to the
appropriate CAFE Model emissions source category. Very recently, EPA
updated its approach to estimating the benefits of changes in
PM2.5 and ozone,631 632 as well as the associated
changes in health impacts per ton. These updates were based on
information drawn from the recent 2019 PM2.5 and 2020 Ozone
Integrated Science Assessments (ISAs), which were reviewed by the Clean
Air Science Advisory Committee (CASAC) and the
public.633 634 EPA has not updated its health incidence
estimates for mobile sources to reflect these updates in time for this
analysis. Instead, based on the recommendation of EPA staff, we use the
same PM2.5 BPT estimates and health incidence values that we
used in the NPRM, to ensure consistency between the values
corresponding to different source sectors. The estimates used are based
on the review of the 2009 PM ISA \635\ and 2012 p.m. ISA Provisional
Assessment \636\ and include
[[Page 25870]]
a mortality risk estimate derived from the Krewski et al. (2009) \637\
analysis of the American Cancer Society (ACS) cohort and nonfatal
illnesses consistent with benefits analyses performed for the analysis
of the final Tier 3 Vehicle Rule (79 FR 23414, April 28, 2014),\638\
the final 2012 p.m. NAAQS Revision (78 FR 3154, Jan. 15, 2013),\639\
and the final 2017-2025 Light-duty Vehicle GHG Rule (77 FR 62624, Oct.
15, 2012).\640\ We expect this lag in updating our health incidence and
BPT estimates to have only a minimal impact on total PM benefits, since
the underlying mortality risk estimate based on the Krewski study is
identical to an updated PM2.5 morality risk estimate derived
from an expanded analysis of the same ACS cohort. We are aware of EPA's
work to update its mobile source BPT and health incidence estimates to
reflect these recent updates for use in future rulemaking analyses, and
we will work further with EPA in future rulemakings to update and
synchronize approaches.
---------------------------------------------------------------------------
\631\ U.S. Environmental Protection Agency (U.S. EPA). 2021a.
Regulatory Impact Analysis for the Final Revised Cross-State Air
Pollution Rule (CSAPR) Update for the 2008 Ozone NAAQS. EPA-452/R-
21-002. March.
\632\ U.S. Environmental Protection Agency (U.S. EPA). 2021b.
Estimating PM2.5- and Ozone-Attributable Health Benefits.
Technical Support Document (TSD) for the Final Revised Cross-State
Air Pollution Rule Update for the 2008 Ozone Season NAAQS. EPA-HQ-
OAR-2020-0272. March.
\633\ U.S. Environmental Protection Agency (U.S. EPA). 2019a.
Integrated Science Assessment (ISA) for Particulate Matter (Final
Report, 2019). U.S. Environmental Protection Agency, Washington, DC,
EPA/600/R-19/188, 2019.
\634\ U.S. Environmental Protection Agency (U.S. EPA). 2019a.
Integrated Science Assessment (ISA) for Ozone and Related
Photochemical Oxidants (Final Report). U.S. Environmental Protection
Agency, Washington, DC, EPA/600/R-20/012, 2020.
\635\ U.S. Environmental Protection Agency (U.S. EPA). 2009.
Integrated Science Assessment for Particulate Matter (Final Report).
EPA-600-R-08-139F. National Center for Environmental Assessment-RTP
Division, Research Triangle Park, NC. December. Available at: http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=216546.
\636\ U.S. Environmental Protection Agency (U.S. EPA). 2012.
Provisional Assessment of Recent Studies on Health Effect of
Particulate Matter Exposure. EPA/600/R-12/056F. National Center for
Environmental Assessment-RTP Division, Research Triangle Park, NC.
December. Available at: https://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=247132.
\637\ Krewski D., M. Jerrett, R.T. Burnett, R. Ma, E. Hughes, Y.
Shi, et al. 2009. Extended Follow-Up and Spatial Analysis of the
American Cancer Society Study Linking Particulate Air Pollution and
Mortality. HEI Research Report, 140, Health Effects Institute,
Boston, MA.
\638\ U.S. Environmental Protection Agency (2014). Control of
Air Pollution from Motor Vehicles: Tier 3 Motor Vehicle Emission and
Fuel Standards Final Rule: Regulatory Impact Analysis, Assessment
and Standards Division, Office of Transportation and Air Quality,
EPA-420-R-14-005, March 2014. Available on the internet: http://www3.epa.gov/otaq/documents/tier3/420r14005.pdf.
\639\ U.S. Environmental Protection Agency. (2012). Regulatory
Impact Analysis for the Final Revisions to the National Ambient Air
Quality Standards for Particulate Matter, Health and Environmental
Impacts Division, Office of Air Quality Planning and Standards, EPA-
452-R-12-005, December 2012. Available on the internet: http://www3.epa.gov/ttnecas1/regdata/RIAs/finalria.pdf.
\640\ U.S. Environmental Protection Agency (U.S. EPA). (2012).
Regulatory Impact Analysis: Final Rulemaking for 2017-2025 Light-
Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average
Fuel Economy.
---------------------------------------------------------------------------
The basis for the health impacts from the petroleum extraction
sector is a 2018 oil and natural gas sector paper written by EPA staff
(Fann et al.), which estimates health impacts for this sector in the
year 2025.\641\ This paper defines the oil and gas sector's emissions
not only as arising from petroleum extraction but also from
transportation to refineries, while the CAFE/GREET component is
composed of only petroleum extraction. After consultation with the
authors of the EPA paper, we determined that these are the best
available estimates for the petroleum extraction sector,
notwithstanding this difference. Specific health incidences per
pollutant were not reported in the paper, so EPA staff sent BenMAP
health incidence files for the oil and natural gas sector upon request.
DOT staff then calculated per ton values based on these files and the
tons reported in the Fann et al. paper.\642\ The only available health
impacts corresponded to the year 2025. Rather than trying to
extrapolate, these 2025 values were used for all the years in the CAFE
Model structure: 2020, 2025, and 2030.\643\ This simplification implies
an overestimate of damages in 2020 and an underestimate in 2030.\644\
---------------------------------------------------------------------------
\641\ Fann, N., Baker, K. R., Chan, E., Eyth, A., Macpherson,
A., Miller, E., & Snyder, J. (2018). Assessing Human Health
PM2.5 and Ozone Impacts from U.S. Oil and Natural Gas
Sector Emissions in 2025. Environmental science & technology,
52(15), 8095-8103 (hereinafter, Fann et al.).
\642\ Nitrate-related health incidents were divided by the total
tons of NOX projected to be emitted in 2025, sulfate-
related health incidents were divided by the total tons of projected
SOX, and EC/OC (elemental carbon and organic carbon)
related health incidents were divided by the total tons of projected
EC/OC. Both Fann et al. and the 2018 EPA source apportionment TSD
define primary PM2.5 as being composed of elemental
carbon, organic carbon, and small amounts of crustal material. Thus,
the EC/OC BenMAP file was used for the calculation of the incidents
per ton attributable to PM2.5.
\643\ These three years are used in the CAFE Model structure
because it was originally based on the estimate provided in the 2018
EPA source apportionment TSD.
\644\ See EPA. 2018. Estimating the Benefit per Ton of Reducing
PM2.5 Precursors from 17 Sectors. https://www.epa.gov/sites/production/files/2018-02/documents/sourceapportionmentbpttsd_2018.pdf, p. 9.
---------------------------------------------------------------------------
We understand that uncertainty exists around the contribution of
VOCs to PM2.5 formation in the modeled health impacts from
the petroleum extraction sector; however, based on feedback to the 2020
final rule, we believe that the updated health incidence values
specific to petroleum extraction sector emissions may provide a more
appropriate estimate of potential health impacts from that sector's
emissions than the previous approach of applying refinery sector
emissions impacts to the petroleum extraction sector. For further
discussion of the BPT estimates corresponding to the health effects
discussed in this section, see Section III.G.2.b)(2).
The petroleum transportation sector and fuel TS&D sector do not
correspond to any one EPA source sector in the 2018 EPA source
apportionment TSD, so we use a weighted average of multiple different
EPA sectors to determine the health impact per ton values for those
sectors. We use a combination of different EPA mobile source sectors
from two different papers, the 2018 EPA source apportionment TSD,\645\
and a 2019 mobile source sectors paper (Wolfe et al.) \646\ to generate
these values. The health incidence per ton values associated with the
refineries sector and electricity generation sector are drawn solely
from the 2018 EPA source apportionment TSD.
---------------------------------------------------------------------------
\645\ Environmental Protection Agency (EPA). 2018. Estimating
the Benefit per Ton of Reducing PM2.5 Precursors from 17
Sectors. https://www.epa.gov/sites/production/files/2018-02/documents/sourceapportionmentbpttsd_2018.pdf.
\646\ Wolfe et al. 2019. Monetized health benefits attributable
to mobile source emissions reductions across the United States in
2025. https://pubmed.ncbi.nlm.nih.gov/30296769/.
---------------------------------------------------------------------------
IPI expressed concern that the agency's domestic fuel refining
share assumptions cause an underestimate in the health effects counted
in this analysis.\647\ For discussion of NHTSA's domestic fuel refining
assumptions, see Section III.G.2.b)(3), TSD Chapter 5.2, and TSD
Chapter 6.2.
---------------------------------------------------------------------------
\647\ IPI, Docket No. NHTSA-2021-0053-1579, at 39.
---------------------------------------------------------------------------
The CAFE Model follows a similar process for computing health
impacts resulting from tailpipe emissions as it does for calculating
health impacts from upstream emissions. Previous rulemakings used the
2018 EPA source apportionment TSD as the source for the health
incidence per ton, matching the CAFE Model tailpipe emissions inventory
to the ``on-road mobile sources sector'' in the TSD. However, a more
recent EPA paper from 2019 (Wolfe et al.) \648\ computes monetized
damage costs per ton values at a more disaggregated level, separating
on-road mobile sources into multiple categories based on vehicle type
and fuel type. Wolfe et al. did not report incidences per ton, but that
information was obtained through communications with EPA staff. The
Center for Biological Diversity, Chesapeake Bay Foundation,
Conservation Law Foundation, Earthjustice, Environmental Law & Policy
Center, Natural Resources Defense Council, Public Citizen, Inc., Sierra
Club, and Union of Concerned Scientists, in their joint summary
comments, stated that the estimates of the benefits of PM2.5
reductions have been improved with the addition of the Wolfe et al.
paper.\649\ We agree, and continue to use these sources in the final
rulemaking analysis as the categories are more expansive and specific
than the original 2018 source.
---------------------------------------------------------------------------
\648\ Wolfe et al. 2019. Monetized health benefits attributable
to mobile source emissions reductions across the United States in
2025. https://pubmed.ncbi.nlm.nih.gov/30296769/.
\649\ CBD et al., Docket No. NHTSA-2021-0053-1572, at 5.
---------------------------------------------------------------------------
The Wisconsin Department of Natural Resources (WDNR) stated that
``NHTSA
[[Page 25871]]
should work with EPA to offset any increases in sulfur dioxide
emissions associated with the rule'' and that ``NHTSA should work with
EPA to offset any short-term increases in NOX and VOC
emissions associated with the rule,'' specifically citing the on-road
emissions that contribute to ozone formation in Wisconsin. Furthermore,
they state that ``NHTSA's analysis should be updated to reflect EPA's
revised area designations for the 2015 ozone NAAQs.'' \650\
---------------------------------------------------------------------------
\650\ WDNR, Docket No. NHTSA-2021-0053-0059, at 2, 4.
---------------------------------------------------------------------------
While this final rulemaking will result in small short-term
increases in criteria pollutants, the number of vehicle re-fueling
events and emissions of certain criteria pollutants and precursors the
emissions impact will vary from area to area depending on factors such
as the composition of the local vehicle fleet and the amount of
gasoline produced in the area. As discussed further in the Final SEIS,
criteria pollutant impacts are by their nature diffuse and
indeterminate, which makes the assessment of any potential mitigation
measures difficult; however, NHTSA does not have jurisdiction to
regulate criteria and air toxic pollutant emissions. However, as
discussed further in the Final SEIS, NHTSA did update the Final SEIS
analysis to reflect EPA's revised area designations for the 2015 ozone
NAAQS, including nonattainment area designations in Wisconsin and the
Chicago area.
The Alliance for Automotive Innovation and CEI expressed the
concern that the analysis overstates health effects. The Alliance
argued that reductions in PM2.5 emissions ``will not provide
public health benefits that are additive to the emissions reductions
accomplished by EPA's mobile-source and stationary-source programs for
criteria air pollutants.'' \651\ CEI objected to counting benefits from
a reduction in PM emissions in areas that are not classified as
nonattainment areas.\652\ As EPA stated in their recent GHG final rule
for MYs 2023-2026 (86 FR 74434, Dec. 30, 2021),\653\ NAAQS are set with
an ``adequate margin of safety'' but this ``does not represent a zero-
risk standard.'' As such, it is important to count health benefits from
reductions in criteria pollutants, regardless of whether they occur in
nonattainment areas or not. Furthermore, the relative magnitude of the
health benefits in our analysis is minimal compared to the other costs
and benefits and does not significantly change net benefits.
---------------------------------------------------------------------------
\651\ Auto Innovators, Docket No. NHTSA-2021-0053-1492, at 90.
\652\ Competitive Enterprise Institute, Docket No. NHTSA-2021-
0053-1546, at 3.
\653\ EPA. Revised 2023 and Later Model Year Light-Duty Vehicle
Greenhouse Gas Emissions Standards: Response to Comments (EPA-420-R-
21-027, December 2021) pp. 15-31.
---------------------------------------------------------------------------
We are aware of other limitations of using national values of
health incidences per ton associated with the BPT approach, which we
discuss extensively in prior rules, the NPRM, and Chapter 5 of the TSD.
That said, we believe that the BPT approach provides a reasonable
estimate of how different levels of CAFE standards may impact public
health.
The methodology for generating values for each emissions category
in the CAFE Model is discussed in further detail in Chapter 5 of the
TSD. The Parameters file contains all of the health impact per ton of
emissions values used in this final rule.
G. Simulating Economic Impacts of Regulatory Alternatives
This section summarizes the agency's approach for measuring the
economic costs and benefits that will result from establishing
alternative CAFE standards for future model years. The benefit and cost
measures the agency uses are important considerations, because as
Office of Management and Budget (OMB) Circular A-4 states, benefits and
costs reported in regulatory analyses must be defined and measured
consistently with economic theory, and should also reflect how
alternative regulations are anticipated to change the behavior of
producers and consumers from a baseline scenario.\654\ For CAFE
standards, those include vehicle manufacturers, buyers of new cars and
light trucks, owners of used vehicles, and suppliers of fuel, all of
whose behavior is likely to respond in complex ways to the level of
CAFE standards that DOT establishes for future model years.
---------------------------------------------------------------------------
\654\ White House Office of Management and Budget, Circular A-4:
Regulatory Analysis, September 17, 2003 (https://obamawhitehouse.archives.gov/omb/circulars_a004_a-4/), Section E.
---------------------------------------------------------------------------
It is important to report the benefits and costs of this final rule
in a format that conveys useful information about how those impacts are
generated and also distinguishes the impacts of those economic
consequences for private businesses and households from the effects on
the remainder of the U.S. economy. A reporting format will accomplish
this objective to the extent that it clarifies who incurs the benefits
and costs of the final rule, and shows how the economy-wide or
``social'' benefits and costs of the final rule are composed of its
direct effects on vehicle producers, buyers, and users, plus the
indirect or ``external'' benefits and costs it creates for the general
public.
Table III-37 and Table III-38 present the incremental economic
benefits and costs of the final rule and the alternatives (described in
detail in Section IV) to increase CAFE standards for MYs 2024-26 at
three percent and seven percent discount rates in a format that is
intended to meet these objectives. The tables include costs that are
transfers between different economic actors--these will appear as both
a cost and a benefit in equal amounts (to separate affected parties).
Societal cost and benefit values shown elsewhere in this document do
not show costs that are transfers for the sake of simplicity but report
the same net societal costs and benefits. The final rule and the
alternatives would increase costs to manufacturers for adding
technology necessary to enable new cars and light trucks to comply with
fuel economy and emission regulations. It may also increase fine
payments by manufacturers who would have achieved compliance with the
less demanding baseline standards. Manufacturers are assumed to
transfer these costs on to buyers by charging higher prices; although
this reduces their revenues, on balance, the increase in compliance
costs and higher sales revenue leaves them financially unaffected.
Since the analysis assumes that manufacturers are left in the same
economic position regardless of the standards, they are excluded from
the tables.
---------------------------------------------------------------------------
\655\ Average SC-GHG values are constructed using a 3 percent
discount rate and are discounted back to present value using a 3
percent discount rate.
---------------------------------------------------------------------------
BILLING CODE 4910-59-P
[[Page 25872]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.105
[[Page 25873]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.106
BILLING CODE 4910-59-C
Compared to the baseline standards, the analysis shows that buyers
of new cars and light trucks will incur higher purchasing prices and
financing costs, which will lead to some buyers dropping out of the new
vehicle market. Drivers of new vehicles will also experience a slight
uptick in the risk of being injured in a crash because of mass
reduction technologies employed to meet the increased standards. While
this effect is not statistically significant, NHTSA provides these
results for transparency, and to demonstrate that their inclusion does
not affect NHTSA's policy decision. Because of the increasing price of
new vehicles, some owners may delay retiring and replacing their older
vehicles with newer models. In effect, this will transfer some driving
that would have been done in newer vehicles under the baseline scenario
to older models within the legacy fleet, thus increasing costs for
injuries (both fatal and less severe) and property damages sustained in
motor vehicle crashes. This stems from the fact that cars and light
trucks have become progressively more protective in crashes over time
(and also slightly less prone to certain types of crashes, such as
rollovers). Thus, shifting some travel from newer to older models would
increase injuries and damages sustained by drivers and passengers
because they are traveling in less safe vehicles and not because it
changes the risk profiles of drivers themselves. These costs are
largely driven by assumptions regarding consumer valuation of fuel
efficiency and an assumption that more fuel-efficient vehicles are less
preferable to consumers than their total cost to improve fuel economy.
The agency examines alternate assumptions regarding consumer valuation,
as well as other assumptions that influence our safety impact estimates
in a sensitivity analysis that can be found in the accompanying FRIA.
In exchange for these costs, consumers will benefit from new cars
and light trucks with better fuel economy. Drivers will experience
lower costs as a consequence of new vehicles' decreased fuel
consumption, and from fewer refueling stops required because of their
increased driving range. They will experience mobility benefits as they
[[Page 25874]]
use newly purchased cars and light trucks more in response to their
lower operating costs. On balance, consumers of new cars and light
trucks produced during the model years subject to this final rule will
experience significant economic benefits.
Table III-37 and Table III-38 also show that the changes in fuel
consumption and vehicle use resulting from this final rule will in turn
generate both benefits and costs to society writ large. These impacts
are ``external,'' in the sense that they are by-products of decisions
by private firms and individuals that alter vehicle use and fuel
consumption but are experienced broadly throughout society rather than
by the firms and individuals who indirectly cause them. In terms of
costs, additional driving by consumers of new vehicles in response to
their lower operating costs will increase the external costs associated
with their contributions to traffic delays and noise levels in urban
areas, and these additional costs will be experienced throughout much
of the society. While most of the risk of additional driving or
delaying purchasing a newer vehicle are internalized by those who make
those decisions, a portion of the costs are borne by other road users.
Finally, since owners of new vehicles will be consuming less fuel, they
will pay less in fuel taxes.
Society will also benefit from more stringent standards. Increased
fuel efficiency will reduce the amount of petroleum-based fuel consumed
and refined domestically, which will decrease the emissions of carbon
dioxide and other greenhouse gases that contribute to climate change,
and, as a result, the U.S. (and the rest of world) will avoid some of
the economic damages from future changes in the global climate.
Similarly, reduced fuel production and use will decrease emissions of
more localized air pollutants (or their chemical precursors), and the
resulting decrease in the U.S. population's exposure to harmful levels
of these pollutants will lead to lower costs from its adverse effects
on health. Decreasing consumption and imports of crude petroleum for
refining lower volumes of gasoline and diesel will also create some
benefits throughout the U.S., in potential gains in energy security as
businesses and households that are dependent on fuel are less subject
to sudden and sharp changes in energy prices.
On balance, Table III-37 and Table III-38 show that both consumers
and society as a whole will experience net economic benefits from the
final rule. The following subsections will briefly describe the
economic costs and benefits considered by the agency. For a complete
discussion of the methodology employed and the results, see TSD Chapter
6 and FRIA Chapter 6, respectively. The safety implications of the
final rule--including the monetary impacts--are addressed in Section
III.H.
1. Private Costs and Benefits
(a) Costs to Consumers
(1) Technology Costs
The final rule and the alternatives would increase costs to
manufacturers for adding technology necessary to enable new cars and
light trucks to comply with fuel economy and emission regulations.
Manufacturers are assumed to transfer these costs on to buyers by
charging higher prices. See Section III.C.6 and TSD Chapter 2.6.
(2) Consumer Sales Surplus
Buyers who would have purchased a new vehicle with the baseline
standards in effect but decide not to do so in response to the changes
in new vehicles' prices due to more stringent standards in place will
experience a decrease in welfare. The collective welfare loss to those
``potential'' new vehicle buyers is measured by the forgone consumer
surplus they would have received from their purchase of a new vehicle
in the baseline.
Consumer surplus is a fundamental economic concept and represents
the net value (or net benefit) a good or service provides to consumers.
It is measured as the difference between what a consumer is willing to
pay for a good or service and the market price. OMB Circular A-4
explicitly identifies consumer surplus as a benefit that should be
accounted for in cost-benefit analysis. For instance, OMB Circular A-4
states the ``net reduction in total surplus (consumer plus producer) is
a real cost to society,'' and elsewhere elaborates that consumer
surplus values be monetized ``when they are significant.'' \656\
---------------------------------------------------------------------------
\656\ OMB Circular A-4, at 37-38.
---------------------------------------------------------------------------
Accounting for the portion of fuel savings that the average new
vehicle buyer demands, and holding all else equal, higher average
prices should depress new vehicle sales and by extension reduce
consumer surplus. The inclusion of consumer surplus is not only
consistent with OMB guidance, but with other parts of the regulatory
analysis. For instance, we calculate the increase in consumer surplus
associated with increased driving that results from the decrease in the
cost per mile of operation under more stringent regulatory
alternatives, as discussed in Section III.G.1.b)(3). The surpluses
associated with sales and additional mobility are inextricably linked
as they capture the direct costs and benefits accrued by purchasers of
new vehicles. The sales surplus captures the welfare loss to consumers
when they forgo a new vehicle purchase in the presence of higher prices
and the additional mobility measures the benefit increased mobility
under lower operating expenses.
The agency estimates the loss of sales surplus based on the change
in quantity of vehicles projected to be sold after adjusting for
quality improvements attributable to fuel economy. For additional
information about consumer sales surplus, see TSD Chapter 6.1.2.
(3) Ancillary Costs of Higher Vehicle Prices
Some costs of purchasing and owning a new or used vehicle scale
with the value of the vehicle. Where fuel economy standards increase
the transaction price of vehicles, they will affect both the absolute
amount paid in sales tax and the average amount of financing required
to purchase the vehicle. Further, where they increase the MSRP, they
increase the appraised value upon which both value-related registration
fees and a portion of insurance premiums are based. The analysis
assumes that the transaction price is a set share of the MSRP, which
allows calculation of these factors as shares of MSRP.
For this final rule, NHTSA has revised its estimates of these
ancillary costs to correct some mistakes in their accounting. First,
NHTSA excludes financing costs from the per-vehicle analysis. The
availability of vehicle financing is, if anything, a benefit to
consumers that would lower the cost to consumers of fuel-economy
technology by spreading out the costs over time. Second, NHTSA has
reduced its estimate of insurance costs to avoid a double-counting
issue it identified. Specifically, a portion of the insurance premium
goes to covering replacement vehicles and including that portion of the
insurance cost would be duplicative with estimates of the upfront
technology cost on the replacement vehicle (which is already captured
in the analysis and discussed above). For a detailed explanation of how
the agency estimates these costs, see TSD Chapter 6.1.1.
These costs are included in the consumer per-vehicle cost-benefit
analysis but are not included in the societal cost-benefit analysis
because they are assumed to be transfers from
[[Page 25875]]
consumers to governments, financial institutions, and insurance
companies.
(b) Benefits to Consumers
(1) Fuel Savings
The primary benefit to consumers of increasing CAFE standards are
the additional fuel savings that accrue to new vehicle owners. Fuel
savings are calculated by multiplying avoided fuel consumption by fuel
prices. Each vehicle of a given body style is assumed to be driven the
same as all the others of a comparable age and body style in each
calendar year. The ratio of that cohort's VMT to its fuel efficiency
produces an estimate of fuel consumption. The difference between fuel
consumption in the baseline, and in each alternative, represents the
gallons (or energy) saved. Under this assumption, our estimates of fuel
consumption from increasing the fuel economy of each individual model
depend only on how much its fuel economy is increased, and do not
reflect whether its actual use differs from other models of the same
body type. Neither do our estimates of fuel consumption account for
variation in how much vehicles of the same body type and age are driven
each year, which appears to be significant (see TSD Chapter 4.3.2).
Consumers save money on fuel expenditures at the average retail fuel
price (fuel price assumptions are discussed in detail in TSD Chapter
4.1.2), which includes all taxes and represents an average across
octane blends. For gasoline and diesel, the included taxes reflect both
the Federal tax and a calculated average state fuel tax. Expenditures
on alternative fuels (E85 and electricity, primarily) are also included
in the calculation of fuel expenditures, on which fuel savings are
based. And while the included taxes net out of the social benefit cost
analysis (as they are a transfer), consumers value each gallon saved at
retail fuel prices including any additional fees such as taxes. See TSD
Chapter 6.1.3 for additional details. In the TSD, the agency considers
the possibility that several of the assumptions made about vehicle use
could lead to imprecision in projecting fuel savings. The agency notes
that these simplifying assumptions are necessary to model fuel savings
and likely have minimal impact to the accuracy of this analysis.
CBD et al. commented that NHTSA underestimates the fuel savings in
the analysis. CBD et al. argued that NHTSA needs to account for any
fuel savings that may be achieved if CAFE standards cause gasoline
prices to fall due to decreasing demand.\657\ The agency acknowledges
that if fuel prices do decrease as a result of this rule, the analysis
could understate the amount of fuel savings. However, given how
pervasive fuel price projections are within the analysis, other
estimates would be incorrect as well. For example, our model assumes
that manufacturers will apply technology if the fuel savings in the
first 30 months exceeds the technology costs. If prices drop as a
result of better fuel economy, our standards would have a larger,
negative impact on sales as fewer technology costs are `worth it' in
the eyes of consumers. It is not readily apparent, then, whether
holding fuel prices constant across alternatives would increase or
decrease the net benefits attributable to the standards. Modeling fuel
prices that respond dynamically is currently outside the ability of the
model. Furthermore, since fuel prices are influenced by many different
factors--many of which are outside the purview of United States--it's
not clear if modeling gas prices dynamically would enhance the agency's
analysis.
---------------------------------------------------------------------------
\657\ CBD et al., Appendix, Docket No. NHTSA-2021-0053-1572, at
31.
---------------------------------------------------------------------------
(2) Refueling Benefit
Increasing CAFE standards, all else being equal, affects the amount
of time drivers spend refueling their vehicles in several ways. First,
they increase the fuel economy of ICE vehicles produced in the future,
which increases vehicle range and decreases the number of refueling
events for those vehicles. Conversely, to the extent that more
stringent standards increase the purchase price of new vehicles, they
may reduce sales of new vehicles and scrappage of existing ones,
causing more VMT to be driven by older and less efficient vehicles,
which require more refueling events for the same amount of VMT driven.
Finally, sufficiently stringent standards may also change the number of
electric vehicles that are produced, and shift refueling to occur at a
charging station or at a residence, rather than at the pump--changing
per-vehicle lifetime expected refueling costs.
We estimate these savings by calculating the amount of refueling
time avoided--including the time it takes to find, refuel, and pay--and
multiplying it by DOT's value of time of travel savings estimate. For a
full description of the methodology, refer to TSD Chapter 6.1.4.
(3) Additional Mobility
Any increase in travel demand provides benefits that reflect the
value to drivers and other vehicle occupants of the added--or more
desirable--social and economic opportunities that become accessible
with additional travel. Under the alternatives in this analysis, the
fuel cost per mile of driving would decrease as a consequence of the
higher fuel economy levels they require, thus increasing the number of
miles that buyers of new cars and light trucks would drive as a
consequence of the well-documented fuel economy rebound effect.
The fact that drivers and their passengers elect to make more
frequent or longer trips to gain access to these opportunities when the
cost of driving declines demonstrates that the benefits they gain by
doing so exceed the costs they incur. At a minimum, the benefits must
equal the cost of the fuel consumed to travel the additional miles (or
they would not have occurred). The cost of that energy is subsumed in
the simulated fuel expenditures, so it is necessary to account for the
benefits associated with those miles traveled here. But the benefits
must also offset the economic value of their (and their passengers')
travel time, other vehicle operating costs, and the economic cost of
safety risks due to the increase in exposure that occurs with
additional travel. The amount by which the benefits of this additional
travel exceeds its economic costs measures the net benefits drivers and
their passengers experience, usually referred to as increased consumer
surplus.
TSD Chapter 6.1.5 explains the agency's methodology for calculating
additional mobility. The benefit of additional mobility over and above
its costs is measured by the change in consumers' surplus. This is
calculated using the rule of one-half, and is equal to one-half of the
change in fuel cost per mile times the increase in vehicle miles
traveled due to the rebound effect.
In contrast to the societal cost-benefit analysis, calculation of
average costs and benefits to consumers is done on a per-vehicle basis
and is intended to describe how alternative standards affect the costs
and benefits of owning vehicles from the consumers' perspective. The
mobility costs and benefits per vehicle are affected by the assumption
that total VMT before adding the rebound effect will be the same in the
baseline and all alternative cases (See TSD Chapter 4.3.1). Because the
standards affect vehicle sales and scrappage which changes the number
of vehicles in the alternative cases, the
[[Page 25876]]
CAFE Model changes VMT per vehicle in the alternative cases to maintain
a constant total non-rebounded VMT. When vehicle sales decrease in the
alternative cases, VMT per vehicle increases. IPI and Drs. Jacobsen and
Liao of the University of California at San Diego (UCSD) commented that
changes in the size and age composition of the vehicle stock will
change total VMT.\658\ IPI suggested VMT will change only ``slightly,''
while the UCSD commenters suggest reallocating only 50 percent of the
difference in non-rebounded VMT between the baseline and alternative
cases. We recognize that the assumption of constant non-rebounded VMT
is an approximation, and we may consider the possibility of refining
this method in the future.
---------------------------------------------------------------------------
\658\ IPI, at 30; Jacobsen and Liao, at 2.
---------------------------------------------------------------------------
When the size of the vehicle stock decreases in the alternative
cases, VMT and fuel cost per vehicle increase. Because maintaining
constant non-rebounded VMT assumes consumers are willing to pay the
full cost of the reallocated vehicle miles, we offset the increase in
fuel cost per vehicle by adding the product of the reallocated VMT and
fuel cost per mile to the mobility value. This corrects an error in the
NPRM per vehicle analysis, which included the fuel cost per vehicle of
reallocated miles but not the mobility benefit per vehicle. Because we
do not estimate other changes in cost per vehicle that could result
from the reallocated miles (e.g., maintenance, depreciation, etc.) we
do not estimate the portion of the transferred mobility benefits that
would correspond to consumers' willingness to pay for those costs. We
do not estimate the consumers' surplus associated with the reallocated
miles because there is no change in total non-rebounded VMT and thus no
change in consumers' surplus per consumer.
2. External Costs and Benefits
(a) Costs
(1) Congestion and Noise
Increased vehicle use associated with the rebound effect also
contributes to increased traffic congestion and highway noise. Although
drivers obviously experience these impacts themselves, they do not
fully value the costs these impacts impose on other road users and
surrounding residents, just as they do not fully value the emissions
impacts of their own driving. Congestion and noise costs are largely
``external'' to the vehicle owners whose decisions about how much,
where, and when to drive more in response to changes in fuel economy
create these costs. Thus, unlike changes in the fuel costs drivers
incur or the safety risks they assume when they decide to travel more,
changes in congestion and noise costs are not offset by corresponding
changes in the benefits drivers experience by making more frequent
trips or traveling to more distant destinations.
While largely external to individual drivers, congestion costs are
limited to road users as a whole; since road users include a
significant fraction of the U.S. population, however, we treat changes
in congestion costs as part of this rule's broader economic impacts on
society instead of as a private cost to those whose choices impose it.
Costs resulting from road and highway noise are even more widely
dispersed, because they are borne partly by surrounding residents,
pedestrians, and other non-road users, and for this reason are also
considered as a cost to the society as a whole.
To estimate the economic costs associated with changes in
congestion and noise caused by differences in miles driven for the
proposal, NHTSA updated FHWA's 1997 Highway Cost Allocation Study's
estimates of marginal congestion costs to reflect changes in three
factors that affect them: The time delays caused by the contribution of
additional travel to congestion, increases in typical vehicle
occupancy, and the hourly value of each occupant's time. The agency
assumed that delay per additional mile driven by cars and light trucks
has increased in proportion to growth in annual vehicle travel per
lane-mile of road and highway capacity in urban areas (where virtually
all congestion occurs) since the date of the original FHWA study. Noise
costs per additional mile driven were assumed to remain constant at
their levels originally estimated by the FHWA study. Both congestion
and noise costs were also updated to reflect changes in the economy-
wide price level since their original publication and make them
comparable to other economic values used in this analysis. The agency
previously relied on this study in its 2010 (75 FR 25324, May 7, 2010),
2011 (76 FR 57106, Sept. 15, 2011), and 2012 (77 FR 62624, Oct. 15,
2012) final rules, and, like the estimates used in the proposal, a
revised version for the 2020 final rule (85 FR 24174, April 30, 2020).
Updating the individual underlying components for congestion costs in
this analysis improves their currency and internal consistency with the
rest of the analysis.
Some commenters objected to the agency's use of increases in
vehicle volumes per mile of roadway to approximate the change in the
incremental contribution to congestion and delays caused by additional
car and light truck use. For example, CARB argued the revised values
led the analysis to overestimate congestion costs. CARB claimed that
the miscalculation arises from the scaling of vehicles per lane
``because (1) it compares a figure for passenger cars to a figure for
light-duty vehicles that includes sport-utility vehicles and vans, and
(2) it is limited to interstate highways instead of all roads.'' \659\
CARB further argued that the revised numbers do not account for changes
in average speeds and improved road designs. California Attorney
General et al. concurred with CARB's comment and suggested using the
1997 estimates updated only for inflation.\660\
---------------------------------------------------------------------------
\659\ CARB, Attachment 2, NHTSA-2021-0053-1521, at 13.
\660\ California Attorney General et al., Detailed Comments,
NHTSA-2021-0053-1499, at 32.
---------------------------------------------------------------------------
The agency disagrees with CARB's argument for several reasons.
First, the agency's scaling of vehicle-miles per lane-mile uses figures
that include all vehicle classes rather than those for light-duty
vehicles alone. SUVs had only begun to enter the fleet in 1997; since
then, they have increasingly substituted for passenger cars, and travel
by both cars and SUVs is included in the figures that the agency
compares for 1997 and more recent years.\661\ Today's SUVs are used
interchangeably with passenger cars, and it is more than reasonable to
assume that an additional SUV mile will produce the same marginal
increase in congestion costs as an additional passenger car mile.
---------------------------------------------------------------------------
\661\ See, e.g., Tom Voelk, Rise of S.U.V.s: Leaving Cars in
Their Dust, With No Signs of Slowing, N.Y. Times, May 21, 2020,
available at https://www.nytimes.com/2020/05/21/business/suv-sales-best-sellers.html.
---------------------------------------------------------------------------
Second, the original 1997 FHWA estimate of congestion costs and the
scaling that NHTSA used to update it both apply to all roads and
highways, and this comparison is consistent with the approach NHTSA has
taken across the last 5 rulemakings. Third, the comment did not explain
the expected direction of changes in speed or provide support for the
commenter's claim that better road design has mitigated the effect of
increased traffic volumes on travel speeds. Further, the commenter's
claims are difficult to reconcile: If we assume that better roads
enable higher speeds despite increased traffic volumes, more frequent
(and possibly more severe) crashes would result, and
[[Page 25877]]
incidents are an important contributor to congestion.\662\
---------------------------------------------------------------------------
\662\ See, e.g., https://safety.fhwa.dot.gov/speedmgt/ref_mats/fhwasa1304/Resources3/08%20-%20The%20Relation%20Between%20Speed%20and%20Crashes.pdf. The agency
also notes that if the average speed has increased, then our safety
costs would require adjustment as well.
---------------------------------------------------------------------------
In response to these comments, the agency also analyzed changes in
estimates of congestion delays reported by the Texas Transportation
Institute (TTI), which are widely cited, use well-documented methods,
and offer the only available measure of long-term trends in the
economic costs of traffic congestion and delays.\663\ TTI's estimates
of congestion delays are derived using well-established patterns of
travel throughout the day and relationships between vehicle travel
volumes and travel speeds for major roads and highways, and more
recently on highly detailed measures of actual hourly travel speeds and
vehicle volumes. The agency's calculations using TTI's detailed
historical database show that from 1997 (the date of the original FHWA
study) through 2017 (the end year used in the agency's update), person-
hours of delay per vehicle-mile traveled increased 57 percent in the
Nation's 100 largest urban areas and 52 percent in all (nearly 500)
U.S. urban areas. More suggestively, incremental hours of delay per
additional vehicle-mile traveled--a more direct measure of the impact
of additional travel on congestion delays and one more comparable to
that reported in the 1997 FHWA study--grew by 86 percent in the largest
areas and by 131 percent in all U.S. urban areas over that same period.
These calculations suggest that the 58 percent increase in person-hours
of delay per additional vehicle-mile of travel reflected in the
agency's updated estimate of incremental congestion costs is
reasonable, so the agency has elected to retain its earlier estimate.
---------------------------------------------------------------------------
\663\ For an overview and links to detailed reports and
documentation, see https://mobility.tamu.edu/umr/.
---------------------------------------------------------------------------
(2) Fuel Tax Revenue
As discussed previously in III.G.1.b)(1), a significant fraction of
the fuel savings experienced by consumers includes avoided fuel taxes,
which average nearly $0.50 per gallon when Federal, state, and local
excise and sales taxes levied on gasoline are included. Fuel taxes are
treated as a transfer within the agency's analysis, which includes an
offsetting loss in revenue to government agencies as a cost of raising
CAFE standards, and thus do not affect net benefits from this rule; the
agency reports this offsetting loss to illustrate the potential impact
on government agencies that rely on fuel tax revenue to support the
activities they fund.\664\
---------------------------------------------------------------------------
\664\ See OMB Circular A-4 for more information on transfer
payments, and how they should be accounted for in regulatory
analysis.
---------------------------------------------------------------------------
CFA erroneously commented that lost gasoline taxes were improperly
included--for the first time--as a cost of the rule.\665\ Not only have
both EPA and NHTSA previously reported changes in gasoline tax payments
by consumers and in revenues to government agencies, but NHTSA's
proposal explains in multiple places that gasoline taxes are considered
a transfer--a cost to governments and an identical benefit to consumers
that has already been accounted for in reported fuel savings--and has
no impact on net benefits. In contrast, Walter Kreucher commented that
billions in gasoline tax revenue would be lost if we finalized stricter
standards.\666\ As indicated above, however, any reduction in tax
revenue received by governments that levy taxes on fuel is exactly
offset by lower fuel tax payments by consumers, so from an economy-wide
standpoint reductions in gasoline tax revenues are simply a transfer of
economic resources and has no effect on net benefits.
---------------------------------------------------------------------------
\665\ CFA, Docket No. NHTSA-2021-0053-1535, at 5.
\666\ Walt Kreucher, Docket No. NHTSA-2021-0053-0013, at 14.
---------------------------------------------------------------------------
(b) Benefits
(1) Reduced Climate Damages
Extracting and transporting crude petroleum, refining it to produce
transportation fuels, and distributing fuel all generate additional
emissions of GHGs and criteria air pollutants beyond those from cars'
and light trucks' use of fuel. By reducing the volume of petroleum-
based fuel produced and consumed, adopting higher CAFE standards will
thus mitigate global climate-related economic damages caused by
accumulation of GHGs in the atmosphere, as well as the more immediate
and localized health damages caused by exposure to criteria pollutants.
Because they fall broadly on the U.S. population--and globally, in the
case of climate damages--reducing them represents an external benefit
from requiring higher fuel economy. The following subsections discuss
the values used to estimate the economic consequences associated with
climate damages and the discount rates applied to those benefits.
(a) Valuation of the Social Cost of Greenhouse Gases
In the proposal, NHTSA estimated the global social benefits from
the reductions in emissions of CO2, CH4, and
N2O expected to result from this rule using the SC-GHG
estimates presented in ``Technical Support Document: Social Cost of
Carbon, Methane, and Nitrous Oxide Interim Estimates under Executive
Order 13990'' (``February 2021 TSD''). These SC-GHG estimates are
interim global values developed pursuant to E.O. 13990 for use in
benefit-cost analysis.
The SC-GHG estimates used in our analysis were developed over many
years, using a transparent process, peer-reviewed methodologies, and
input from the public. Specifically, in 2009, an interagency working
group (IWG) that included experts from the DOT and other executive
branch agencies and offices was established to support agencies in
using the most comprehensive available science and to promote
consistency in the SC-GHG values used across agencies. The IWG
published SCC estimates in 2010 that were developed using three peer-
reviewed Integrated Assessment Models relating CO2 and other
GHG emissions to climate change and its potential economic impacts, and
updated these estimates in 2013 using new versions of each IAM. In
August 2016, the IWG published estimates of the social cost of methane
(SC-CH4) and nitrous oxide (SC-N2O) using
methodologies consistent with the underlying the SCC estimates. E.O.
13990 (issued on January 20, 2021) re-established an IWG, and directed
it to publish interim SC-GHG values for CO2, CH4,
and N2O within thirty days. Furthermore, the E.O. tasked the
IWG with updating the methodologies used in calculating these SC-GHG
values. The E.O. instructed the IWG to utilize ``the best available
economics and science,'' and incorporate principles of ``climate risk,
environmental justice, and intergenerational equity.'' \667\ The E.O.
also instructed the IWG to take into account the recommendations from
the NAS committee convened on this topic published in 2017.\668\ The
February 2021 TSD provides a complete discussion of the IWG's initial
review conducted under E.O. 13990, and
[[Page 25878]]
NHTSA incorporates that discussion by reference into this preamble.
---------------------------------------------------------------------------
\667\ Executive Order on Protecting Public Health and the
Environment and Restoring Science to Tackle the Climate Crisis.
(2021). Available at https://www.whitehouse.gov/briefing-room/presidential-actions/2021/01/20/executive-order-protecting-public-health-and-environment-and-restoring-science-to-tackle-climate-crisis/.
\668\ National Academy of Sciences (NAS). (2017). Valuing
Climate Damage: Updating Estimation of the Social Cost of Carbon
Dioxide. Available at https://www.nap.edu/catalog/24651/valuing-climate-damages-updating-estimation-of-the-social-cost-of.
---------------------------------------------------------------------------
NHTSA is using the IWG's interim values, published in February 2021
in a technical support document, for this CAFE analysis.\669\ As a
member of the IWG, DOT has thoroughly reviewed the inputs and
methodological choices for these estimates, and DOT affirms that, in
its expert judgment, the Interim Estimates reflect the best available
science and economics and are the most appropriate values to use in the
analysis of this rule. This use of the IWG estimates is the same
approach as that taken in DOT regulatory analyses between 2009 and
2016, and is consistent with the proposal.
---------------------------------------------------------------------------
\669\ Interagency Working Group on Social Cost of Greenhouse
Gases, United States Government. (2021). Technical Support Document:
Social Cost of Carbon, Methane, and Nitrous Oxide Interim Estimates
under Executive Order 13990, available at https://www.whitehouse.gov/wp-content/uploads/2021/02/TechnicalSupportDocument_SocialCostofCarbonMethaneNitrousOxide.pdf.
---------------------------------------------------------------------------
NHTSA indicated in the NPRM that if the Interagency Working Group
issued revised estimates of climate damages in time for NHTSA to
evaluate whether to incorporate them into this final rule, NHTSA would
consider using them. The IWG has not issued revised estimates.
The following section provides further discussion of the discount
rates that NHTSA uses in its regulatory analysis. For a full discussion
of the agency's quantification of GHGs, see TSD Chapter 6.2.1 and the
FRIA.
(b) Discount Rates for Climate Related Benefits
A standard function of regulatory analysis is to evaluate tradeoffs
between impacts that occur at different points in time. Many Federal
regulations involve costly upfront investments that generate future
benefits in the form of reductions in health, safety, or environmental
damages. To evaluate these tradeoffs, the analysis must account for the
social rate of time preference--the broadly observed social preference
for benefits that occur sooner versus those that occur further in the
future.\670\ This is accomplished by discounting impacts that occur
further in the future more than impacts that occur sooner.
---------------------------------------------------------------------------
\670\ This preference is observed in many market transactions,
including by savers that expect a return on their investments in
stocks, bonds, and other equities; firms that expect positive rates
of return on major capital investments; and banks that demand
positive interest rates in lending markets.
---------------------------------------------------------------------------
OMB Circular A-4 affirmed the appropriateness of accounting for the
social rate of time preference in regulatory analyses and recommended
discount rates of 3 and 7 percent for doing so. The recommended 3
percent discount rate was chosen to represent the ``consumption rate of
interest'' approach, which discounts future costs and benefits to their
present values using the rate at which consumers appear to make
tradeoffs between current consumption and equal consumption
opportunities deferred to the future. OMB Circular A-4 reports an
inflation-adjusted or ``real'' rate of return on 10-year Treasury notes
of 3.1 percent between 1973 and its 2003 publication date and
interprets this as approximating the rate at which society is
indifferent between consumption today and in the future.
The 7 percent rate reflects the opportunity cost of capital
approach to discounting, where the discount rate approximates the
forgone return on private investment if the regulation were to divert
resources from capital formation.\671\ OMB Circular A-4 cites pre-tax
rates of return on capital as part of its selection of the 7 percent
rate.\672\ The IWG rejected the use of the opportunity cost of capital
approach to discounting reductions in climate-related damages,
concluding that the ``consumption rate of interest is the correct
discounting concept to use when future damages from elevated
temperatures are estimated in consumption-equivalent units as is done
in the IAMs used to estimate the SC-GHG (National Academies 2017).''
\673\ In fact, Circular A-4 indicates that discounting at the
consumption rate of interest is the ``analytically preferred method''
when effects are presented in consumption-equivalent units.\674\ DOT
concurs that in light of Circular A-4's guidance on discount rates
spanning displacement of investments and/or consumption, and given the
considerations that climate damages are modeled in consumption
equivalent units and intergenerational equity, the use of consumption
based discount rates is superior for estimating SC-GHG.
---------------------------------------------------------------------------
\671\ As the IWG explained, use of the 7 percent opportunity
cost of capital approach in fact ``at best creat[es] a lower bound
on the estimate of net benefits that would only be met in an extreme
case where regulatory costs fully displace investment.'' Interagency
Working Group on Social Cost of Greenhouse Gases, United States
Government, Technical Support Document: Social Cost of Carbon,
Methane, and Nitrous Oxide, Interim Estimates under Executive Order
13990, February 2021. NHTSA agrees and observes that this rule does
not represent such an ``extreme case.'' NHTSA's analysis assumes
that most of the rule's costs and benefits, including technology
costs passed through to consumers, will affect consumption choices.
The focus on consumption rates is therefore especially appropriate.
\672\ OMB Circular A-4.
\673\ Interagency Working Group on Social Cost of Greenhouse
Gases, United States Government, Technical Support Document: Social
Cost of Carbon, Methane, and Nitrous Oxide, Interim Estimates under
Executive Order 13990, February 2021.
\674\ OMB, Circular A-4. See also Declaration of Dominic J.
Mancini. Submitted in Support of Defendants' Motion for a Stay
Pending Appeal, Louisiana v. Biden, Case No. 2:21-cv-01074-JDC-KK
(W.D. La., filed Feb. 19, 2022) (confirming the appropriateness of
this approach to discounting).
---------------------------------------------------------------------------
As the IWG states, ``GHG emissions are stock pollutants, where
damages are associated with what has accumulated in the atmosphere over
time, and they are long lived such that subsequent damages resulting
from emissions today occur over many decades or centuries depending on
the specific greenhouse gas under consideration.'' \675\ OMB Circular
A-4 states that impacts occurring over such intergenerational time
horizons require special treatment:
---------------------------------------------------------------------------
\675\ Ibid.
Special ethical considerations arise when comparing benefits and
costs across generations. Although most people demonstrate time
preference in their own consumption behavior, it may not be
appropriate for society to demonstrate a similar preference when
deciding between the well-being of current and future generations.
Future citizens who are affected by such choices cannot take part in
making them, and today's society must act with some consideration of
their interest.\676\
---------------------------------------------------------------------------
\676\ OMB Circular A-4.
Furthermore, NHTSA notes that in 2015, OMB--along with the rest of
the IWG--articulated that ``Circular A-4 is a living document, which
may be updated as appropriate to reflect new developments and
unforeseen issues,'' and that ``the use of 7 percent is not considered
appropriate for intergenerational discounting. There is wide support
for this view in the academic literature, and it is recognized in
Circular A-4 itself.'' \677\ Following this statement from OMB, and in
light of the need to weigh welfare to current and future generations,
it would be inappropriate to apply an opportunity cost of capital rate
to estimate SC-GHG.
---------------------------------------------------------------------------
\677\ Interagency Working Group on the Social Cost of Carbon,
United States Government, Response to Comments: Social Cost of
Carbon for Regulatory Impact Analysis under Executive Order 12866,
July 2015. Note that OMB, as a co-chair of the IWG, published the
request for comments.
---------------------------------------------------------------------------
In addition to the ethical considerations, Circular A-4 also
identifies uncertainty in long-run interest rates as another reason why
it is appropriate to use lower rates to discount intergenerational
impacts, since recognizing such uncertainty causes the appropriate
discount rate to decline gradually over progressively longer time
horizons. Circular A-4 also acknowledges the difficulty in estimating
appropriate discount rates for
[[Page 25879]]
``intergenerational'' time horizons, noting that ``[p]rivate market
rates provide a reliable reference for determining how society values
time within a generation, but for extremely long time periods no
comparable private rates exist.'' \678\ The social costs of distant
future climate damages--and by implication, the value of reducing them
by lowering emissions of GHGs--are highly sensitive to the discount
rate, and the present value of reducing future climate damages grows at
an increasing rate as the discount rate used in the analysis declines.
This ``non-linearity'' means that even if uncertainty about the exact
value of the long-run interest rate is equally distributed between
values above and below the 3 percent consumption rate of interest, the
probability-weighted (or ``expected'') present value of a unit
reduction in climate damages will be higher than the value calculated
using a 3 percent discount rate. The effect of such uncertainty about
the correct discount rate can be accounted for by using a lower
``certainty-equivalent'' rate to discount distant future damages,
defined as the rate that produces the expected present value of a
reduction in future damages implied by the distribution of possible
discount rates around what is believed to be the most likely single
value.
---------------------------------------------------------------------------
\678\ Ibid.
---------------------------------------------------------------------------
The IWG identifies ``a plausible range of certainty-equivalent
constant consumption discount rates: 2.5, 3, and 5 percent per year,''
each intended to reflect the effect of uncertainty surrounding
alternative estimates of the correct discount rate. The IWG's
justification for its selection of these rates is summarized in this
excerpt from its 2021 guidance:
The 3 percent value was included as consistent with estimates
provided in OMB's Circular A-4 (OMB 2003) guidance for the
consumption rate of interest. . . . The upper value of 5 percent was
included to represent the possibility that climate-related damages
are positively correlated with market returns, which would imply a
certainty equivalent value higher than the consumption rate of
interest. The low value, 2.5 percent, was included to incorporate
the concern that interest rates are highly uncertain over time. It
represents the average certainty-equivalent rate using the mean-
reverting and random walk approaches from Newell and Pizer (2003)
starting at a discount rate of 3 percent. Using this approach, the
certainty equivalent is about 2.2 percent using the random walk
model and 2.8 percent using the mean reverting approach. Without
giving preference to a particular model, the average of the two
rates is 2.5 percent. Additionally, a rate below the consumption
rate of interest would also be justified if the return to
investments in climate mitigation are negatively correlated with the
overall market rate of return. Use of this lower value was also
deemed responsive to certain judgments based on the prescriptive or
normative approach for selecting a discount rate and to related
ethical objections that have been raised about rates of 3 percent or
higher.
Because the certainty-equivalent discount rate will lie
progressively farther below the best estimate of the current rate as
the time horizon when future impacts occur is extended, the IWG's
recent guidance also suggests that it may be appropriate to use a
discount rate that declines over time to account for interest rate
uncertainty, as has been recommended by NAS and EPA's Science Advisory
Board.\679\ The IWG noted that it will consider these recommendations
and the relevant academic literature on declining rates in developing
its final guidance on the social cost of greenhouse gases.
---------------------------------------------------------------------------
\679\ Interagency Working Group on Social Cost of Greenhouse
Gases, United States Government, Technical Support Document: Social
Cost of Carbon, Methane, and Nitrous Oxide, Interim Estimates under
Executive Order 13990, February 2021.
---------------------------------------------------------------------------
The IWG 2021 interim guidance also presented new evidence on the
consumption-based discount rate suggesting that a rate lower than 3
percent may be appropriate. For example, the IWG replicated OMB
Circular A-4's original 2003 methodology for estimating the consumption
rate using the average return on 10-year Treasury notes over the last
30 years and found a discount rate close to 2 percent. They also
presented rates over a longer time horizon, finding an average rate of
2.3 percent from 1962 to the present. Finally, they summarized results
from surveys of experts on the topic and found a ``surprising degree of
consensus'' for using a 2 percent consumption rate of interest to
discount future climate-related impacts.\680\
---------------------------------------------------------------------------
\680\ Ibid.
---------------------------------------------------------------------------
NHTSA notes that the primary analysis of the NPRM estimated
benefits from reducing emissions of CO2 and other GHGs using
per-ton values of reducing their emissions that incorporated a 2.5
percent discount rate for distant future climate damages, while it
discounted costs and non-climate related benefits using a 3 percent
rate. NHTSA also presented cost and benefits estimates in the primary
analysis that reflected unit values of reducing GHG emissions
constructed using a 3 percent discount rate for reductions in climate-
related damages, while discounting costs and non-climate related
benefits at 7 percent. NHTSA believed at the time this approach
represented an appropriate treatment of the intergenerational issues
presented by emissions that result in climate-related damages over a
very-long time horizon, and was within scope of the IWG's Technical
Support Document: Social Cost of Carbon, Methane, and Nitrous Oxide
that recommends discounting future climate damages at rates of 2.5, 3,
and 5 percent.\681\
---------------------------------------------------------------------------
\681\ Interagency Working Group on Social Cost of Greenhouse
Gases, United States Government, Technical Support Document: Social
Cost of Carbon, Methane, and Nitrous Oxide, Interim Estimates under
Executive Order 13990, February 2021.
---------------------------------------------------------------------------
In addition, NHTSA emphasized the importance and value of
considering the benefits calculated using all four SC-GHG estimates for
each of three greenhouse gases. NHTSA included the social costs of
CO2, CH4, and N2O calculated using the
four different estimates recommended in the February 2021 TSD (model
average at 2.5 percent, 3 percent, and 5 percent discount rates; 95th
percentile at 3 percent discount rate) in the FRIA.
The IWG TSD does not address the question of how agencies should
combine its estimates of benefits from reducing GHG emissions that
reflect these alternative discount rates with the discount rates for
nearer-term benefits and costs prescribed in OMB Circular A-4. However,
the February 2021 TSD identifies 2.5 percent as the ``average
certainty-equivalent rate using the mean-reverting and random walk
approaches from Newell and Pizer (2003) starting at a discount rate of
3 percent.'' \682\ As such, NHTSA believed using a 2.5 percent discount
rate for climate-related damages was consistent with the IWG TSD.
---------------------------------------------------------------------------
\682\ Ibid.
---------------------------------------------------------------------------
As indicated above, NHTSA's PRIA presented cost and benefit
estimates using a 2.5 percent discount rate for reductions in climate-
related damages and 3 percent for non-climate related impacts. NHTSA
also presented cost and benefits estimates using a 3 percent discount
rate for reductions in climate-related damages alongside estimates of
non-climate related impacts discounted at 7 percent. This latter
pairing of a 3 percent rate for discounting benefits from reducing
climate-related damages with a 7 percent discount rate for non-climate
related impacts is consistent with NHTSA's past practice.\683\ However,
NHTSA's pairing in the PRIA of 2.5 percent for climate-related damage
reductions with 3 percent for non-climate related impacts was novel.
---------------------------------------------------------------------------
\683\ See, e.g., the 2012 and 2020 final CAFE rules.
---------------------------------------------------------------------------
[[Page 25880]]
In this final rule, NHTSA has not selected a primary discount rate
for the social cost of greenhouse gases and instead presents non-GHG
related impacts of the final rule discounted at 3 and 7 percent
alongside estimates of the social cost of greenhouse gases reflecting
each of the three discount rates presented by the IWG. This approach
was selected because, as NHTSA pointed out in the NPRM, the IWG does
not specify which of the discount rates it recommends should be
considered the agency's primary estimate. The agency's analysis showing
our primary non-GHG impacts at 3 and 7 percent alongside climate-
related benefits discounted at each rate recommended by the IWG may be
found in FRIA Chapter 6.5.6. For the sake of simplicity, most tables
throughout this analysis pair both the 3 percent and the 7 percent
discount rates with the social costs of greenhouse gases discounted at
a 3 percent rate. To calculate the present value of climate benefits,
we also use the same discount rate as the rate used to discount the
value of damages from future GHG emissions, for internal
consistency.\684\ We believe that this approach provides policymakers
with a range of costs and benefits associated with the rule using a
reasonable range of discounting approaches and associated climate
benefits, as well as the 95th percentile value that illustrates the
potential for climate change to cause damages that are much higher than
the ``best guess'' damage estimates. This approach is also consistent
with the options outlined by NAS's 2017 recommendations on how SC-GHG
estimates can ``be combined in RIAs with other cost and benefits
estimates that may use different discount rates.'' NAS reviewed
``several options,'' including ``presenting all discount rate
combinations of other costs and benefits with [SC-GHG] estimates.''
---------------------------------------------------------------------------
\684\ This approach follows the same approach that the IWG's
February 2021 TSD recommended ``to ensure internal consistency--
i.e., future damages from climate change using the SC-GHG at 2.5
percent should be discounted to the base year of the analysis using
the same 2.5 percent rate.''
---------------------------------------------------------------------------
(c) Comments and Responses About the Agency's Choice of Social Cost of
Carbon Estimates and Discount Rates
California Attorney General et al. commented that the 3 percent
discount rate was too high, referencing the discussion in the IWG's
interim guidance showing rates on 10-year Treasury notes hovering
around 2 percent over the last 30 years. Our Children's Trust commented
that the use of any discount rate on reductions in future climate
damages is unconstitutional because it treats them ``as less valuable
or not equal under the eyes of the law when it comes to life, liberty,
personal security and a climate system that sustains human life, among
other unalienable rights.'' AFPM argued that we should discount the
benefit of reduced climate-related costs at the same rate as is used to
discount other costs and benefits.\685\
---------------------------------------------------------------------------
\685\ AFPM, NHTSA-2021-0053-1530, at 19-21.
---------------------------------------------------------------------------
As noted above, NHTSA presented and considered a range of discount
rates, including 2.5 percent and 5 percent. The above discussion also
explained why it is important to adjust the discounting approach in the
context of intergenerational effects and uncertainty about long-run
interest rates. NHTSA disagrees, however, with the argument that the
use of discounting where there are intergenerational effects is a
violation of the Constitution. The impacts on future generations are
reflected in the estimates used in this analysis.
IPI et al. commented in general support of the agency's approach to
estimating SC-GHG. They argued that the agency should acknowledge that
the IWG's estimates are appropriate but may underestimate the effects
of climate change,\686\ and that the transparent and rigorous
methodology employed by IWG was based on the available science which
adds credibility to their estimates.\687\ Their comment continued by
arguing that the agency should continue to use a global estimate of
SCC-GHG because doing so is supported by science and a domestic
estimate would understate U.S. extraterritorial interests, damages such
as security threats and transboundary damages that spillover into the
U.S., and harm U.S. citizens and assets that are extraterritorial.\688\
Finally, IPI et al. commented that the agency's approach to discounting
climate-related benefits was appropriate, but argued that the agency
should consider aligning with EPA's methodology of reporting climate
benefits at 3 percent for the majority of the tables and include a
sensitivity analysis at a 2 percent discount rate.\689\ Many of the
points raised by IPI et al. are aligned with the agency's approach in
both the proposal and final rule.
---------------------------------------------------------------------------
\686\ IPI et al., Docket No. NHTSA-2021-0053-1547, at 4-7.
\687\ Id. at 31-41.
\688\ Id. at 7-14.
\689\ Id. at 14-31.
---------------------------------------------------------------------------
Competitive Enterprise Institute recommended against the agency's
use of the Interagency Working Group's Interim Estimates of the social
cost of carbon. CEI argued that the degree of global warming mitigation
projected by NHTSA is too small to generate climate benefits valued at
the scale valued by NHTSA using the IWG Interim Estimates. CEI also
argued that the 7 percent discount rate is the appropriate discount
rate for climate damage reduction benefits and that using a lower rate
would justify mitigation projects with a lower rate of return than
could be found in private markets. CEI's rationale was that investing
in higher rate of return projects today would pass along more wealth to
future generations, making them better able to overcome the adversity
posed by potential climate change. They argued that the SC-GHG is
highly sensitive to the time horizon of the analysis and that the SC-
GHG drops significantly if the time horizon for estimating climate
damages is shortened from 300 years to 150 years, and suggested that
the outer years of the 300-year time-horizon were speculative. CEI also
argued that the IWG uses an outdated equilibrium climate sensitivity
distribution and that more recent studies present distributions with
lower modal and central values. They argued that CO2
emissions have important benefits to agriculture and plant growth
through carbon fertilization, which increases internal plant water use
efficiency. Finally, they argued that the IWG's assumptions regarding
human adaptation mitigating the costs of climate change and projected
baseline carbon emissions were unduly pessimistic.
Estimating the social costs of future climate damages caused by
emissions of greenhouse gases, or SC-GHG, requires analysts to make a
number of projections that necessarily involve uncertainty--for
example, about the likely future pattern of global emissions of GHGs--
and to model multifaceted scientific phenomena, including the effect of
cumulative emissions and atmospheric concentrations of GHGs on climate
measures including global surface temperatures and precipitation
patterns. Each of these entail critical judgements about complex
scientific and modeling questions. Doing so requires specialized
technical expertise, accumulated experience, and expert judgment, and
highly trained, experienced, and informed analysts can reasonably
differ in their judgements.
CEI's comments raise questions about the IWG's selection of the
specific assumptions and parameter values it used to produce the
estimates of the social costs of various GHGs that NHTSA relies on in
this regulatory analysis, and contends that using alternative
assumptions and values would reduce the IWG's recommended
[[Page 25881]]
values significantly. However, the agency notes that the IWG's
membership includes experts in climate science, estimation of climate-
related damages, and economic valuation of those impacts, and that
these members applied their collective expertise to review and evaluate
available empirical evidence and alternative projections of important
measures affecting the magnitude and cost of such damages. The agency
also notes that the IWG members employed a collaborative, consensus-
based process to arrive at their collective judgements about the most
reliable assumptions and parameter values. In addition, the IWG uses
its consensus assumptions and estimates in conjunction with three
different widely recognized, peer-reviewed models of climate economic
impacts, and its recommended values represent a synthesis of the costs
each one estimates on the basis of that common set of inputs. Finally,
DOT uses its own judgment in applying the estimates in this analysis.
Thus, the agency believes that the SC-GHG estimates developed by
the IWG have two important advantages over other available estimates:
First, they are the product of consensus estimates of the critical
inputs necessary to estimate damage costs for GHGs; and second, they
synthesize the results of multiple estimation methods represented in
different widely regarded models. As a consequence, NHTSA views the
IWG's recommended values as the most reliable among those that were
available for it to use in its analysis. While the agency acknowledges
that--as CEI notes--selecting certain input assumptions and parameter
estimates different from those the IWG chose could reduce those values,
it also agrees with the IWG that equally and perhaps more plausible
assumptions and parameter values would have resulted in estimated SC-
GHG values that were far higher than those the group ultimately
recommended. Furthermore, due to omitted damage categories, NHTSA
concurs with the IWG that its estimates are likely conservative
underestimates. Unlike the IWG's work, we feel that CEI, Children's
Trust, and the other commenters did not address the inherent
uncertainty in estimating the SC-GHG. Specifically, we note that any
alternative model that attempts to project the costs of GHGs over the
coming decades--and centuries--will be subject to the same uncertainty
and criticisms raised by commenters. Commenters essentially ask NHTSA
to replace this working group's expertise in favor of specific
alternative perspectives presented outside of the full context of the
IWG's significantly technical and multifaceted assessments.
Furthermore, these alternative estimates are reliant on the commenters'
specific set of assumptions of the future being correct.\690\ The IWG's
analysis considered the possibility that its assumptions were either
too conservative or extreme, and based its guidance on a robust review
of potential outcomes.
---------------------------------------------------------------------------
\690\ For example, CEI argued that the IWG estimates ``err[ed]
on the side of alarm and regulatory ambition.'' However, if CEI is
being overly optimistic about how mankind can deal with a changing
climate or the possibility that carbon may have some benefits on
agriculture, IWG's estimate could be an accurate--or even
underestimate--of the SC-GHG.
---------------------------------------------------------------------------
CEI commented that the probability distribution function the IWG
uses to simulate the equilibrium climate sensitivity is outdated and
that more recent empirical work suggests the distribution should have a
lower central tendency. However, CEI's comment overlooked the seminal
work published in 2021 by the Intergovernmental Panel on Climate Change
(IPCC)--an organization of expert scientists with 195 members chartered
by the United Nation and the World Meteorological Organization that
reviews the scientific work of thousands of contributors all over the
world and provides a comprehensive summary ``about what is known about
the drivers of climate change, its impacts and future risks, and how
adaptation and mitigation can reduce those risks.'' \691\ This work was
subjected to a transparent review by experts and governments all around
the world to ``ensure an objective and complete assessment and to
reflect a diverse range of views and expertise.'' \692\ The IPCC's most
recent report states that ``[i]mproved knowledge of climate processes,
paleoclimate evidence and the response of the climate system to
increasing radiative forcing gives a best estimate of equilibrium
climate sensitivity of 3 degrees Celsius.'' \693\ This is the same
value the IWG's probability distribution function uses as the median
estimate of equilibrium climate sensitivity. While the IWG may choose
to revisit the distribution it uses for simulating the equilibrium
climate sensitivity in a future forthcoming update, it is clear that
the distribution used for the interim values is reasonable and
scientifically defensible.
---------------------------------------------------------------------------
\691\ Intergovernmental Panel on Climate Change website, https://www.ipcc.ch/about/.
\692\ Ibid.
\693\ IPCC, 2021: Summary for Policymakers. In: Climate Change
2021: The Physical Science Basis. Contribution of Working Group I to
the Sixth Assessment Report of the Intergovernmental Panel on
Climate Change [Masson-Delmotte, V., P. Zhai, A. Pirani, S.L.
Connors, C. P[eacute]an, S. Berger, N. Caud, Y. Chen, L. Goldfarb,
M. I. Gomis, M. Huang, K. Leitzell, E. Lonnoy, J.B.R. Matthews, T.K.
Maycock, T. Waterfield, O. Yelek[ccedil]i, R. Yu and B. Zhou
(eds.)]. Cambridge University Press. In Press. SPM-13.
---------------------------------------------------------------------------
CEI also commented that we should use an SC-GHG in our main
analysis that only reflects damages to the United States. As an initial
matter, such an estimate would undermine the many rationales for a
global estimate articulated by the IWG, which emphasizes the value of a
global analysis to sufficiently and comprehensively estimate climate
damages. NHTSA believes that continued reliance on the IWG's
recommendations in this respect remains appropriate for all of the
reasons outlined above, which underscore the reasonableness of the
IWG's consensus-based approach.
However, even beyond the recommendations of the IWG, NHTSA agrees
with the IWG that climate change is a global problem and that the
global SC-GHG values are appropriate for this analysis. Emitting
greenhouse gases creates a global externality, in that GHG emitted in
one country mix uniformly with other gases in the atmosphere and the
consequences of the resulting increased concentration of GHG are felt
all over the world.
The effects of climate change are global and affect the United
States through many different pathways. These include through
destabilization that affects our national security, economic impacts
due to interlinked global economies, in danger and risk to U.S.
military assets abroad, harm to soldiers stationed outside the United
States, increased migration to the United States due to climate events
like drought, the provision of disaster aid in response to disasters
caused by climate change, interruptions to supply chains from extreme
weather events, and in many other ways. Given methodological
challenges, it has not yet been possible to derive a robust social cost
estimate that isolates impacts to the United States and its inhabitants
and, in many respects, such an estimate represents an artificial
distinction that fails to account for the comingling of interests
throughout the world. The models used both for the Interim Estimates
and for the 2020 rule's SC-GHG value do not organize all of the
relevant economic and welfare impacts by country, and as such, it is
not possible to develop robust estimates of U.S.-specific damages. As
the Government Accountability Office concluded in a June 2020 report
examining the SC-GHG values used in the 2020 rule, the models ``were
not
[[Page 25882]]
premised or calibrated to provide estimates of the social cost of
carbon based on domestic damages.'' \694\ Further, the report noted
that NAS found that country-specific social costs of carbon estimates
were ``limited by existing methodologies, which focus primarily on
global estimates and do not model all relevant interactions among
regions.'' \695\ It is also important to note that the 2020 rule's SC-
GHG values were never peer reviewed, and when their use in a specific
regulatory action was challenged, a Federal court determined that use
of a U.S.-only value had been ``soundly rejected by economists as
improper and unsupported by science,'' and that the values themselves
omitted key U.S.-specific damages including to supply chains, U.S.
assets and companies, and geopolitical security. California v.
Bernhardt, 472 F.Supp.3d 573, 613-14 (N.D. Cal. 2020).
---------------------------------------------------------------------------
\694\ GAO, Social Cost of Carbon: Identifying a Federal Entity
to Address the National Academies' Recommendations Could Strengthen
Regulatory Analysis, GAO-20-254 (June 2020) at 29.
\695\ Id. at 26.
---------------------------------------------------------------------------
Furthermore, the United States cannot address the domestic
consequences of climate change for the United States by itself.
Instead, we need other nations to take action to reduce their own
domestic emissions and to consider the benefits of their emission
reductions to the United States. In order to ensure other nations take
similar actions to reduce GHG emissions, the United States is actively
involved in developing and implementing international commitments to
secure reductions in GHG emissions. If the United States fails to
consider the benefits (and harms) of its actions to other countries,
our bargaining position is significantly weakened. It is hard to argue
that a large emitter like China, for example, should consider the
global consequences of its actions--including to the United States--if
the United States fails to do so. As a result, the United States may
fail to secure sufficient emission reduction commitments from its
counterpart s to reduce adverse consequences from climate change that
will affect the United States if it were to use U.S.-specific values
for the SC-GHG. A wide range of scientific and economic experts have
emphasized the issue of reciprocity as support for considering global
damages of GHG emissions. Using a global estimate of damages in U.S.
analyses of regulatory actions allows the United States to continue to
actively encourage other nations, including emerging major economies,
to take significant steps to reduce emissions. The only way to achieve
an efficient allocation of resources for emissions reduction on a
global basis--and so benefit the United States and its citizens--is for
all countries to base their policies on global estimates of damages.
Further, in practice, data and modeling limitations naturally
restrain the ability of estimates of SC-GHG to include all of the
important physical, ecological, and economic impacts of climate change,
such that the estimates are a partial accounting of climate change
impacts and will therefore tend to be underestimates of the marginal
benefits of abatement. As an empirical matter, the development of a
U.S.-specific SC-GHG is greatly complicated by the relatively few
region- or country-specific estimates of the SC-GHG in the literature.
Importantly, due to methodological constraints, NHTSA is not aware
of a robust analysis that isolates damages to the United States. Due to
these constraints, the SC-GHG value used in the 2020 final rule is an
underestimate of damages to the United States, and as such is
inappropriately low for purposes of informing the current analysis.
However, NHTSA explored an analysis incorporating a U.S.-specific
social cost of carbon as promoted by commenters such as CEI in order to
comply with a preliminary injunction issued by the United States
District Court for the Western District of Louisiana on February 11,
2022, that enjoined NHTSA from, among other activities, ``[a]dopting,
employing, treating as binding, or relying upon any [SC-GHG] estimates
based on global effects,'' as well as from ``adopting, employing,
treating as binding, or relying upon the work product of the [IWG].''
\696\ When NHTSA considered that analysis, the agency determined that
the selected standards continue to remain maximum feasible.
---------------------------------------------------------------------------
\696\ Louisiana v. Biden, Order, No. 2:21-CV-01074, ECF No. 99
(W.D. La. Feb. 11, 2022). That injunction was subsequently stayed.
Louisiana v. Biden, Order, No. 22-30087, Doc. No. 00516242341 (5th
Cir. Mar. 16, 2022).
---------------------------------------------------------------------------
Even with the underestimate of climate benefits, the analysis still
contained numerous quantitative indicia that the new standards remained
appropriate. For instance, fuel savings for the preferred alternative
still exceeded the price increase due to the rule by $290.\697\
Likewise, a calendar year accounting using a 3 percent discount rate
still revealed a net benefit for the preferred alternative of $28.1
billion. Moreover, these figures--like any cost-benefit analysis
results in a CAFE rulemaking--offered only one informative data point
in addition to the host of considerations that NHTSA must balance by
statute when determining maximum feasible standards. Even taking the
severely reduced climate benefit estimates into account, the overall
balance of other significant qualitative and quantitative
considerations and factors support the selection of the preferred
alternative, as described at length throughout this final rule.
Accordingly, even the limited perspective of impacts urged by
commenters such as CEI would not alter the standards necessitated in
this rulemaking.
---------------------------------------------------------------------------
\697\ This final rule is estimated to increase the price of
model year 2029 vehicles by $1,087 and save consumer $1,387.
---------------------------------------------------------------------------
NHTSA believes that the three issues raised by CEI and specifically
addressed in this section on the IWG's interim values--regarding the
use of opportunity cost of capital discounting, the use of global
values for the social costs of greenhouse gases, and the probability
distribution function of equilibrium climate sensitivity--are
representative of their comments overall in that they choose to
highlight areas of uncertainty and dynamics that would tend to reduce
the social cost of carbon. In each case, CEI has chosen to ignore
sources of uncertainty and dynamics that may increase the social cost
of carbon and asserts scientific authority only where the cited papers
or dynamics would tend to reduce it.
Contrary to the position put forward by Children's Trust that it is
unlawful to discount the estimated costs of SC-GHG, we also believe
that discounting the SC-GHG estimate to develop a present value of the
benefits of reducing GHG emissions is consistent with the law, and that
the discounting approach used by the IWG is reasonable. Unsurprisingly,
when the cost-benefit analysis is the predominant basis for an agency's
decision, courts have previously reviewed and affirmed rules that
discount climate-related costs.\698\ Courts have likewise advised
agencies to approach cost-benefit analyses with impartiality, to ensure
that important factors are captured in the analysis, including climate
benefits,\699\ and to ensure that the decision rests ``on a
consideration of the relevant factors.'' \700\ NHTSA has followed these
principles here.
---------------------------------------------------------------------------
\698\ See, e.g., E.P.A. v. EME Homer City Generation, L.P., 572
U.S. 489 (2015).
\699\ CBD v. NHTSA, 538 F.3d 1172, 1197 (9th Cir. 2008).
\700\ State Farm, 463 U.S. 29, 43 (1983) (internal quotation
marks omitted).
---------------------------------------------------------------------------
For these reasons, NHTSA believes that the Interim Estimates
employed in
[[Page 25883]]
this analysis and the results they produce are the most reliable
estimates of what are inherently uncertain values it could have
selected, and that the range of values used to examine the sensitivity
of its results adequately incorporate the range of uncertainty
surrounding the values used in its central analysis.
(2) Reduced Health Damages
The CAFE Model estimates monetized health effects associated with
emissions from three criteria pollutants: NOX,
SOX, and PM2.5. As discussed in Section III.F
above, although other criteria pollutants are currently regulated, we
only calculate impacts from these three pollutants since they are known
to be emitted regularly from mobile sources, have the most adverse
effects to human health, and are based on EPA papers that estimate the
benefits per ton of reducing these pollutants.
CBD et al. stated that NHTSA improved the monetization of
PM2.5 attributable to fuel economy standards (discussed
further below); however, the commenters also argued that NHTSA should
monetize benefits from reductions in ozone and air toxics.\701\
---------------------------------------------------------------------------
\701\ CBD et al., Docket No. NHTSA-2021-0053-1572, at 5.
---------------------------------------------------------------------------
As we discussed in the proposal, other pollutants, especially those
that are precursors to ozone, are difficult to model due to the
complexity of their formation in the atmosphere, and EPA does not
calculate benefit-per-ton estimates for these. We will continue to
explore this concept for future analyses. Chapter 5.4 of the TSD
includes a section on uncertainty related to monetizing health impacts.
The Final SEIS also includes a section describing the health effects of
ozone and air toxics (see Section 4.1.1.2).
The CAFE Model computes the monetized impacts associated with
health damages from each pollutant by multiplying monetized health
impact per ton values by the total tons of these pollutants, which are
emitted from both upstream and tailpipe sources. Chapter 5 of the TSD
accompanying this final rule includes a detailed description of the
emission factors that inform the CAFE Model's calculation of the total
tons of each pollutant associated with upstream and tailpipe emissions.
These monetized health impacts per ton values are closely related
to the health incidence per ton values described above in Section III.F
and in detail in Chapter 5.4 of the TSD. We use the same EPA sources
that provide health incidence values to determine which monetized
health impacts per ton values to use as inputs in the CAFE Model. Like
the estimates associated with health incidences per ton of criteria
pollutant emissions, we use multiple EPA papers and conversations with
EPA staff to appropriately account for monetized damages for each
pollutant associated with the source sectors included in the CAFE
Model, based on which papers contain the most up-to-date data
corresponding to the relevant source sectors.\702\ The various emission
source sectors included in the EPA papers do not always correspond
exactly to the emission source categories used in the CAFE Model.\703\
In those cases, we map multiple EPA sectors to a single CAFE source
category and compute a weighted average of the health impact per ton
values.
---------------------------------------------------------------------------
\702\ Environmental Protection Agency (EPA). 2018. Estimating
the Benefit per Ton of Reducing PM2.5 Precursors from 17
Sectors. https://www.epa.gov/sites/production/files/2018-02/documents/sourceapportionmentbpttsd_2018.pdf; Wolfe et al. 2019.
Monetized health benefits attributable to mobile source emissions
reductions across the United States in 2025. https://pubmed.ncbi.nlm.nih.gov/30296769/; Fann et al. 2018. Assessing Human
Health PM2.5 and Ozone Impacts from U.S. Oil and Natural
Gas Sector Emissions in 2025. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6718951/.
\703\ The CAFE Model's emission source sectors follow a similar
structure to the inputs from GREET. See Chapter 5.2 of the TSD
accompanying this notice for further information.
---------------------------------------------------------------------------
CBD et al. stated that the estimates of the benefits of
PM2.5 reductions were improved by the addition of the Wolfe
et al. paper to the sources used by NHTSA.\704\ We agree, and continue
to use these sources in the final rulemaking analysis as they allow us
to map sectors to categories that are more expansive and specific than
the original 2018 source.
---------------------------------------------------------------------------
\704\ CBD et al., at 5.
---------------------------------------------------------------------------
EPA uses the value of a statistical life (VSL) to estimate
premature mortality impacts, and a combination of willingness to pay
estimates and costs of treating the health impact for estimating the
morbidity impacts.\705\ EPA's 2018 TSD, ``Estimating the Benefit per
Ton of Reducing PM2.5 Precursors from 17 Sectors,'' \706\
(referred to here as the 2018 EPA source apportionment TSD) contains a
more detailed account of how health incidences are monetized. It is
important to note that the EPA sources cited frequently refer to these
monetized health impacts per ton as ``benefits per ton,'' since they
describe these estimates in terms of emissions avoided. In the CAFE
Model input structure, these are generally referred to as monetized
health impacts or damage costs associated with pollutants emitted, not
avoided, unless the context states otherwise.
---------------------------------------------------------------------------
\705\ Although EPA and DOT's VSL values differ, DOT staff
determined that using EPA's VSL was appropriate here, since it was
already included in these monetized health impact values, which were
best suited for the purposes of the CAFE Model.
\706\ See Environmental Protection Agency (EPA). 2018.
Estimating the Benefit per Ton of Reducing PM2.5
Precursors from 17 Sectors. https://www.epa.gov/sites/production/files/2018-02/documents/sourceapportionmentbpttsd_2018.pdf.
---------------------------------------------------------------------------
The Competitive Enterprise Institute questioned the use of the
specific EPA studies that inform the BPT values that NHTSA uses, namely
the Six Cities Study.\707\ We report only one BPT estimate in this
final rule, based on the Krewski et al. study, to be consistent with
EPA in their final GHG rule. We consulted with EPA staff at the Office
of Air Quality Planning and Standards (OAQPS) on the most appropriate
benefit per ton values to use for the various upstream and tailpipe
emission categories. EPA bases its benefits analyses on peer-reviewed
studies of air quality and health effects and peer-reviewed studies of
the monetary values of public health and welfare improvements. Very
recently, EPA updated its approach to estimating the benefits of
changes in PM2.5 and ozone.708 709 These updates
were based on information drawn from the recent 2019 PM2.5
and 2020 Ozone Integrated Science Assessments (ISAs), which were
reviewed by the Clean Air Science Advisory Committee (CASAC) and the
public.710 711 EPA has not updated its mobile source BPT
estimates to reflect these updates in time for this analysis. Instead,
based on the recommendation of EPA staff, we use the same
PM2.5 BPT estimates that we used in the NPRM to ensure
consistency between the values corresponding to different source
sectors. The BPT estimates used are based on the review of the 2009 PM
ISA \712\ and 2012 PM ISA Provisional
[[Page 25884]]
Assessment \713\ and include a mortality risk estimate derived from the
Krewski et al. (2009) \714\ analysis of the American Cancer Society
(ACS) cohort and nonfatal illnesses consistent with benefits analyses
performed for the analysis of the final Tier 3 Vehicle Rule,\715\ the
final 2012 PM NAAQS Revision,\716\ and the final 2017-2025 Light-duty
Vehicle GHG Rule.\717\ We expect this lag in updating our BPT estimates
to have only a minimal impact on total PM benefits, since the
underlying mortality risk estimate based on the Krewski study is
identical to an updated PM2.5 morality risk estimate derived
from an expanded analysis of the same ACS cohort. We are aware of EPA's
work to update its mobile source BPT estimates to reflect these recent
updates for use in future rulemaking analyses, and will work further
with EPA in future rulemakings to update and synchronize approaches to
BPT estimates.
---------------------------------------------------------------------------
\707\ Competitive Enterprise Institute, Docket No. NHTSA-2021-
0053-1546, at 3.
\708\ U.S. Environmental Protection Agency (U.S. EPA). 2021a.
Regulatory Impact Analysis for the Final Revised Cross-State Air
Pollution Rule (CSAPR) Update for the 2008 Ozone NAAQS. EPA-452/R-
21-002. March.
\709\ U.S. Environmental Protection Agency (U.S. EPA). 2021b.
Estimating PM2.5- and Ozone-Attributable Health Benefits.
Technical Support Document (TSD) for the Final Revised Cross-State
Air Pollution Rule Update for the 2008 Ozone Season NAAQS. EPA-HQ-
OAR-2020-0272. March.
\710\ U.S. Environmental Protection Agency (U.S. EPA). 2019a.
Integrated Science Assessment (ISA) for Particulate Matter (Final
Report, 2019). U.S. Environmental Protection Agency, Washington, DC,
EPA/600/R-19/188, 2019.
\711\ U.S. Environmental Protection Agency (U.S. EPA). 2019a.
Integrated Science Assessment (ISA) for Ozone and Related
Photochemical Oxidants (Final Report). U.S. Environmental Protection
Agency, Washington, DC, EPA/600/R-20/012, 2020.
\712\ U.S. Environmental Protection Agency (U.S. EPA). 2009.
Integrated Science Assessment for Particulate Matter (Final Report).
EPA-600-R-08-139F. National Center for Environmental Assessment--RTP
Division, Research Triangle Park, NC. December. Available at: http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=216546.
\713\ U.S. Environmental Protection Agency (U.S. EPA). 2012.
Provisional Assessment of Recent Studies on Health Effect of
Particulate Matter Exposure. EPA/600/R-12/056F. National Center for
Environmental Assessment--RTP Division, Research Triangle Park, NC.
December. Available at: https://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=247132.
\714\ Krewski D., M. Jerrett, R.T. Burnett, R. Ma, E. Hughes, Y.
Shi, et al. 2009. Extended Follow-Up and Spatial Analysis of the
American Cancer Society Study Linking Particulate Air Pollution and
Mortality. HEI Research Report, 140, Health Effects Institute,
Boston, MA.
\715\ U.S. Environmental Protection Agency (2014). Control of
Air Pollution from Motor Vehicles: Tier 3 Motor Vehicle Emission and
Fuel Standards Final Rule: Regulatory Impact Analysis, Assessment
and Standards Division, Office of Transportation and Air Quality,
EPA-420-R-14-005, March 2014. Available on the internet: http://www3.epa.gov/otaq/documents/tier3/420r14005.pdf.
\716\ U.S. Environmental Protection Agency. (2012). Regulatory
Impact Analysis for the Final Revisions to the National Ambient Air
Quality Standards for Particulate Matter, Health and Environmental
Impacts Division, Office of Air Quality Planning and Standards, EPA-
452-R-12-005, December 2012. Available on the internet: http://www3.epa.gov/ttnecas1/regdata/RIAs/finalria.pdf.
\717\ U.S. Environmental Protection Agency (U.S. EPA). (2012).
Regulatory Impact Analysis: Final Rulemaking for 2017-2025 Light-
Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average
Fuel Economy.
---------------------------------------------------------------------------
Auto Innovators also suggested additional sensitivity analysis of
BPT inputs, citing the EPA Science Advisory Board's ``recommended
sensitivity analyses of alternative values of the dose-response
function, differential toxicity by type of particle, and spatially-
dependent VSL values.'' \718\ We include other BPT values in one health
effects sensitivity case described in Chapter 7 of the FRIA. Further
sensitivity cases were not deemed necessary for the purposes of this
analysis, since criteria pollutant health impacts make up a very small
portion of overall benefits.
---------------------------------------------------------------------------
\718\ Auto Innovators, Docket No. NHTSA-2021-0053-1492, at 92.
---------------------------------------------------------------------------
Our Children's Trust objected to using discount rates when
monetizing health benefits, stating that ``to apply a discount rate to
monetized health impacts is also completely inappropriate and unlawful
and discriminates against children.'' \719\ The health impacts of
exposure to criteria pollutants occur well after exposure to air
pollution (i.e., the impacts have long ``latency periods''), and
therefore it is appropriate to reflect some difference in timing
(through discounting) in the monetized values.
---------------------------------------------------------------------------
\719\ Our Children's Trust, Docket No. NHTSA-2021-0053-1587, at
3.
---------------------------------------------------------------------------
We disagree with Our Children's Trust's assertion that applying a
discount rate to health benefits is illegal. Our Children's Trust did
not provide any specific laws that we were allegedly violating, nor are
we aware of any such law. Guidance from OMB Circular A-4 recommends
using discount rates of 3 and 7 percent in benefit-cost analyses and
has been used for regulatory analyses for decades, including in the
evaluation of regulations with health impacts similar to those of this
final rule.
However, OMB Circular A-4 also acknowledges the ethical
considerations involved in analyzing impacts occurring over
intergenerational time horizons:
Special ethical considerations arise when comparing benefits and
costs across generations. Although most people demonstrate time
preference in their own consumption behavior, it may not be
appropriate for society to demonstrate a similar preference when
deciding between the well-being of current and future generations.
Future citizens who are affected by such choices cannot take part in
making them, and today's society must act with some consideration of
their interest.\720\
---------------------------------------------------------------------------
\720\ OMB Circular A-4.
Factoring in competing social interests presents additional
difficulties in weighing these ethical considerations. As of this time,
we include health benefits at the 3 percent as well as 7 percent
discount rate and will consider the question of lower discount rates
for health benefits in future analyses.
The CAFE Model health impacts inputs are based partially on the
structure of the 2018 EPA source apportionment TSD, which reports
benefits per ton values for the years 2020, 2025, and 2030. For the
years in between the source years used in the input structure, the CAFE
Model applies values from the closest source year. For instance, the
model applies 2020 monetized health impact per ton values for calendar
years 2020-2022 and applies 2025 values for calendar years 2023-2027.
For some of the monetized health damage values, in order to match the
structure of other impacts costs, we developed proxies for 7 percent
discounted values for specific source sectors by using the ratio
between a comparable sector's 3 and 7 percent discounted values. In
addition, we used implicit price deflators from the Bureau of Economic
Analysis (BEA) to convert different monetized estimates to 2018
dollars, to be consistent with the rest of the CAFE Model inputs.
This process is described in more detail in Chapter 6.2.2 of the
TSD accompanying this final rule. In addition, the CAFE Model
documentation contains more details of the model's computation of
monetized health impacts. All resulting emissions damage costs for
criteria pollutants are located in the Criteria Emissions Cost
worksheet of the Parameters file.
(3) Reduction in Petroleum Market Externalities
By amending existing standards, this action will reduce domestic
consumption of gasoline, producing a corresponding decrease in the
Nation's demand for crude petroleum, a commodity that is traded
actively in a worldwide market. U.S. consumption and imports of
petroleum products have three potential effects on the domestic economy
that are often referred to collectively as ``energy security
externalities.'' Increases in their magnitude are sometimes cited as
possible social costs of increased U.S. demand for petroleum, and
analogously, any reduction in their value in response to lower U.S.
consumption or imports of petroleum represent potential economic
benefits.
First, the U.S. accounts for a sufficiently large (although
declining) share of global petroleum demand such that changes in
domestic consumption of petroleum products can affect global petroleum
prices. Any increase in global petroleum prices that results from
higher U.S. gasoline demand will cause a transfer of revenue from
consumers of petroleum to oil suppliers worldwide, because consumers
throughout the world are ultimately subject to the higher global price
that results. Although this transfer is simply a shift of resources
that produces no change in global economic welfare, the financial
[[Page 25885]]
drain it produces on the U.S. economy is sometimes cited as an external
cost of increased U.S. petroleum consumption because consumers of
petroleum products are unlikely to consider it. Similarly, a decline in
U.S. consumption of petroleum-derived transportation fuel will reduce
global petroleum demand and exert some downward pressure on worldwide
prices. Although the resulting savings to worldwide consumers of
petroleum products is again a transfer--this time from oil producers to
consumers--it may reduce the financial drain on the U.S. economy caused
by domestic oil production and imports.
As the U.S. approaches self-sufficiency in petroleum production
(the Nation became a net exporter of petroleum in 2020), any effect of
reduced domestic demand on global petroleum prices increasingly results
in a transfer from U.S. petroleum producers to domestic consumers of
refined products.\721\ Thus not only does it leave net U.S. welfare
unaffected, it also ceases even to be a financial burden on the U.S.
economy. In fact, as the U.S. becomes a larger net petroleum exporter,
any transfer from global consumers to petroleum producers would become
a financial benefit to the U.S. economy, although uncertainty in the
Nation's long-term import-export balance makes it difficult to project
precisely how these effects might change in response to increased
consumption.
---------------------------------------------------------------------------
\721\ See https://www.eia.gov/energyexplained/oil-and-petroleum-products/imports-and-exports.php (accessed March 17, 2022).
---------------------------------------------------------------------------
Higher U.S. petroleum consumption also increases domestic
consumers' exposure to oil price shocks and by doing so impose
potential costs on all U.S. petroleum users (including those outside
the light duty vehicle sector, whose consumption would be unaffected by
this final rule) from possible interruptions in the global supply of
petroleum or rapid fluctuations in global oil prices. These potential
costs arise from petroleum users' need to pay more for oil-based
products, to switch energy sources, or adjust production methods
rapidly in response to reduced supplies or higher prices, which they
cannot recover once supplies are restored or prices return to pre-
disruption levels, and from losses in economic output while supplies
are disrupted. Because users of petroleum products are unlikely to
consider the effect of their increased purchases on the risk of these
effects, their probability-weighted (or ``expected'') economic value is
often cited as an external cost of increased U.S. consumption of
petroleum products. Conversely, reducing domestic consumption of
refined products reduces exposure to supply disruptions or rapid price
changes and petroleum users' costs for adjusting rapidly to them, which
will reduce the external economic costs associated with domestic
petroleum consumption. When U.S. oil consumption is linked to the
globalized and tightly interconnected oil market, as it is now, the
only means of reducing the exposure of U.S. consumers to global oil
shocks is to reduce their consumption. Thus the reduction in oil
consumption driven by fuel economy standards creates an energy security
benefit.
This benefit is the original purpose behind the CAFE standards. Oil
prices are inherently volatile, in part because geopolitical risk
affects prices. International conflicts, sanctions, civil conflicts
targeting oil production infrastructure, pandemic-related economic
upheaval, and cartels have all had dramatic and sudden effects on oil
prices in recent years. U.S. net exporter status does not insulate U.S.
drivers from higher gas prices, because those prices are currently
largely determined by oil prices set in the globally integrated market.
Given these dynamics, the effective policies to protect consumers from
oil price spikes are those that reduce the oil-intensity of the
economy, including fuel economy standards.
Finally, some analysts argue that domestic demand for imported
petroleum may also influence U.S. military spending. Because the
increased cost of military activities would not be reflected in the
prices drivers pay at the gas pump, increased military spending to
secure oil imports is often represented as a third category of external
costs form increased U.S. petroleum consumption. NHTSA has received
extensive comments to past actions on this topic.
Each of these three factors would be expected to decrease--albeit
by a limited magnitude--as a consequence of decreasing U.S. petroleum
consumption resulting from more stringent CAFE standards. TSD Chapter
6.2.4 provides a comprehensive explanation of the agency's analysis of
these three impacts and explains how it values potential economic
benefits from reducing each one. The agency's proposed rule also
presented this same explanation and drew numerous comments, most
asserting that the value the agency attached to reducing the expected
economic costs of oil supply disruptions and price volatility was too
low.
As one illustration of the comments that the agency received on
this issue, the Applied Economics Clinic (AEC) argued on behalf of the
California Attorney General and the CARB that the expert assessment of
the likelihood of petroleum supply disruptions underlying the agency's
estimate of macroeconomic disruption costs estimated disruption
probabilities that were far too low to be consistent with recent
experience, causing the agency's cost estimate to be unrealistically
low. AEC also noted that NHTSA's estimates were presented as a single
value without acknowledging the range of uncertainty customarily
estimated to surround it, and that other estimates reported in the same
source on which NHTSA relies for its disruption costs are significantly
higher. AEC argued that the agency should return to using the estimates
of disruption probabilities and expected costs from Oak Ridge National
Laboratories (ORNL) that it had relied on in previous analyses, whose
central value it estimated at more than twice the figure the agency
used to analyze its proposed rule. However, the agency notes that both
ORNL's estimates of supply disruption costs and the alternative
estimates presented in the source NHTSA relies on use exactly the same
type of expert elicitation of the probabilities and magnitudes of
disruptions used in the study from which NHTSA's cost estimates were
derived, and also reflect less up to date assumptions about other
factors such as petroleum prices and global petroleum supply
elasticities that affect its cost estimates. For these reasons, the
agency's analysis of this final rule continues to rely on its earlier
estimates.
In addition, AEC argues that net financial transfers between U.S.
suppliers and consumers of petroleum products are unlikely to be zero
in any single year because of year-to-year variation in U.S. gross
imports and exports of petroleum, and that the agency's analysis should
explicitly account for forecast variation in these volumes. The agency
notes that this would force it to rely on inherently uncertain
forecasts of U.S. and global petroleum production and demand, and in
any case, would be unlikely to produce a significantly different
outcome from the analysis presented here because AEC's assumption
depends primarily on the Nation's net imports over the entire period it
spans. Discounting of net transfers projected to occur in distant
future years would also reduce their present values, particularly or
distant future years.
Finally, AEC also argues that even if net dollar-valued revenue
transfers
[[Page 25886]]
between U.S. consumers and suppliers are zero, their net welfare
impacts will not necessarily be neutral and should be accounted for.
The agency notes that while this assertion is correct, accounting for
the true welfare rather than the financial consequences of revenue
transfers would require detailed information on the income
distributions of U.S. consumers of petroleum products and of equity
holders (and other investors) in domestically based oil companies, as
well as estimates of the marginal utility of income and its variation
over the income spectrum. This level of detail is well beyond the scope
of the agency's analyses of other, much more significant economic
impacts of this final rule, and employing it would complicate the
analysis and its interpretation enormously without a commensurate
improvement in its usefulness to decision-makers or the public.
In the proposal, the agency reviewed its previous assumption that
90 percent of any reduction in domestic petroleum refining to produce
gasoline that results from the proposal would reduce U.S. petroleum
imports, with the remaining 10 percent reducing domestic production.
The California Attorney General requested that we revisit this
assumption, asserting that only a small portion of U.S. gasoline demand
is supplied by foreign-refined oils today. The agency neglected to make
this change in the analysis supporting the proposal, and has refrained
from revising the analysis for the final rule. While we believe that
there remains a strong case to assume that any reduction in refining of
crude petroleum to produce gasoline would reduce U.S. oil imports,
rather than changing U.S. petroleum output, we are going to continue to
evaluate assumption given the concerns raised by the California
Attorney General. In the interim, we will continue to assume that 90
percent of any reduction in domestic petroleum refining to produce
gasoline that results from the proposal would reduce U.S. petroleum
imports, with the remaining 10 percent reducing domestic production. We
conducted a sensitivity analysis to scope the difference between the
two assumptions and observed that the difference in estimated total and
net benefits is less than 0.1 percent when viewed from either the model
year or calendar year perspective and discounted at either 3 or 7
percent.\722\
---------------------------------------------------------------------------
\722\ See FRIA Chapter 7.
---------------------------------------------------------------------------
(4) Changes in Labor
As vehicle prices rise, we expect consumers to purchase fewer
vehicles than they would have at lower prices. If manufacturers produce
fewer vehicles as a consequence of lower demand, manufacturers may need
less labor to produce their fleet and dealers may need less labor to
sell the vehicles. Conversely, as manufacturers add equipment to each
new vehicle, the industry will require labor resources to develop,
sell, and produce additional fuel-saving technologies.\723\ We also
account for the possibility that new standards could shift the relative
shares of passenger cars and light trucks in the overall fleet. Since
the production of different vehicles involves different amounts of
labor, this shift impacts the quantity of estimated labor.
---------------------------------------------------------------------------
\723\ For the purposes of this analysis, DOT assumes a linear
relationship between labor and production volumes.
---------------------------------------------------------------------------
The analysis considers the direct labor effects that the CAFE
standards have across the automotive sector. The facets include (1)
dealership labor related to new light-duty vehicle unit sales; (2)
assembly labor for vehicles, engines, and transmissions related to new
vehicle unit sales; and (3) labor related to mandated additional fuel
savings technologies, accounting for new vehicle unit sales. The labor
utilization analysis is intentionally narrow in its focus and does not
represent an attempt to quantify the overall labor or economic effects
of this rulemaking because adjacent employment factors and consumer
spending factors for other goods and services are uncertain and
difficult to predict. We do not consider how direct labor changes may
affect the macro economy and potentially change employment in adjacent
industries. For instance, we do not consider possible labor changes in
vehicle maintenance and repair, nor changes in labor at retail gas
stations. We also do not consider possible labor changes due to raw
material production, such as production of aluminum, steel, copper, and
lithium, nor does the agency consider possible labor impacts due to
changes in production of oil and gas, ethanol, and electricity.
Auto Innovators recommended NHTSA consider the geographic
differences in employment losses and gains in its labor analysis and
present additional results based on such regional differences. Auto
Innovators pointed out that the impacts of BEVs on U.S. employment,
specifically in gasoline engine and transmission plants and supply
chains, as well as in the petroleum and biofuels sector, may differ
based on region. They also noted that the employment impacts of BEV
production elsewhere should be studied.\724\ As discussed above,
NHTSA's labor utilization analysis is intentionally narrow in focus and
all effects are reported at a national level. While we appreciate the
benefits of identifying how employment may shift between geographic
areas as different suites of technologies are employed, identifying
where to deploy resources and trainings within the Nation is outside
the scope of this rulemaking. We may consider expanding the scope of
the labor utilization analysis or reporting subnational results in
future rulemaking analyses.
---------------------------------------------------------------------------
\724\ Auto Innovators, at 122.
---------------------------------------------------------------------------
All labor effects are estimated and reported at a national level,
in person-years, assuming 2,000 hours of labor per person-year.\725\
These labor hours are not converted into monetized values because we
assume that the labor costs are included into a new vehicle's
purchasing price. The analysis estimates labor effects from the
forecasted CAFE Model technology costs and from review of automotive
labor for the MY 2020 fleet. The agency uses information about the
locations of vehicle assembly, engine assembly, and transmission
assembly, and the percent of U.S. content of vehicles collected from
American Automotive Labeling Act (AALA) submissions for each vehicle in
the reference fleet.\726\ The analysis assumes the portion of parts
that are made in the U.S. will remain constant for each vehicle as
manufacturers add fuel-savings technologies. This should not be
misconstrued as a prediction that the percentage of U.S.-made parts--
and by extension U.S. labor--will remain constant, but rather that the
agency does not have a clear basis to project where future productions
may shift. The analysis also uses data from the 2016 National
Automotive Dealers Association (NADA) annual report to derive
dealership labor estimates. We considered using data from NADA's 2020
report but concluded that 2020 was too affected by COVID-19 to be an
appropriate basis to project future dealership labor values.
---------------------------------------------------------------------------
\725\ The agencies recognize a few local production facilities
may contribute meaningfully to local economies, but the analysis
reports only on national effects.
\726\ 49 CFR part 583.
---------------------------------------------------------------------------
In sum, the analysis shows that the increased labor from production
of new technologies used to meet the Preferred Alternative will
outweigh any decreases attributable to the change in new vehicle sales.
For a full description of the process the agency uses to estimate labor
impacts, see TSD Chapter 6.2.5.
[[Page 25887]]
3. Costs and Benefits not Quantified
In addition to the costs and benefits described above, Table III-37
and Table III-38 each include two line-items without values. The first
is maintenance and repair costs. Many of the technologies manufacturers
apply to vehicles to meet CAFE standards are sophisticated and costly.
The technology costs capture only the initial or ``upfront'' costs to
incorporate this equipment into new vehicles; however, if the equipment
is costlier to maintain or repair--which is likely either because the
materials used to produce the equipment are more expensive or the
equipment is significantly more complex than less fuel efficient
alternatives and requires more time and labor--then consumers will also
experience increased costs throughout the lifetime of the vehicle to
keep it operational. The agency does not calculate the additional cost
of repair and maintenance currently because it lacks a basis for
estimating the incremental change attributable to the standards. NHTSA
sought comment on how to estimate these costs from the public but did
not receive any suggestions.
The second item is the potential tradeoff with other vehicle
attributes that could create an opportunity cost for some consumers. In
addition to fuel economy, potential buyers of new cars and light trucks
value other features such as their seating and cargo-carrying capacity,
ride comfort, safety, and performance. Changing some of these other
features, however, can sometimes affect vehicles' fuel economy, so
manufacturers will carefully consider any tradeoffs among them when
deciding how to comply with stricter CAFE standards. Currently the
analysis assumes that these vehicle attributes will not change as a
result of these rules,\727\ but in practice manufacturers may make
practical design changes to meet the standards and minimize their
compliance costs.
---------------------------------------------------------------------------
\727\ See TSD Chapter 2.4.5.
---------------------------------------------------------------------------
If manufacturers do so, they may lower compliance costs relative to
those estimated here,\728\ but the change to other attributes could in
theory involve an opportunity cost to consumers who value specific
attributes, if those consumers cannot purchase a vehicle with those
attributes. Similarly, if manufacturers could use the same technology
to either improve efficiency or improve performance relative to current
attributes, and choose to use the technology only to improve
efficiency, the consumer may not experience the performance
enhancement. Of course, unless automakers reach an absolute technology
limit, which has not been observed, and unless there is a technical or
engineering constraint that makes it impossible or much more expensive
to add additional performance features after increasing fuel economy,
they can still improve other vehicle attributes while improving fuel
economy--as is always the case, those improvements would come at a
cost, but that cost would be borne only by consumers who value the
attribute improvement more than its cost. Because fuel efficiency
improvements can save consumers money on net by reducing fuel
expenditures, assuming consumers are completely financing their vehicle
purchases, the fuel economy improvements can only reduce a consumer's
``budget'' for other vehicle attributes to the extent that the monthly
car payment increases due to the improvements by more than the fuel
savings the technologies deliver.
---------------------------------------------------------------------------
\728\ See Kate S. Whitefoot et al., Compliance by Design:
Influence of Acceleration Trade-Offs on CO2 Emissions and
Costs of Fuel Economy and Greenhouse Gas Regulations, 51 Env't Sci.
& Tech. 10,307 (2018); Gloria Helfand & Reid Dorsey-Palmateer, The
Energy Efficiency Gap in EPA's Benefit-Cost Analysis of Vehicle
Greenhouse Gas Regulations: A Case Study, 6 J. Benefit Cost Analysis
432 (2015).
---------------------------------------------------------------------------
The agency has previously attempted to model the potential
opportunity cost associated with changes in other vehicle attributes in
sensitivity analyses. In those other rulemakings, the agency
acknowledged that it is extremely difficult to quantify the potential
changes to other vehicle attributes. To accurately do so requires
extensive projections about which and how much of other attributes will
be altered and a detailed accounting of how much value consumers
assigned to those attributes. The agency modeled the opportunity cost
associated with changes in other vehicle attributes using published
empirical estimates of tradeoffs between higher fuel economy and
improvements to other attributes, together with estimates of the values
buyers attach to those attributes. The agency does not believe this is
an appropriate methodology since there is considerable uncertainty in
the literature about how much fuel economy consumers are willing to pay
for and how consumers value other vehicle attributes. We note, for
example, a recent EPA-commissioned study that ``found very little
useful consensus'' regarding ``estimates of the values of various
vehicle attributes,'' which ultimately were ``of little use for
informing policy decisions.'' \729\
---------------------------------------------------------------------------
\729\ EPA, Consumer Willingness to Pay for Vehicle Attributes:
What is the Current State of Knowledge? (2018).
---------------------------------------------------------------------------
As noted above, an analysis of opportunity costs optimally would
need to assess compliance with these standards while allowing
manufacturers to adjust vehicle attributes. This requires detailed
information about how much different consumers value various vehicle
attributes, which is not currently available. Such an analysis could
show lower compliance costs for the standards, but could identify any
opportunity costs where consumers value other vehicle attributes that
are not incorporated into the vehicle they purchase.
Still, there is some evidence that consumers are myopic with
respect to future savings well beyond any attribute tradeoff.
Gillingham et al. (2021) use an error in fuel efficiency marketing and
subsequent change in the market equilibrium price for the vehicles in
question to assess the willingness to pay for fuel efficiency and find
that consumers are only willing to pay $0.16-0.39 per discounted value
of a dollar of fuel savings. The intriguing feature of this study is
that it uses identical cars made by Hyundia and Kia, which means the
features of the car with and without the reported fuel savings are
identical and the discount paid for future fuel saving cannot be
attributed to an omitted feature. Therefore, the undervaluation
observed in this study is not due to consumers valuing other vehicle
attributes more than fuel economy. The findings of this paper are
consistent with consumers displaying myopia--a term they use to
``describe a range of behavioral phenomena that could cause
undervaluation.''
In comments to the NPRM, IPI provided extensive comments on this
topic. IPI cited the 2019 EPA Automotive Trends Report as showing that
horsepower and fuel economy have both steadily improved since 2008, and
cited EPA's finding in the Midterm Evaluation that simultaneous
improvements in fuel economy and other vehicle attributes likely
indicates that any historical trade-off between the two is far less
likely to be present in the context of advanced vehicle engines. IPI
also stated that many technologies that improve fuel economy also
improve other vehicle attributes, and those benefits would offset any
opportunity costs. Further, IPI stated that:
Economic research has long recognized the various implicit subsidies
and externalities
[[Page 25888]]
imposed on society by vehicles. These include: Accidents, road
congestion, road and parking construction and maintenance costs, the
space used for parking, and pollution. Drivers with higher
horsepower vehicles are much more likely to speed--by 10 miles per
hour or more--increasing the risk of accidents, damages, and
fatalities. Vehicles with features that allow faster acceleration
also cause a greater number of and more consequential accidents.
Vehicles with internal combustion engines are more dangerous than
those with electric engines due to the latter's additional crumple
space. Heavier vehicles also increase the cost of road maintenance
and repair. Vehicles with greater acceleration also may be driven in
ways that consume more fuel and so emit more pollution. And as
discussed below, certain status features like horsepower impose
negative positional externalities on other drivers.\730\
---------------------------------------------------------------------------
\730\ IPI, Docket No. NHTSA-2021-0053-1579-A1, at 22.
IPI further states that these negative externalities associated
with other vehicle attributes would also offset opportunity costs
associated with reduced deployment of these attributes where valued by
consumers.
CFA commented that the agency should include a $.90 macroeconomic
stimulus for every dollar of net reduction in driving expenses.\731\
CFA did not provide any details or support for their claim, nor did it
describe how to handle factors like up-front costs. We find CFA's
argument without support.
---------------------------------------------------------------------------
\731\ CFA, Docket No. NHTSA-2021-0053-1535, at 5.
---------------------------------------------------------------------------
A number of commenters argued that the agency should include the
ancillary costs of electric vehicles, such as building additional
charging stations,\732\ improving the grid,\733\ and potential tax
credits given to individuals that purchase electric vehicles.\734\ As
noted elsewhere in this rule and within many of the same comments, many
of these issues are already being addressed by government at the
Federal and state-level. Counting those costs here would be duplicative
to include those costs in this rulemaking.
---------------------------------------------------------------------------
\732\ See, e.g., NATSO and SIGMA, NHTSA-2021-0053-1570, at 10.
\733\ Walter Kreucher, NHTSA-2021-0053-0013, at 14.
\734\ Id. At 14.
---------------------------------------------------------------------------
H. Simulating Safety Effects of Regulatory Alternatives
The primary objective of CAFE standards is to achieve maximum
feasible fuel economy, thereby reducing fuel consumption. In setting
standards to achieve this intended effect, the potential of the
standards to affect vehicle safety is also considered. As a safety
agency, we have long considered the potential for adverse safety
consequences when establishing CAFE standards.
This safety analysis includes the comprehensive measure of safety
impacts from three factors:
1. Changes in Vehicle Mass. Similar to previous analyses, we
calculate the safety impact of changes in vehicle mass made to
reduce fuel consumption and comply with the standards. Statistical
analysis of historical crash data indicates reducing mass in heavier
vehicles generally improves safety, while reducing mass in lighter
vehicles generally reduces safety. Our crash simulation modeling of
vehicle design concepts for reducing mass revealed similar effects.
These observations align with the role of mass disparity in crashes;
when vehicles of different masses collide, the smaller vehicle will
experience a larger change in velocity (and, by extension, force),
which increases the risk to its occupants. As discussed below, in
our analysis, any effect of changes in mass on vehicle safety was
not sufficiently precisely estimated to distinguish it from zero
statistically.
2. Impacts of Vehicle Prices on Fleet Turnover. Vehicles have
become safer over time through a combination of new safety
regulations and voluntary safety improvements. We expect this trend
to continue as emerging technologies, such as advanced driver
assistance systems, are incorporated into new vehicles. Safety
improvements will likely continue regardless of changes to CAFE
standards. As discussed in Section III.E.2, technologies added to
comply with fuel economy standards have an impact on vehicle prices,
therefore slowing the acquisition of newer vehicles and retirement
of older ones. The delay in fleet turnover caused by the effect of
new vehicle prices affects safety by slowing the penetration of new
safety technologies into the fleet.
The standards also influence the composition of the light-duty
fleet. As the safety provided by light trucks, SUVs and passenger
cars responds differently to technology that manufacturers employ to
meet the standards--particularly mass reduction--fleets with
different compositions of body styles will have varying numbers of
fatalities, so changing the share of each type of light-duty vehicle
in the projected future fleet impacts safety outcomes.
3. Increased driving because of better fuel economy. The
``rebound effect'' predicts consumers will drive more when the cost
of driving declines. More stringent standards reduce vehicle
operating costs, and in response, some consumers may choose to drive
more. Additional driving increases exposure to risks associated with
motor vehicle travel, and this added exposure translates into higher
fatalities and injuries.
The contributions of the three factors described above generate the
differences in safety outcomes among regulatory alternatives.\735\ Our
analysis makes extensive efforts to allocate the differences in safety
outcomes between the three factors. Fatalities expected during future
years under each alternative are projected by deriving a fleet-wide
fatality rate (fatalities per vehicle mile of travel) that incorporates
the effects of differences in each of the three factors from baseline
conditions and multiplying it by that alternative's expected VMT.
Fatalities are converted into a societal cost by multiplying fatalities
with the DOT-recommended value of a statistical life (VSL) supplemented
by economic impacts that are external to VSL measurements. Traffic
injuries and property damage are also modeled directly using the same
process and valued using costs that are specific to each injury
severity level.
---------------------------------------------------------------------------
\735\ The terms safety performance and safety outcome are
related but represent different concepts. When we use the term
safety performance, we are discussing the intrinsic safety of a
vehicle based on its design and features, while safety outcome is
used to describe whether a vehicle has been involved in an accident
and the severity of the accident. While safety performance
influences safety outcomes, other factors such as environmental and
behavioral characteristics also play a significant role.
---------------------------------------------------------------------------
All three factors influence predicted fatalities, but only two of
them--changes in vehicle mass and in the composition of the light-duty
fleet in response to changes in vehicle prices--impose increased risks
on drivers and passengers that are not compensated for by accompanying
benefits. In contrast, increased driving associated with the rebound
effect is a consumer choice that reveals the benefit of additional
travel. Consumers who choose to drive more have apparently concluded
that the utility of additional driving exceeds the additional costs for
doing so, including the crash risk that they perceive additional
driving involves. As discussed in Chapter 7 of the accompanying TSD,
the benefits of rebound driving are accounted for by offsetting a
portion of the added safety costs.
We categorize safety outcome through three measures of light-duty
vehicle safety: Fatalities to occupants occurring in crashes, serious
injuries sustained by occupants, and the number of vehicles involved in
crashes that cause property damage but no injuries. Counts of
fatalities to occupants of automobiles and light trucks are obtained
from the Fatal Accident Reporting System (FARS). Estimates of the
number of serious injuries to drivers and passengers of light-duty
vehicles are tabulated from the General Estimates System (GES), an
annual sampling of motor vehicle crashes occurring throughout the U.S.
Weights for different types of crashes were used to expand the samples
of each type to
[[Page 25889]]
estimates of the total number of crashes occurring during each year.
Finally, estimates of the number of automobiles and light trucks
involved in property damage-only (PDO) crashes each year were also
developed using GES.
1. Changes in Vehicle Mass
Similar to previous analyses, we calculate the safety impact of
changes in vehicle mass made to reduce fuel consumption and comply with
the standards. While reduction in mass should have a beneficial safety
effect overall by reducing average fleet mass, a statistical analysis
of historical crash data indicates that reducing mass in heavier
vehicles generally improves safety, while reducing mass in lighter
vehicles generally reduces safety. Our crash simulation modeling of
vehicle design concepts for reducing mass revealed similar effects.
These observations align with the role of mass disparity in crashes:
When vehicles of different masses collide, the smaller vehicle will
experience a larger change in velocity (and, by extension, force),
which increases the risk to its occupants. As discussed below, while
NHTSA's current analysis did not find a statistically significant
relationship between mass and safety, it did find results that are
directionally consistent with previous NHTSA and other studies,
illustrating a common pattern across all studies is that changes in
mass disparity are associated with changes in motor vehicle safety:
Increased disparity increases fatality risk, while decreased disparity
decreases risk. The historical relationship may be changing, however,
and merits ongoing study, which NHTSA is pursuing.
2. Mass Reduction Impacts
Vehicle mass reduction can be one of the more cost-effective means
of improving fuel economy, particularly for makes and models not
already built with much high-strength steel or aluminum closures or
low-mass components. Manufacturers have stated that they will continue
to reduce vehicle mass to meet more stringent standards, and therefore,
this expectation is incorporated into the modeling analysis supporting
the standards. Safety trade-offs associated with mass-reduction have
occurred in the past, particularly before CAFE standards were
attribute-based; past safety trade-offs may have occurred because
manufacturers chose at the time, in response to CAFE standards, to
build smaller and lighter vehicles. In cases where fuel economy
improvements were achieved through reductions in vehicle size and mass,
the smaller, lighter vehicles did not fare as well in crashes as
larger, heavier vehicles, on average. We now, however, use attribute-
based standards, in part to reduce or eliminate the incentive to
downsize vehicles to comply with CAFE standards, but we must continue
to be mindful of the possibility of related safety trade-offs.
For this final rule, we employed the modeling technique developed
in the 2016 Puckett and Kindelberger report to analyze the updated
crash and exposure data by examining the cross sections of the societal
fatality rate per billion vehicle miles of travel (VMT) by mass and
footprint, while controlling for driver age, gender, and other factors,
in separate logistic regressions for five vehicle groups and nine crash
types.\736\ We utilized the relationships between weight and safety
from this analysis, expressed as percentage increases in fatalities per
100-pound weight reduction (which is how mass reduction is applied in
the technology analysis; see Section III.D.4, to examine the weight
impacts applied in this CAFE analysis. The effects of mass reduction on
safety were estimated relative to (incremental to) the regulatory
baseline in the CAFE analysis, across all vehicles for MY 2021 and
beyond.
---------------------------------------------------------------------------
\736\ 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. 2016-0068). Washington, DC: National Highway Traffic Safety
Administration.
---------------------------------------------------------------------------
In computing the impact of changes in mass on safety, we are faced
with competing challenges. Research has consistently shown that mass
reduction affects ``lighter'' and ``heavier'' vehicles differently
across crash types. The 2016 Puckett and Kindelberger report found mass
reduction concentrated among the heaviest vehicles is likely to have a
beneficial effect on overall societal fatalities, while mass reduction
concentrated among the lightest vehicles is likely to have a
detrimental effect on fatalities. This represents a relationship
between the dispersion of mass across vehicles in the fleet and
societal fatalities: Decreasing dispersion is associated with a
decrease in fatalities. Mass reduction in heavier vehicles is more
beneficial to the occupants of lighter vehicles than it is harmful to
the occupants of the heavier vehicles. Mass reduction in lighter
vehicles is more harmful to the occupants of lighter vehicles than it
is beneficial to the occupants of the heavier vehicles.
To accurately capture the differing effect on lighter and heavier
vehicles, we split vehicles into lighter and heavier vehicle
classifications in the analysis. However, this poses a challenge of
producing statistically meaningful results. There are limited relevant
crash data to use for the analysis. Each partition of the data reduces
the number of observations per vehicle classification and crash type,
and thus reduces the statistical robustness of the results. The
methodology we employed was designed to balance these competing forces
as an optimal trade-off to accurately capture the impact of mass-
reduction across vehicle curb weights and crash types while preserving
the potential to identify robust estimates.
The boundary between ``lighter'' and ``heavier'' cars is 3,201
pounds (which is the median mass of MY 2004-2011 cars in fatal crashes
in CY 2006-2012, up from 3,106 pounds for MY 2000-2007 cars in CY 2002-
2008 in the 2012 NHTSA safety database, and up from 3,197 pounds for MY
2003-2010 cars in CY 2005-2011 in the 2016 NHTSA safety database).
Likewise, for truck-based LTVs, curb weight is a two-piece linear
variable with the boundary at 5,014 pounds (again, the MY 2004-2011
median, higher than the median of 4,594 pounds for MY 2000-2007 LTVs in
CY 2002-2008 and the median of 4,947 pounds for MY 2003-2010 LTVs in CY
2005-2011). CUVs and minivans are grouped together in a single group
covering all curb weights of those vehicles; as a result, curb weight
is formulated as a simple linear variable for CUVs and minivans.
Historically, CUVs and minivans have accounted for a relatively small
share of new-vehicle sales over the range of the data, resulting in
fewer crash data available than for cars or truck-based LTVs. In sum,
vehicles are distributed into five groups by class and curb weights:
Passenger cars <3,201 pounds; passenger cars 3,201 pounds or greater;
truck-based LTVs <5,014 pounds; truck-based LTVs 5,014 pounds or
greater; and all CUVs and minivans.
Table III-39 presents the estimated percent increase in U.S.
societal fatality risk per ten billion VMT for each 100-pound reduction
in vehicle mass, while holding footprint constant, for each of the five
vehicle classes.
[[Page 25890]]
[GRAPHIC] [TIFF OMITTED] TR02MY22.107
Techniques developed in the 2011 (preliminary) and 2012 (final)
Kahane reports have been retained to test statistical significance and
to estimate 95 percent confidence bounds (sampling error) for mass
effects and to estimate the combined annual effect of removing 100
pounds of mass from every vehicle (or of removing different amounts of
mass from the various classes of vehicles), while holding footprint
constant. Confidence bounds estimate only the sampling error internal
to the data used in the specific analysis that generated the point
estimate. Point estimates are also sensitive to the modification of
components of the analysis, as discussed at the end of this section.
However, this degree of uncertainty is methodological in nature rather
than statistical.
None of the estimated effects has 95-percent confidence bounds that
exclude zero, and thus are not statistically significant at the 95-
percent confidence level. We have evaluated these results and provided
them for the purposes of transparency. Sensitivity analyses have
confirmed that the exclusion of these statistically insignificant
results would not affect our policy determination, because the net
effects of mass reduction on safety costs are small relative to
predominant estimated benefit and cost impacts. Among the estimated
effects, the most important effects of mass reduction are, as expected,
concentrated among the lightest and heaviest vehicles. Societal
fatality risk is estimated to: (1) Increase by 1.2 percent if mass is
reduced by 100 pounds in the lighter cars; and (2) decrease by 0.61
percent if mass is reduced by 100 pounds in the heavier truck-based
LTVs. These estimates align with the predominant view regarding the
relationship between mass disparity in the vehicle fleet and societal
fatalities: All else being equal, making the heaviest vehicles lighter
(i.e., reducing mass disparity in the fleet) will reduce societal
fatalities, while making the lightest vehicles lighter (i.e.,
increasing mass disparity) will increase societal fatalities. IPI
commented that we ``should give additional weight to externalities such
as pedestrian fatalities and the impact of increased weight
distribution between vehicles.'' \737\ Pedestrian fatalities are
weighted within the above analysis directly proportional to their
frequency among all societal fatalities involving light-duty vehicles.
Any change to the weighting of pedestrian fatalities would thus involve
valuing the societal cost of a pedestrian fatality as being worth a
different amount from other fatalities involving light-duty vehicle
crashes. IPI did not provide a basis to support their proposal to value
fatalities based on occupancy status differently. We are confident that
the current (and historical) specification of relationships among
vehicle curb weights and societal fatality risk represents the role of
mass disparity in societal fatality risk appropriately, by scaling
societal fatality risk as a positive function of mass disparity through
the intuitive coefficients for the lightest and heaviest vehicles (and
through muted coefficients for vehicles with mass closer to the
median).
---------------------------------------------------------------------------
\737\ IPI, Docket No. NHTSA-2021-0053-1579, at 3, 22.
---------------------------------------------------------------------------
The ACC commented that groups including NAS/NASEM have noted that
future improvements in vehicle design could weaken the relationship
between mass disparity and societal fatality rates over time.\738\ We
acknowledge this view, and remain confident that our approach is the
best available representation of the relationship between mass
disparity and societal fatality rates subject to the data available for
analysis, and note again that in our analysis, any effect of changes in
mass on vehicle safety was not sufficiently precisely estimated to
distinguish it from zero at all standard confidence levels used in the
scientific literature.
---------------------------------------------------------------------------
\738\ ACC, Docket No. NHTSA-2021-0053-1564-A1, at 7.
---------------------------------------------------------------------------
Multiple commenters proposed that, due to the limited statistical
significance of the estimates, it would be more appropriate to assume
that changes in vehicle mass in response to CAFE standards will have no
effect on societal fatalities.\739\ NHTSA's current analysis did not
find a statistically significant relationship between mass and safety.
This may reflect the effects of a decreased sample size (the current
study was based on 32 percent fewer fatal cases than the Kahane 2012
study), as well as possible mitigating effects from newer safety
technologies or vehicle designs. While not finding statistical
significance, NHTSA's current study did find results that are
directionally consistent with previous NHTSA studies and a fleet
simulation study by George Washington University.\740\ The common
pattern across all studies is that changes in mass disparity are
associated with changes in motor vehicle safety: Increased disparity
increases fatality risk, while decreased disparity decreases risk. The
agency will
[[Page 25891]]
continue to conduct research on the effects of mass disparity on
vehicle safety in an effort to identify the impacts of evolving vehicle
fleets.
---------------------------------------------------------------------------
\739\ IPI, at 30-1; Consumer Reports, Docket No. NHTSA-2021-
0053-1576, Appendix 9, at 17-8; CARB, Docket No. NHTSA-2021-0053-
1537, Appendix 11, at 269; CBD et al., Docket No. NHTSA-2021-0053-
1572, Appendix 2, at 20; CBD et al., Appendix 1, at 4.
\740\ In response to questions of whether designs and materials
of more recent model year vehicles may have weakened the historical
statistical relationship between mass, size, and safety, NHTSA
updated its public database for statistical analysis consisting of
crash data. The database incorporates the full range of real-world
crash types. NHTSA also sponsored a study conducted by George
Washington University to develop a fleet simulation model and study
the impact and relationship of light-weighted vehicle design with
crash injuries and fatalities. That study is discussed in detail in
Chapter 7.1.5 of the TSD. The study focused on vehicles from MY 2001
to MY 2011, as discussed in the TSD, and found results that are
directionally consistent with NHTSA's statistical analyses of
vehicle mass and fatality risk.
---------------------------------------------------------------------------
We have assessed whether the inclusion of these results would
affect the overall analysis. Because the impacts are very small, we
concluded that it does not have a significant effect on the analysis or
any effect on the choice of standards. Given this conclusion, we
maintain that it is reasonable for the analysis to use the best
available estimates of the impacts of mass reduction that results from
changes in mass disparities on crash fatalities, even if the estimates
are not statistically significant at the 95-percent confidence level.
The estimated statistical significance is limited, but the results
offer some evidence that the relevant point estimates are meaningfully
different from zero (e.g., approximately five to six times more likely
to be non-zero than zero). They are also consistent with a time series
of estimates that represent a relationship that is consistent with
predominant views regarding mass disparity. We believe it would be
inappropriate to ignore these data or to use values of zero for the
rulemaking analysis. Specifically, negative point estimates for heavier
LTVs and positive point estimates for lighter passenger cars have been
found consistently across prior rulemakings. Smaller estimates
corresponding to vehicles near the median of the fleet curb weight
distribution are likely to be less informative due to both statistical
(i.e., small coefficients are less likely to be statistically
significant for a given level of sampling error) and physical (i.e., a
given change in mass will have a smaller effect on societal fatalities
for vehicles near the median mass) concerns.
An additional factor supporting continuing to quantify the safety
impacts related to changes in mass is the sensitivity analysis
including passenger cars with AWD summarized below; when including cars
with AWD, the estimated coefficients are likewise consistent with
previous NHTSA analyses and have statistical significance near the 95-
percent confidence level. Chapter 5 of the FRIA discusses four
sensitivity analyses that were presented for public comment in the
NPRM. We did not identify any comments on the alternative approaches;
in turn, we will defer the decision whether to incorporate the results
into the CAFE Model to subsequent rulemakings. The relevant alternative
with respect to statistical significance centers on aligning passenger
cars with the rest of the sample by including cars that are equipped
with AWD. In previous analyses, passenger cars with AWD were excluded
from the analysis because they represented a sufficiently low share of
the vehicle fleet that statistical relationships between AWD status and
societal fatality risk were highly prone to being conflated with other
factors associated with AWD status (e.g., location, luxury vehicle
status). However, the share of AWD passenger cars in the fleet has
grown. Approximately one-quarter of the passenger cars in the database
have AWD, compared to an approximately five-percent share in the MY
2000-2007 database. Furthermore, all other vehicle types in the
analysis include AWD as an explanatory variable. Thus, we find
expanding the sample size to include a considerable portion of the
real-world fleet (i.e., passenger cars with AWD) to be a meaningful
consideration.
Including passenger cars with AWD in the analysis has little effect
on the point estimate for lighter passenger cars (increase in societal
fatality rates of approximately 1.1 percent per 100-pound mass
reduction, versus 1.2 percent in the central analysis). However, this
revision has a strong effect on the point estimate for heavier
passenger cars (increase in societal fatality rates of between 0.84 and
0.89 percent per 100-pound mass reduction, versus 0.42 percent in the
central analysis). This result supports a hypothesis that, after taking
AWD status into account, mass reduction in heavier passenger cars is a
more important driver of societal fatality rates than previously
estimated. Although this result could be spurious, estimated 95-percent
confidence bounds (from -0.57 to 2.80 percent for lighter passenger
cars, and from -0.14 to 1.82 percent for heavier passenger cars for the
CYs evaluated in the sensitivity analysis) indicate that accounting for
AWD status reduces uncertainty in the point estimate.
A more detailed description of the mass-safety analysis can be
found in Chapter 7 of the accompanying TSD.
3. Sales/Scrappage Impacts
The sales and scrappage responses to higher vehicle prices
discussed in Section III.E.2 have important safety consequences and
influence safety through the same basic mechanism, fleet turnover. In
the case of the scrappage response, delaying fleet turnover keeps
drivers in older vehicles which tend to be less safe than newer
vehicles.\741\ Similarly, the sales response slows the rate at which
newer vehicles, and their associated safety improvements, enter the on-
road population. The sales response also influences the mix of vehicles
on the road--with more stringent CAFE standards leading to a higher
share of light trucks sold in the new vehicle market, assuming all else
is equal. This occurs because there is diminishing value to marginal
improvements in fuel economy (there are fewer gallons to be saved), and
as the difference in consumption between light trucks and passenger
cars diminishes, the other attributes of the trucks will likely lead to
increases in their market share--especially under lower gas prices.
Light trucks have higher rates of fatal crashes when interacting with
passenger cars and, as earlier discussed, different directional
responses to mass reduction technology based on the existing mass and
body style of the vehicle.
---------------------------------------------------------------------------
\741\ See Passenger Vehicle Occupant Injury Severity by Vehicle
Age and Model Year in Fatal Crashes, Traffic Safety Facts Research
Note, DOT-HS-812-528, National Highway Traffic Safety
Administration, April 2018, and The Relationship Between Passenger
Vehicle Occupant Injury Outcomes and Vehicle Age or Model Year in
Police-Reported Crashes, Traffic Safety Facts Research Note, DOT-HS-
812-937, National Highway Traffic Safety Administration, March,
2020.
---------------------------------------------------------------------------
Any effects on fleet turnover (either from delayed vehicle
retirement or deferred sales of new vehicles) will affect the
distribution of both ages and model years present in the on-road fleet.
Because each of these vintages carries with it inherent rates of fatal
crashes, and newer vintages are generally safer than older ones,
changing that distribution will change the total number of on-road
fatalities under each regulatory alternative. Similarly, the dynamic
fleet share model captures the changes in the fleet's composition of
cars and trucks. As cars and trucks have different fatality rates,
differences in fleet composition across the alternatives will affect
fatalities.
At the highest level, the agency calculates the impact of the sales
and scrappage effects by multiplying the VMT of a vehicle by the
fatality risk of that vehicle. For this analysis, calculating VMT is
rather simple: The agency uses the distribution of miles calculated in
TSD Chapter 4.3. The trickier aspect of the analysis is creating
fatality rate coefficients. The fatality risk measures the likelihood
that a vehicle will be involved in a fatal accident per mile driven.
The agency calculates the fatality risk of a vehicle based on the
vehicle's model year, age, and style, while controlling for factors
which are
[[Page 25892]]
independent of the intrinsic nature of the vehicle, such as behavioral
characteristics. Using this same approach, the agency designed separate
models for fatalities, non-fatal injuries, and property damaged
vehicles.
The fatality risk projections described above capture the
historical evolution of safety. Given that modern technologies are
proliferating faster than ever and offer greater safety benefits than
traditional safety improvements, the agency augmented the fatality risk
projections with knowledge about forthcoming safety improvements. The
agency applied detailed empirical estimates of the market uptake and
improving effectiveness of crash avoidance technologies to estimate
their effect on the fleet-wide fatality rate, including explicitly
incorporating both the direct effect of those technologies on the crash
involvement rates of new vehicles equipped with them, as well as the
``spillover'' effect of those technologies on improving the safety of
occupants of vehicles that are not equipped with these
technologies.\742\
---------------------------------------------------------------------------
\742\ These technologies included Forward Collision Warning
(FCW), Crash Imminent Braking (CIB), Dynamic Brake Support (DBS),
Pedestrian AEB (PAEB), Rear Automatic Braking, Semi-automatic
Headlamp Beam Switching, Lane Departure Warning (LDW), Lane Keep
Assist (LKA), and Blind Spot Detection (BSD). While Autonomous
vehicles offer the possibility of significantly reducing or
eventually even eliminating the effect of human error in crash
causation, a contributing factor in roughly 94 percent of all
crashes, there is insufficient information and certainty regarding
autonomous vehicles eventual impact to include them in this
analysis.
---------------------------------------------------------------------------
The agency's approach to measuring these impacts is to derive
effectiveness rates for these advanced crash-avoidance technologies
from safety technology literature. The agency then applies these
effectiveness rates to specific crash target populations for which the
crash avoidance technology is designed to mitigate and adjusted to
reflect the current pace of adoption of the technology, including the
public commitment by manufactures to install these technologies. The
products of these factors, combined across all 6 advanced technologies,
produce a fatality rate reduction percentage that is applied to the
fatality rate trend model discussed above, which projects both vehicle
and non-vehicle safety trends. The combined model produces a projection
of impacts of changes in vehicle safety technology as well as
behavioral and infrastructural trends. A much more detailed discussion
of the methods and inputs used to make these projections of safety
impacts from advanced technologies is included in Chapter 7 of the
accompanying TSD.
Securing America's Future Energy commented that our analysis should
account for improvements in safety over time as crash-avoidance
technologies become more prevalent in the vehicle fleet.\743\ We agree
with this approach, and have accounted for this expected effect in this
and the previous rulemaking by projecting baseline fatality and injury
rates to decrease as a function of the adoption of crash-avoidance
technologies.
---------------------------------------------------------------------------
\743\ Securing America's Future Energy, Docket No. NHTSA-2021-
0053-1513-A1, at 14-15.
---------------------------------------------------------------------------
4. Rebound Effect Impacts
The additional VMT resulting from the rebound effect is accompanied
by more exposure to risk, though rebound miles are not imposed on
consumers by regulation. They are a freely chosen activity resulting
from reduced vehicle operational costs and reflect the perceived
benefit of additional tr