The 2021 EPA Automotive Trends Report: Greenhouse Gas Emissions, Fuel Economy, and Technology since 1975


The
EPA Automotive
Trends Report	
Greenhouse Gas Emissions,
Fuel Economy, and
Technology since 1975
United States
Environmental Protection
Agency
EPA-420-R-21-023 November 2021

-------
This technical report does not necessarily represent final EPA decisions, positions, or validation of
compliance data reported to EPA by manufacturers. It is intended to present technical analysis of issues
using data that are currently available and that may be subject to change. Historic data have been
adjusted, when appropriate, to reflect the result of compliance investigations by EPA or any other
corrections necessary to maintain data integrity.
The purpose of the release of such reports is to facilitate the exchange of technical information and to
inform the public of technical developments. This edition of the report supersedes all previous versions.
jm
lift
ES2

-------
Table of Contents
1.	Introduction	1
A.	What's New This Year	1
B.	Manufacturers in this Report	2
C.	Fuel Economy and C02 Metrics in this Report	3
2.	Fleetwide Trends Overview	5
A.	Overall Fuel Economy and C02 Trends	5
B.	Production Trends	8
C.	Manufacturer Fuel Economy and C02 Emissions	9
3.	Vehicle Attributes	14
A.	Vehicle Class and Type	14
B.	Vehicle Weight	19
C.	Vehicle Power	23
D.	Vehicle Footprint	27
E.	Vehicle Type and Attribute Tradeoffs	30
4.	Vehicle Technology	37
A.	Technology Overview	37
B.	Vehicle Propulsion	40
C.	Vehicle Drivetrain	62
D.	Technology Adoption	67
5.	Manufacturer GHG Compliance	76
A.	Footprint-Based C02 Standards	78
B.	Model Year Performance	81
C.	GHG Program Credits and Deficits	107
D.	End of Year GHG Program Credit Balances	117
Appendices: Methods and Additional Data
A.	Sources of Input Data
B.	Harmonic Averaging of Fuel Economy Values
C.	Fuel Economy and C02 Metrics
D.	Historical Changes in the Database and Methodology
E.	Electric Vehicle and Plug-In Hybrid Metrics
F.	Authors and Acknowledgments
i

-------
List of Figures
Figure 2.1. Estimated Real-World Fuel Economy and C02 Emissions 5
Figure 2.2. Trends in Fuel Economy and C02 Emissions Since Model Year 1975 6
Figure 2.3. Distribution of New Vehicle C02 Emissions by Model Year 7
Figure 2.4. New Vehicle Production by Model Year 9
Figure 2.5. Changes in Estimated Real-World Fuel Economy and C02 Emissions by Manufacturer 10
Figure 3.1. Regulatory Classes and Vehicle Types Used in This Report 15
Figure 3.2. Production Share and Estimated Real-World Fuel Economy 16
Figure 3.3. Vehicle Type Distribution by Manufacturer for Model Year 2020 17
Figure 3.4. Car-Truck Classification of SUVs with Inertia Weights of 4000 Pounds or Less 18
Figure 3.5. Average New Vehicle Weight by Vehicle Type 20
Figure 3.6. Inertia Weight Class Distribution by Model Year 21
Figure 3.7. Relationship of Inertia Weight and C02 Emissions 22
Figure 3.8. Average New Vehicle Horsepower by Vehicle Type 24
Figure 3.9. Horsepower Distribution by Model Year 25
Figure 3.10. Relationship of Horsepower and C02 Emissions 26
Figure 3.11. Calculated 0-to-60 Time by Vehicle Type 27
Figure 3.12. Footprint by Vehicle Type for Model Year 2008-2021 28
Figure 3.13. Footprint Distribution by Model Year 29
Figure 3.14. Relationship of Footprint and C02 Emissions 30
Figure 3.15. Relative Change in Fuel Economy, Weight, Horsepower, and Footprint 31
Figure 4.1. Vehicle Energy Flow 37
Figure 4.2. Manufacturer Use of Emerging Technologies for Model Year 2020 39
Figure 4.3. Production Share by Engine Technology 41
Figure 4.4. Gasoline Engine Production Share by Number of Cylinders 43
Figure 4.5. Percent Change for Specific Gasoline Engine Metrics 45
Figure 4.6. Engine Metrics for Different Gasoline Technology Packages 47
Figure 4.7. Gasoline Turbo Engine Production Share by Vehicle Type 49
Figure 4.8. Gasoline Turbo Engine Production Share by Number of Cylinders 49
Figure 4.9. Distribution of Gasoline Turbo Vehicles by Displacement and Horsepower, Model Year
2011, 2014, and 2020 50
Figure 4.10. Non-Hybrid Stop/Start Production Share by Vehicle Type 52
Figure 4.11. Non-Hybrid Stop/Start Production Share by Number of Cylinders 52
Figure 4.12. Gasoline Hybrid Engine Production Share by Vehicle Type 53
Figure 4.13. Gasoline Hybrid Engine Production Share by Number of Cylinders 53
Figure 4.14. Production Share of EVs, PHEVs, and FCVs 56
Figure 4.15. Electric Vehicle Production Share by Vehicle Type 57
Figure 4.16. Plug-In Hybrid Vehicle Production Share by Vehicle Type 57
Figure 4.17. Charge Depleting Range and Fuel Economy for EVs and PHEVs 58
Figure 4.18. Diesel Engine Production Share by Vehicle Type 60

-------
Figure 4.19. Diesel Engine Production Share by Number of Cylinders 60
Figure 4.20. Percent Change for Specific Diesel Engine Metrics 61
Figure4.21 .Transmission Production Share 64
Figure 4.22. Average Number of Transmission Gears 65
Figure 4.23. Comparison of Manual and Automatic Transmission Real-World Fuel Economy for
Comparable Vehicles 66
Figure 4.24. Front-, Rear-, and Four-Wheel Drive Production Share 67
Figure 4.25. Industry-Wide Car Technology Penetration after First Significant Use 69
Figure 4.26. Manufacturer Specific Technology Adoption over Time for Key Technologies 71
Figure 5.1. The GHG Compliance Process 76
Figure 5.2. 2012-2020 Model Year C02 Footprint Target Curves 78
Figure 5.3. Changes in "2-Cycle" Tailpipe C02 Emissions by Manufacturer (g/mi) 83
Figure 5.4. Model Year 2020 Production of EVs, PHEVs, and FCVs 86
Figure 5.5. Model Year 2020 Advanced Technology Credits by Manufacturer 86
Figure 5.6. HFO-1234yf Adoption by Manufacturer 89
Figure 5.7. Fleetwide A/C Credits by Credit Type 91
Figure 5.8. Total A/C Credits by Manufacturer for Model Year 2020 91
Figure 5.9. Off-Cycle Menu Technology Adoption by Manufacturer, Model Year 2020 93
Figure 5.10. Total Off-Cycle Credits by Manufacturer for Model Year 2020 102
Figure 5.11. Performance and Standards by Manufacturer, Model Year 2020 108
Figure 5.12. Early Credits by Manufacturer 114
Figure 5.13. Total Credits Transactions 116
Figure 5.14. Manufacturer Credit Balance After Model Year 2020 119
Figure 5.15. Industry Performance and Standards, Credit Generation and Use 123
List of Tables
Table 1.1. Model Year 2020 Manufacturer Definitions 3
Table 1.2. Fuel Economy and C02 Metrics Used in this Report 4
Table 2.1. Production, Estimated Real-World C02, and Fuel Economy for Model Year 1975-2021 11
Table 2.2. Manufacturers and Vehicles with the Highest Fuel Economy, by Year 12
Table 2.3. Manufacturer Estimated Real-World Fuel Economy and C02 Emissions for Model Year
2019-2021 13
Table 3.1. Vehicle Attributes by Model Year 32
Table 3.2. Estimated Real-World Fuel Economy and C02 by Vehicle Type 33
Table 3.3. Model Year 2020 Vehicle Attributes by Manufacturer 34
Table 3.4. Model Year 2020 Estimated Real-World Fuel Economy and C02 by Manufacturer and
Vehicle Type 35
Table 3.5. Footprint by Manufacturer for Model Year 2019-2021 (ft2) 36
Table 4.1. Production Share by Powertrain 72

-------
Table 4.2. Production Share by Engine Technologies 73
Table 4.3. Production Share by Transmission Technologies 74
Table 4.4. Production Share by Drive Technology 75
Table 5.1. Manufacturer Footprint and Standards for Model Year 2020 80
Table 5.2. Production Multipliers by Model Year 85
Table 5.3. Model Year 2020 Off-Cycle Technology Credits from the Menu, by Manufacturer and
Technology (g/mi) 98
Table 5.4. Model Year 2020 Off-Cycle Technology Credits from an Alternative Methodology, by
Manufacturer and Technology (g/mi) 101
Table 5.5. Manufacturer Performance in Model Year 2020, All (g/mi) 104
Table 5.6. Industry Performance by Model Year, All (g/mi) 104
Table 5.7. Manufacturer Performance in Model Year 2020, Car (g/mi) 105
Table 5.8. Industry Performance by Model Year, Car (g/mi) 105
Table 5.9. Manufacturer Performance in Model Year 2020, Truck (g/mi) 106
Table 5.10. Industry Performance by Model Year, Truck (g/mi) 106
Table 5.11. Credits Earned by Manufacturers in Model Year 2020, All 110
Table 5.12. Total Credits Earned in Model Years 2009-2020, All 110
Table 5.13. Credits Earned by Manufacturers in Model Year 2020, Car 111
Table 5.14. Total Credits Earned in Model Years 2009-2020, Car 111
Table 5.15. Credits Earned by Manufacturers in Model Year 2020, Truck 112
Table 5.16. Total Credits Earned in Model Years 2009-2020, Truck 112
Table 5.17. Example of a Deficit Offset with Credits from Previous Model Years 117
Table 5.18. Final Credit Balance by Manufacturer for Model Year 2020 (Mg) 120
Table 5.19. Distribution of Credits by Expiration Date (Mg) 121
iv

-------
1. Introduction
This annual report is part of the U.S. Environmental Protection Agency's (EPA) commitment
to provide the public with information about new light-duty vehicle greenhouse gas (GHG)
emissions, fuel economy, technology data, and auto manufacturers' performance in
meeting the agency's GHG emissions standards.
EPA has collected data on every new light-duty vehicle model sold in the United States
since 1975, either from testing performed by EPA at the National Vehicle Fuel and
Emissions Laboratory in Ann Arbor, Michigan, or directly from manufacturers using official
EPA test procedures. These data are collected to support several important national
programs, including EPA criteria pollutant and GHG standards, the U.S. Department of
Transportation's National Highway Traffic Safety Administration (NHTSA) Corporate
Average Fuel Economy (CAFE) standards, and vehicle Fuel Economy and Environment
labels. This expansive data set allows EPA to provide a uniquely comprehensive analysis of
the automotive industry over the last 45 years.
A. What's New This Year
This report is updated each year to reflect the most recent data, best methodology, and
any relevant regulatory changes. This version of the report contains the most up to date
data available to EPA for all model years and supersedes all previous reports. The major
updates for this year are as follows:
• In August 2021, EPA and NHTSA proposed rules to revise the existing light-duty GHG
and fuel economy standards for model years 2023-2026 and 2024-2026, respectively.
In addition, EPA is reconsidering the withdrawal of California's waiver to enforce
greenhouse gas standards for cars and light trucks. Since these proposals have not
been finalized, they are not reflected in this report. Any applicable regulatory changes
finalized by EPA and NHTSA will be included in future versions of this report.
• Fiat Chrysler Automobiles (FCA) and the PSA group finalized a merger to become a
new company called Stellantis. The merger does not impact the grouping of any
manufacturers or makes in this report; however, the manufacturer name Stellantis,
instead of FCA, is now used throughout the report to reflect this change.
• For model year 2020, Volvo and Lotus submitted data as one manufacturer for
compliance with the light-duty GHG program, since both companies are majority
owned by Zhejiang Geely Holding Group Co., Ltd (Geely). EPA determinations related
1

-------
to this merger are subject to change and will be updated in future reports as
necessary.
• This report shows projected model year 2021 data that was generally provided to EPA
by manufacturers during the outbreak of COVID-19. Given the impacts of COVID-19
and worldwide supply chain issues, and their associated impacts on the automobile
industry, the projected model year 2021 data may change significantly before being
finalized.
• EPA continues to update detailed data from this report, including all years of the light-
duty GHG standards, to the EPA Automotive Trends website. We encourage readers to
visit https://www.epa.gov/automotive-trends and explore the data. EPA will continue
to add content and tools on the web to allow transparent access to public data.
B. Manufacturers in this Report
The underlying data for this report include every new light-duty vehicle offered for sale in
the United States. These data are presented by manufacturer throughout this report, using
model year 2020 manufacturer definitions determined by EPA and NHTSAfor
implementation of the GHG emission standards and CAFE program. For simplicity, figures
and tables in the executive summary and in Sections 1-4 show only the top 14
manufacturers, by production. These manufacturers produced at least 150,000 vehicles
each in the 2020 model year and accounted for more than 97% of all production. The
compliance discussion in Section 5 includes all manufacturers, regardless of production
volume. Table 1.1 lists all manufacturers that produced vehicles in the U.S. for model year
2020, including their associated makes, and their categorization for this report. Only vehicle
brands produced in model year 2020 are shown in this table; however, this report contains
data on many other manufacturers and brands that have produced vehicles for sale in the
U.S. since 1975.
When a manufacturer grouping changes under the GHG and CAFE programs, EPA applies
the new manufacturer definitions to all prior model years for the analysis of estimated real-
world CO2 emission and fuel economy trends in Sections 1 through 4 of this report. This
maintains consistent manufacturer and make definitions over time, which enables better
identification of long-term trends. However, the compliance data that are discussed in
Section 5 of this report maintain the previous manufacturer definitions where necessary to
preserve the integrity of compliance data as accrued.
2

-------
Table 1.1. Model Year 2020 Manufacturer Definitions

Manufacturer
Makes in the U.S. Market

BMW
BMW, Mini, Rolls Royce

Ford
Ford, Lincoln, Roush, Shelby

GM
Buick, Cadillac, Chevrolet, GMC

Honda
Acura, Honda

£?
3
+J
U
Mazda
Mazda
TO
_l
£
3
Mercedes
Maybach, Mercedes, Smart

£
CO
Nissan
Infiniti, Nissan

Stellantis
Subaru
Tesla
Toyota
Volkswagen (VW)
Alfa Romeo, Chrysler, Dodge, Fiat, Jeep, Maserati, Ram
Subaru
Tesla
Lexus, Scion, Toyota
Audi, Bentley, Bugatti, Lamborghini, Porsche, Volkswagen

l/l
Jaguar Land Rover
Jaguar, Land Rover

0)
L.
Mitsubishi
Mitsubishi
-------
are measured using EPA city and highway test procedures (the "2-cycle" tests), and
fleetwide averages are calculated by weighting the city and highway test results by 55% and
45%, respectively. These procedures are required for compliance; however, they no longer
accurately reflect real-world driving. Compliance data may also encompass optional
performance credits and adjustments that manufacturers can use towards meeting their
emissions standards.
The second category is estimated real-world data, which is measured using additional
laboratory tests to capture a wider range of operating conditions (including hot and cold
weather, higher speeds, and faster accelerations) encountered by an average driver. This
expanded set of tests is referred to as "5-cycle" testing. City and highway results are
weighted 43% city and 57% highway, consistent with fleetwide driver activity data. The city
and highway values are the same values found on new vehicle fuel economy labels,
however the label combined value is weighted 55% city and 45% highway. Unlike
compliance data, the method for calculating real-world data has evolved over time, along
with technology and driving habits.
Table 1.2. Fuel Economy and CO2 Metrics Used in this Report
C02 and Fuel Economy
Data Category
Purpose
Current
City/ Highway
Weighting
Current Test
Basis

Basis for manufacturer

Compliance
compliance with standards
55% / 45%
2-cycle

Best estimate of real-world

Estimated Real-World
performance
43% / 57%
5-cycle
This report will show estimated real-world data except for the discussion specific to the
GHG regulations in Section 5 and Executive Summary Figures ES-6 through ES-8. The
compliance CO2 data must not be compared to the real-world C02data presented
elsewhere in this report. Appendices C and D present a more detailed discussion of the fuel
economy and CO2 data used in this report.
This report does not provide data about NHTSA's CAFE program. For more information
about CAFE and manufacturer compliance with the CAFE fuel economy standards, see the
CAFE Public Information Center, which can be accessed at
https://one.nhtsa.gov/cafe pic/CAFE PIC Home.htm.
-------
2. Fleetwide Trends Overview
The automotive industry continues to make progress towards lower tailpipe CO2 emissions
and higher fuel economy in recent years. This section provides an update on the estimated
real-world tailpipe CO2 emissions and fuel economy for the overall fleet, and for
manufacturers based on final model year 2020 data. The unique, historical data on which
this report is based also provide an important backdrop for evaluating the more recent
performance of the industry. Using that data, this section will also explore basic fleetwide
trends in the automotive industry since EPA began collecting data in model year 1975.
A. Overall Fuel Economy and CO2 Trends
In model year 2020, the industry
achieved record low new vehicle
CO2 emissions and record high fuel
economy, as shown in Figure 2.1.
Average estimated real-world CO2
tailpipe emissions fell by 7 g/mi to
349 g/mi, while estimated real-world
fuel economy increased 0.5 mpg to
25.4 mpg compared to the previous
year.1
Since model year 2004, CO2emissions
have decreased 24%, or 112 g/mi, and
fuel economy has increased 32%, or
6.1 mpg. Over that time, CO2
emissions and fuel economy have
improved in thirteen out of sixteen
years. The trends in CO2 emissions
and fuel economy since 1975 are
shown in Figure 2.1.
Preliminary data suggest that CO2
emissions and fuel economy in model
1 EPA generally uses unrounded values to calculate values in the text, figures, and tables in this report. This
approach results in the most accurate data but may lead to small apparent discrepancies due to rounding.
E
O)
to
c
o
CO

E
o
c
o
o
LU
2
i—
o
ca
$
12.5

•

25.4 MPG

MY 2020

1975 1985 1995 2005 2015 2025
Model Year
5
-------
year 2021 will remain near the levels achieved in 2020. The preliminary model year 2021
data are based on production estimates provided to EPA by manufacturers months before
the vehicles go on sale. The data are a useful indicator, however there is always uncertainty
associated with such projections, and we caution the reader against focusing only on these
data. Projected data are shown in Figure 2.1 as a dot because the values are based on
manufacturer projections rather than final data.
While the most recent annual changes often receive the most public attention, the greatest
value of the Trends database is to document long-term trends. The magnitude of changes
in annual CO2 emissions and fuel economy tend to be small relative to longer, multi-year
trends. Figure 2.2 shows fleetwide estimated real-world CO2 emissions and fuel economy
for model years 1975-2020. Over this timeframe there have been three basic phases: 1) a
rapid improvement of CO2 emissions and fuel economy between 1975 and 1987, 2) a
period of slowly increasing CO2 emissions and decreasing fuel economy through 2004, and
3) decreasing CO2 emissions and increasing fuel economy through the current model year.
Figure 2.2. Trends in Fuel Economy and CO2 Emissions Since Model Year 1975
-------
Vehicle CO2 emissions and fuel economy are inversely related for gasoline and diesel
vehicles, but not for electric vehicles. Since gasoline and diesel vehicles have made up the
vast majority of vehicle production since 1975, Figure 2.2 shows an inverted, but highly
correlated relationship between CO2 emissions and fuel economy. Electric vehicles, which
account for a small but growing portion of vehicle production, have zero tailpipe CO2
emissions, regardless of fuel economy (as measured in miles per gallon equivalent, or
mpge). If electric vehicles continue to capture a larger market share, the overall
relationship between fuel economy and tailpipe C02 emissions will change.
Another way to look at CO2 emissions over time is to examine how the distribution of new
vehicle emission rates have changed. Figure 2.3 shows the distribution of real-world
tailpipe C02 emissions for all vehicles produced within each model year. Half of the vehicles
produced each year are clustered within a small band around the median CO2 emission
rate, as shown in blue. The remaining vehicles show a much wider spread, especially in the
best and worst 5% of production each year. The lowest C02-emitting vehicles have all been
hybrids or electric vehicles since the first hybrid was introduced in model year 2000. The
highest C02-emitting vehicles are generally low volume performance vehicles or large
trucks.
Figure 2.3. Distribution of New Vehicle CO2 Emissions by Model Year2
1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025
Model Year
2 Electric vehicles prior to 2011 are not included in this figure due to limited data. However, those vehicles were
available in small numbers only.
7
-------
It is important to note that the methodology used in this report for calculating estimated
real-world fuel economy and CO2 emission values has changed over time to reflect
changing vehicle technology and operation. For example, the estimated real-world fuel
economy for a 1980s vehicle is somewhat higher than it would be if the same vehicle were
being produced today. These changes are small for most vehicles, but larger for very high
fuel economy vehicles. See Appendix C and D for a detailed explanation of fuel economy
metrics and their changes over time.
B. Production Trends
This report is based on the total number of vehicles produced by manufacturers for sale in
the United States by model year. Model year is the manufacturer's annual production
period which includes January 1 of the same calendar year. A typical model year for a
vehicle begins in fall of the preceding calendar year and runs until late in the next calendar
year. However, model years vary among manufacturers and can occur between Janurary 2
of the preceding calendaryear and the end of the calendaryear. Model year production
data is the most direct way to analyze emissions, fuel economy, technology, and
compliance trends because vehicle designs within a model year do not typically change.
The use of model year production may lead to some short-term discrepencies with other
sources, which typically report calendar year sales; however, sales based on the calendar
year generally encompass more than one model year, which complicates any analysis.
Since the inception of this report, production of vehicles for sale in the United States has
grown about 0.5% year over year, but there have been significant swings up or down in any
given model year due to the impact of multiple market forces. For example, in model year
2009 the Great Recession resulted in the lowest model year production since the start of
this report, at 9.3 million vehicles. Production rebounded over the next several model
years, reaching an all-time high of more than 17 million vehicles in model year 2017. Model
year 2020 production fell 15% from the previous year, as the COVID-19 pandemic had
wide-ranging impacts on the economy and vehicle production. Production is expected to
rebound somewhat in model year 2021, but the ongoing COVID-19 pandemic, as well as
supply chain disruptions affecting the availability of semiconductors and other components
continue to challenge the industry. Figure 2.4 shows the production trends by model year
for model years 1975 to 2020.
8
-------
Figure 2.4. New Vehicle Production by Model Year
20,000-
O
O
o
^ 15,000-
O
o
d
TJ
2
CL
§ 10,000-
c
c
<
5,000-
C. Manufacturer Fuel Economy and CO2
Emissions
Along with the overall industry, most manufacturers have improved new vehicle CO2
emission rates and fuel economy in recent years. Manufacturer trends over the last five
years are shown in Figure 2.5. This span covers the approximate length of a vehicle
redesign cycle, and it is likely that most vehicles have undergone design changes in this
period, resulting in a more accurate depiction of recent manufacturer trends than focusing
on a single year. Changes over this time period can be attributed to both vehicle design
and changing vehicle production trends. The change in production trends, and the impact
on the trends shown in Figure 2.5 are discussed in more detail in the next section.
For model year 2020 alone, Tesla's all-electric fleet had by far the lowest tailpipe CO2
emissions and highest fuel economy of all large manufacturers. Tesla was followed by
Honda, Subaru, and Hyundai. Stellantis had the highest new vehicle average CO2 emissions
and lowest fuel economy of the large manufacturers in model year 2020, followed by Ford
and GM.
1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025
Model Year
9
-------
Figure 2.5. Changes in Estimated Real-World Fuel Economy and CO2 Emissions by Manufacturer
Fuel Economy (MPG), 2015 - 2020 C02 Emissions (g/mi), 2015 - 2020
Tesla
Honda
Subaru
Hyundai
Nissan
Mazda
Kia
Toyota
BMW
VW
Mercedes
GM
Ford
Stellantis
All Manufacturers
—r~
60
97.1
119.1
~
—1—
80
—1—
100
120

28.5-~29.1
28.1 >-28.5
27.5 —~28.4
27.9-428.0
27.9-4 29.2

25
z

25.!
i^-26.1

25
!3,

23.4>S
22.2-^23.0
22.8 ~23.C
21.3*4-21.8
u
ZU.L)

I-

24.6
->
25.4

50
100
150
200
305-4-312
312<317
312*4-324
316^317

OU'
r jia
320-4 3'
¦1

341 ~
339 —
347
~3J

4
379-4381
386-4—
386-438
4(

399
9
)9-^418

349
4-
360

20
24
28
32
300
350
400
450

-------
Table 2.1. Production, Estimated Real-World CO2, and Fuel Economy for Model Year 1375-2021

Production
Real-World
Real-World

Production
Real-World
Real-World
Model Year
(000)
C02 (g/mi)
FE(MPG)
Model Year
(000)
C02 (g/mi)
FE(MPG)
1975
10,224
681
13.1
2000
16,571
450
19.8
1976
12,334
625
14.2
2001
15,605
453
19.6
1977
14,123
590
15.1
2002
16,115
457
19.5
1978
14,448
562
15.8
2003
15,773
454
19.6
1979
13,882
560
15.9
2004
15,709
461
19.3
1980
11,306
466
19.2
2005
15,892
447
19.9
1981
10,554
436
20.5
2006
15,104
442
20.1
1982
9,732
425
21.1
2007
15,276
431
20.6
1983
10,302
426
21.0
2008
13,898
424
21.0
1984
14,020
424
21.0
2009
9,316
397
22.4
1985
14,460
417
21.3
2010
11,116
394
22.6
1986
15,365
407
21.8
2011
12,018
399
22.3
1987
14,865
405
22.0
2012
13,449
377
23.6
1988
15,295
407
21.9
2013
15,198
368
24.2
1989
14,453
415
21.4
2014
15,512
369
24.1
1990
12,615
420
21.2
2015
16,739
360
24.6
1991
12,573
418
21.3
2016
16,278
359
24.7
1992
12,172
427
20.8
2017
17,016
357
24.9
1993
13,211
426
20.9
2018
16,259
353
25.1
1994
14,125
436
20.4
2019
16,139
356
24.9
1995
15,145
434
20.5
2020
13,721
349
25.4
1996
13,144
435
20.4
2021 (prelim)

348
25.3
1997
14,458
441
20.2

1998
14,456
442
20.1

1999
15,215
451
19.7

To explore this data in more depth, please see the report website at https://www.epa.gov/automotive-trends.
11
-------
Table 2.2. Manufacturers and Vehicles with the Highest Fuel Economy, by Year
Overall Vehicle with

Manufacturer
Manufacturer
Highest Fuel Economy4

with Highest Fuel Economy

with Highest
with Lowest

Real-

Fuel Economy3
Fuel Economy

World FE
Engine

World FE
Model Year
(mpg)
(mpg)
Vehicle
(mpg)
Type
Gasoline Vehicle
(mpg)
1975
Honda
Ford
Honda Civic
28.3
Gas
Honda Civic
28.3
1980
VW
Ford
VW Rabbit
40.3
Diesel
Nissan 210
36.1
1985
Honda
Mercedes
GM Sprint
49.6
Gas
GM Sprint
49.6
1990
Hyundai
Mercedes
GM Metro
53.4
Gas
GM Metro
53.4
1995
Honda
Stellantis
Honda Civic
47.3
Gas
Honda Civic
47.3
2000
Hyundai
Stellantis
Honda Insight
57.4
Hybrid
GM Metro
39.4
2005
Honda
Ford
Honda Insight
53.3
Hybrid
Honda Civic
35.1
2006
Mazda
Ford
Honda Insight
53.0
Hybrid
Toyota Corolla
32.3
2007
Toyota
Mercedes
Toyota Prius
46.2
Hybrid
Toyota Yaris
32.6
2008
Hyundai
Mercedes
Toyota Prius
46.2
Hybrid
Smart Fortwo
37.1
2009
Toyota
Stellantis
Toyota Prius
46.2
Hybrid
Smart Fortwo
37.1
2010
Hyundai
Mercedes
Honda FCX
60.2
FCV
Smart Fortwo
36.8
2011
Hyundai
Mercedes
BMW Active E
100.6
EV
Smart Fortwo
35.7
2012
Hyundai
Stellantis
Nissan-i-MiEV
109.0
EV
Toyota iQ
36.8
2013
Hyundai
Stellantis
Toyota IQ
117.0
EV
Toyota iQ
36.8
2014
Mazda
Stellantis
BMW i3
121.3
EV
Mitsubishi Mirage
39.5
2015
Mazda
Stellantis
BMW i3
121.3
EV
Mitsubishi Mirage
39.5
2016
Mazda
Stellantis
BMW i3
121.3
EV
Mazda 2
37.1
2017
Honda
Stellantis
Hyundai loniq
132.6
EV
Mitsubishi M
rage
41.5
2018
Tesla
Stellantis
Hyundai loniq
132.6
EV
Mitsubishi M
rage
41.5
2019
Tesla
Stellantis
Hyundai loniq
132.6
EV
Mitsubishi M
rage
41.6
2020
Tesla
Stellantis
Tesla Model 3 SR+
138.6
EV
Mitsubishi M
rage
41.6
2021 (prelim)
Tesla
GM
Tesla Model 3 SR+
139.9
EV
Mitsubishi M
rage
40.1
3 Manufacturers below the 150,000 threshold for "large" manufacturers are excluded in years they did not meet the threshold.
4 Vehicles are shown based on estimated real-world fuel economy as calculated for this report. These values will differ from values found on the fuel
economy labels at the time of sale. For more information on fuel economy metrics see Appendix C.
12
-------
Table 2.3. Manufacturer Estimated Real-World Fuel Economy and CO2 Emissions for Modr Y*\ r 2019-2021
MY 2019 Final MY 2020 Final MY2021 Preliminary
FE Change C02 Change

Real-World
Real-World
Real-World
from
Real-World
from
Real-World
Real-World

FE
C02
FE
MY 2019
C02
MY 2019
FE
C02
Manufacturer
(mpg)
(g/mi)
(mpg)
(mpg)
(g/mi)
(g/mi)
(mpg)
(g/mi)
BMW
26.2
337
25.5
-0.6
347
10
25.5
345
Ford
22.5
395
23.0
0.5
386
-9
22.7
390
GM
22.5
395
23.0
0.6
386
-9
21.5
415
Honda
28.9
307
29.1
0.2
305
-2
29.0
305
Hyundai
28.5
311
28.4
-0.1
312
1
29.0
305
Kia
28.1
316
27.7
-0.4
320
4
28.1
316
Mazda
27.8
320
27.9
0.1
319
-1
27.4
324
Mercedes
23.7
374
23.4
-0.3
379
5
24.0
370
Nissan
27.0
329
27.9
1.0
317
-11
29.2
303
Stellantis
21.2
418
21.3
0.1
418
-1
21.6
410
Subaru
28.4
312
28.5
0.0
312
0
28.4
312
Tesla
118.0
0
119.1
1.1
0
0
125.7
0
Toyota
25.8
345
27.0
1.2
329
-16
26.1
339
VW
26.1
338
25.0
-1.2
354
16
25.9
333
All Manufacturers
24.9
356
25.4
0.5
349
-7
25.3
348
To explore this data in more depth, please see the report website at https://www.epa.gov/automotive-trends.
13
-------
3. Vehicle Attributes
Vehicle CO2 emissions and fuel economy are strongly influenced by vehicle design
parameters, including weight, power, acceleration, and size. In general, vehicles that are
larger, heavier, and more powerful typically have lower fuel economy and higher CO2
emissions than other comparable vehicles. This section focuses on several key vehicle
design attributes that impact CO2 emissions and fuel economy and evaluates the impact of
a changing automotive marketplace on overall fuel economy.
A. Vehicle Class and Type
Manufacturers offer a wide variety of light-duty vehicles in the United States. Under the
CAFE and GHG regulations, new vehicles are separated into two distinct regulatory classes,
cars and trucks, and each vehicle class has separate GHG and fuel economy standards.
Vehicles that weigh more than 6,000 pounds gross vehicle weight5 (GVW) or have four-
wheel drive and meet various off-road requirements, such as ground clearance, qualify as
trucks. Vehicles that do not meet these requirements are considered cars.
Pickup trucks, vans, and minivans are all considered trucks under the regulatory
definitions, while sedans, coupes, and wagons are generally classified as cars. Sport utility
vehicles (SUVs), fall into both categories. Based on the CAFE and GHG regulatory
definitions, most two-wheel drive SUVs under 6,000 pounds GVW are classified as cars,
while most SUVs that have four-wheel drive or are above 6,000 pounds GVW are
considered trucks. SUV models that are less than 6,000 pounds GVW can have both car and
truck variants, with two-wheel drive versions classified as cars and four-wheel drive
versions classified as trucks. As the fleet has changed over time, the line drawn between
car and truck classes has also evolved. This report uses the current regulatory car and truck
definitions, and these changes have been propagated back throughout the historical data.
This report further separates the car and truck regulatory classes into five vehicle type
categories based on their body style classifications under the fuel economy labeling
program. The regulatory car class is divided into two vehicle types: sedan/wagon and car
SUV. The sedan/wagon vehicle type includes minicompact, subcompact, compact, midsize,
large, and two-seater cars, hatchbacks, and station wagons. Vehicles that are SUVs under
the labeling program and cars under the CAFE and GHG regulations are classified as car
5 Gross vehicle weight is the combined weight of the vehicle, passengers, and cargo of a fully loaded vehicle.
14
-------
SUVs in this report. The truck class is divided into three vehicle types: pickup, minivan/van,
and truck SUV. Vehicles that are SUVs under the labeling program and trucks under the
CAFE and GHG regulations are classified as truck SUVs. Figure 3.1 shows the two regulatory
classes and five vehicle types used in this report. The distinction between these five vehicle
types is important because different vehicle types have different design objectives, and
different challenges and opportunities for improving fuel economy and reducing CO2
emissions.
Figure 3.1. Regulatory Classes and Vehicle Types Used in This Report
Regulatory Class Vehicle Type
Fuel Economy and CO2 by Vehicle Type
The production volume of the different vehicle types has changed significantly over time.
Figure 3.2 shows the production shares of each of the five vehicle types since model year
1975. The overall new vehicle market continues to move away from the sedan/wagon
vehicle type towards a combination of truck SUVs, car SUVs, and pickups. Sedan/wagons
were the dominant vehicle type in 1975, when more than 80% of vehicles produced were
sedan/wagons. Since then, their production share has generally been falling, reaching a
new low in model year 2020 at only 31 % of all production, or far less than half of the
market share they held in model year 1975.
Vehicles that could be classified as a car SUV or truck SUV were a very small part of the
production share in 1975 but now account for more than half of all new vehicles produced.
In model year 2020, truck SUVs reached a record high 39% of production and car SUVs
reached a record high of 13% of production. The production share of pickups has
15
-------
fluctuated over time, peaking at 19% in 1994 and then falling to 10% in 2012. Pickups have
generally increased in recent years and accounted for 14% of the market in model year
2020. Minivan/vans captured less than 5% of the market in 1975, increased to 11 % in
model year 1995 but have fallen since to 3% of vehicle production. The projected 2021 data
shows a continuing shift away from sedan/wagons and towards truck SUVs and pickups.
In model year 2020, 44% of the fleet were cars and 56% were trucks. This was the highest
percentage of trucks on record and a significant change from 1975. In Figure 3.2, the
dashed line between the car SUVs and truck SUVs shows the split in car and truck
regulatory class.
Figure 3.2. Production Share and Estimated Real-World Fuel Economy
0
CO
.c
(1.1
c
o
•*-'
o
T3
O
100%
75%
50%
25%
0%
Sedan/Wagon
Car SUV
1975
1985
1995 2005
Model Year
2015
2025
Sedan
Wagon
Car
SUV
Truck
SUV
Minivan
Van
Pickup
1975 1985 1995 2005 2015 2025
Model Year
Figure 3.2 also shows estimated real-world fuel economy for each vehicle type since 1975.
All five vehicle types are at record low CO2 emissions and record high fuel economy in
model year 2020. Minivan/Vans, car SUVs, and sedan/wagons all increased fuel economy
by 0.9 mpg, while truck SUVs increased by 0.3 mpg, and pickups increased by 0.2 mpg. All
of the vehicle types, except for pickups, now achieve fuel economy more than double what
they achieved in 1975. In the preliminary model year 2021 data (shown as a dot on Figure
3.2), four of the five vehicle types are expected to improve fuel economy and one,
sedan/wagons, is projected to remain the same as MY 2020.
Overall fuel economy trends depend on the trends within the five vehicle types, but also on
the market share of each of the vehicle types. The trend away from sedan/wagons, which
16
-------
remain the vehicle type with the highest fuel economy and lowest CO2 emissions, and
towards vehicle types with lower fuel economy and higher CO2 emissions, has offset some
of the fleetwide benefits that otherwise would have been achieved from the improvements
within each vehicle type.
Vehicle Type by Manufacturer
The model year 2020 production breakdown by vehicle type for each manufacturer is
shown in Figure 3.3. There are clear variations in production distribution by manufacturer.
Nearly 65% of Tesla's production was sedan/wagons, which is the highest of any
manufacturer. For other vehicle types, Hyundai had the highest percentage of car SUVs at
46%, Subaru had the highest percentage of truck SUVs at 82%, Ford had the highest
percentage of pickups at 31 %, and Stellantis had the highest percentage of minivan/vans at
10%.
Sedan/wagon market penetration fell 2% across the industry in model year 2020, with
reductions from eleven out of fourteen manufacturers. Tesla had the largest change, as
sedan/wagons fell from 91 % of production in model year 2019 to 65% in model year 2020.
Mercedes had the second largest change, with sedan/wagons falling from 51 % of
production to 33%, followed by VW, which saw sedan/wagons fall from 50% to 40% of
production. All three companies increased their relative production of car SUVs or truck
SUVs.
Figure 3.3. Vehicle Type Distribution by Manufacturer for Model Year 2020
^ Lower average C02 Emissions ^
100%
g 75%
_£Z
cn
.0 50%
"G
15
T3
O
25%
0%
W &v/
Vehicle Type
Sedan/Wagon
| Car SUV
Truck SUV
Minivan/Van
I Pickup
17
-------
A Closer Look at SUVs
SUV Classification
Over the last 30 years, the production share of SUVs in the United States has increased in all
but six years and now accounts for more than 50% of all vehicles produced (see Figure 3.2).
This includes both the car and truck SUV vehicle types.
Based on the regulatory definitions of cars and trucks, SUVs that are less than 6,000 pounds
GVW can be classified as either cars or trucks, depending on design requirements such as
minimum angles and clearances, and whether the vehicle has 2-wheel drive or 4-wheel drive.
This definition can lead to similar vehicles having different car or truck classifications, and
different requirements under the GHG and CAFE regulations. One particular trend of interest
is the classification of SUVs as either car SUVs or truck SUVs.
This report does not track GVW, but instead tracks weight using inertia weight classes, where
inertia weight is the weight of the empty vehicle, plus 300 pounds (see weight discussion on
the next page). Figure 3.4 shows the breakdown of SUVs into the car and truck categories
over time for vehicles with an inertia weight of 4,000 pounds or less. Vehicles in the 4,500-
pound inertia weight class and higher were excluded, as these vehicles generally exceed
6,000 pounds GVW and are classified as trucks. The relative percentage of SUVs with an
inertia weight of 4,000 pounds or less that meet the current regulatory truck definition in
model year 2020 is slightly higher than in model year 2000 and is projected to reach a new
high in model year 2021.
Figure 3.4. Car-Truck Classification of SUVs with Inertia Weights of 4000 Pounds or Less
2000 2005 2010 2015 2020
Model Year
18
-------
For some manufacturers, changes in the mix of vehicle types they produce has also led to a
significant impact on their overall new vehicle CO2 emissions and fuel economy. Over the
last five years, as shown in Figure 2.5, Toyota achieved the largest reduction in CO2
emissions of any manufacturer, at 27 g/mi. Toyota decreased emissions in all vehicle types,
and their mix of vehicle types produced in model years 2015 and 2020 was similar. Kia
achieved the second largest reduction in overall CO2 tailpipe emissions, even as their
production share of more efficient sedan/wagons fell from 75% to 53%, and less efficient
truck SUVs increased from 3% to 36% of all production. GM had the third largest reduction
in overall CO2 tailpipe emissions, while their production share of sedan/wagons fell from
31 % to 13% and truck SUVs increased from 28% to 39% of all production.
Over the same five-year period, VW and Mazda tied for the largest increase in new vehicle
CO2 emissions at 15 g/mi. VW achieved reductions in CO2 emissions in both sedan/wagon
and truck SUVs vehicle types; however, that was more than offset by a reduction in their
production share of sedan/wagons, from 79% to 40%, and a corresponding increase in
truck SUVs from 19% to 58%. Mazda had similar trends, as a drop in sedan/wagon
production share from 57% to 24%, a corresponding increase in car and truck SUV
production share, and an increase in sedan/wagon C02 emissions offset reductions in C02
emissions from both car SUVs and truck SUVs.
B. Vehicle Weight
Vehicle weight is a fundamental vehicle attribute, both because it can be related to utility
functions such as vehicle size and features, and because higher weight, other things being
equal, will increase CO2 emissions and decrease fuel economy. All vehicle weight data in
this report are based on inertia weight classes. Each inertia weight class represents a range
of loaded vehicle weights, or vehicle curb weights6 plus 300 pounds. Vehicle inertia weight
classes are in 250-pound increments for classes below 3,000 pounds, while inertia weight
classes over 3,000 pounds are divided into 500-pound increments.
Vehicle Weight by Vehicle Type
Figure 3.5 shows the average new vehicle weight for all vehicle types since model year
1975. From model year 1975 to 1981, average vehicle weight dropped 21 %, from 4,060
pounds per vehicle to about 3,200 pounds; this was likely driven by both increasing fuel
6 Vehicle curb weight is the weight of an empty, unloaded vehicle.
19
-------
economy standards (which, at the time, were universal standards, and not based on any
type of vehicle attribute) and higher gasoline prices.
From model year 1981 to model year 2004, the trend reversed, and average new vehicle
weight began to slowly but steadily climb. By model year 2004, average new vehicle weight
had increased 28% and reached 4,111 pounds per vehicle, in part because of the increasing
truck share. Average vehicle weight in model year 2020 was only about 1.3% above 2004
but has increased slowly over the last several years and is currently at the highest point on
record, at 4,166 pounds. Preliminary model year 2021 data suggest that weight will
continue to increase.
Figure 3.5. Average New Vehicle Weight by Vehicle Type
Sedan/Wagon
ALL
Car SUV

5500

5000

4500

4000

3500

3000
(/)

2500
-1—»

n)
5500
£

5000

4500

4000

3500

3000

2500

•V

/
•
\
/
/
V
L-
y
q
% t
;e MY 1975

0
Sin<
-14%'
Since MY 1975
I
-6%
Since MY 1975
Truck SUV
Minivan/Van

7% t
Since MY 1975

Pickup

r\
/

\
/

28% I
Since MY 1975

1975 1985 1995 2005 2015 20251975 1985 1995 2005 2015 20251975 1985 1995 2005 2015 2025
Model Year
In model year 1975, the difference between the heaviest and lightest vehicle types was
about 215 pounds, or about 5% of the average new vehicle. By model year 2020, the
difference between the heaviest and lightest vehicle types had increased to more than
1,600 pounds, or almost 40% of the average new vehicle weight. Over that time, the weight
of an average new sedan/wagon fell 14% while the weight of an average new pickup
20
-------
increased 28%. In 1975, the average new sedan/wagon outweighed the average new pickup
by about 45 pounds, but the different weight trends over time for each of these vehicle
types led to a very different result in model year 2020, with the average new pickup
outweighing the average new sedan/wagon by more than 1,600 pounds. Pickups did have a
drop of over 300 pounds per average new vehicle in model year 2015, the year the Ford F-
150 was redesigned with a largely aluminum body, and a drop of almost 150 pounds in
model year 2019, which correlates with a redesign of the Ram 1500.
Figure 3.6 shows the annual production share of different inertia weight classes for new
vehicles since model year 1975. In model year 1975 there were significant sales in all
weight classes from <2,750 pounds to 5,500 pounds. In the early 1980s the largest vehicles
disappeared from the market, and light cars <2,750 pounds inertia weight briefly captured
more than 25% of the market. Since then, cars in the <2,750-pound inertia weight class
have all but disappeared, and the market has moved towards heavier vehicles.
Interestingly, the heaviest vehicles in model year 1975 were mostly large cars, whereas the
heaviest vehicles today are largely trucks, with a few luxury vehicles and vans.
Figure 3.6. Inertia Weight Class Distribution by Model Year
100%
75%
<1)
CO
_c
CO
o 50%
o
3
T3
O
25%
0%
Weight
¦
<2750

2750

3000

3500

4000

4500

5000

5500

6000
¦
>6000
1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025
Model Year
21
-------
Vehicle Weight and CO2 Emissions
Heavier vehicles require more energy to move than lower-weight vehicles and, if all other
factors are the same, will have lower fuel economy and higher CO2 emissions. The wide
array of technology available in modern vehicles complicates this comparison, but it is still
useful to evaluate the relationship between vehicle weight and C02 emissions, and how
these variables have changed over time.
Figure 3.7 shows estimated real-world CO2 emissions as a function of vehicle inertia weight
for model year 19787 and model year 2020. On average, C02 emissions increase linearly
with vehicle weight for both model years, although the rate of change as vehicles get
heavier is different. At lower weights, vehicles from model year 2020 produced about two
thirds of the CO2 emissions of 1978 vehicles. The difference between model year 2020 and
1978 increases for heavier vehicles, as the heaviest model year 2020 vehicles produce
about half of the C02 emissions of 1978 vehicles.
Figure 3.7. Relationship of Inertia Weight and CO2 Emissions
Inertia Weight (lbs)
7 Model year 1978 was the first year for which complete horsepower data are available, therefore it will be used
for several historical comparisons for consistency.
22
-------
Electric vehicles, which do not produce any tailpipe CO2 emissions regardless of weight, are
visible along the 0 g/mi axis of Figure 3.7. As more electric vehicles are introduced into the
market, the relationship between average vehicle C02 emissions and inertia weight will
continue to evolve.
C. Vehicle Power
Vehicle power, measured in horsepower (hp), has changed dramatically since model year
1975. In the early years of this report, horsepower fell, from an average of 137 hp in model
year 1975 to 102 hp in model year 1981. Since model year 1981, however, horsepower has
increased almost every year. The average new vehicle in model year 2020 produced 80%
more power than a new vehicle in model year 1975, and 141 % more power than an
average new vehicle in model year 1981. The average new vehicle horsepower is at a
record high, increasing from 245 hp in model year 2019 to 246 hp in model year 2020. The
preliminary value for model year 2021 is 252 hp, which would be another record-high for
horsepower.
Whh ir IVwer by Vehicle Type
As with weight, the changes in horsepower are also quite different among vehicle types, as
shown in Figure 3.8. Horsepower for sedan/wagons increased 51 % between model year
1975 and 2020, 74% for truck SUVs, 77% for car SUVs, 80% for minivan/vans, and 143% for
pickups. Increases in horsepower have been more variable over the last decade, but the
general trend continues to be increasing horsepower. The projected model year 2021 data
shows another expected increase of about 6 hp across all new vehicles.
23
-------
Figure 3.8. Average New Vehicle Horsepower by Vehicle Type
350
300
250
200
150
I 100
o
Cl
-------
Figure 3.9. Horsepower Distribution by Model Year
100% - ^
1980 1990 2000
Model Year
2010
2020
Horsepower

>450
¦
400-450

350-400
¦
300-350

250-300
¦
200-250

150-200

100-150
¦
50-100
Vehicle Power and CO2 Emissions
The relationship between vehicle power, C02 emissions, and fuel economy has become
more complex as new technology and vehicles have emerged in the marketplace. In the
past, higher power generally increased CO2 emissions and decreased fuel economy,
especially when new vehicle production relied exclusively on gasoline and diesel internal
combustion engines. As shown in Figure 3.10, model year 1978 vehicles with increased
horsepower generally had increased C02 emissions. In model year 2020, C02 emissions
increased with increased vehicle horsepower at a much lower rate than in model year
1978, such that model year 2020 vehicles nearly all had lower CO2 emissions than their
model year 1978 counterparts with the same amount of power. Technology improvements,
including turbocharged engines and hybrid packages, have reduced the incremental C02
emissions associated with increased power. Electric vehicles are present along the 0 g/mi
line in Figure 3.10 because they produce no tailpipe CO2 emissions, regardless of
horsepower, further complicating this analysis for modern vehicles.
25
-------
Figure 3.10. Relationship of Horsepower and CO2 Emissions
900
J
O)
CM
o
0 600 -
1
1
CD
-------
The calculated 0-to-60 time for model year 2021 is projected to fall further, to in seconds.
The long-term downward trend in 0-to-60 times is consistent across all vehicle types,
although it appears to be diverging some in more recent years. The average 0-to-60 time
for pickups continues to decrease steadily, while 0-to-60 times for car SUVs have changed
less. The continuing decrease in pickup truck 0-to-60 times is likely due to their increasing
power, as shown in Figure 3.8. While much of that power is intended to increase towing
and hauling capacity, it also decreases 0-to-60 times.
Figure 3.11. Calculated 0-to-60 Time by Vehicle Type
18
15
12
9
U)
TJ
C
15
12
9
1975 1985 1995 2005 2015 20251975 1985 1995 2005 2015 20251975 1985 1995 2005 2015 2025
Model Year
D. Vehicle Footprint
Vehicle footprint is an important attribute since it is the basis for the current C02 emissions
and fuel economy standards. Footprint is the product of wheelbase times average track
width (the area defined by where the centers of the tires touch the ground). This report
provides footprint data beginning with model year 2008, although footprint data from
model years 2008-2010 were aggregated from various sources and EPA has less
ALL Sedan / Wagon Car SUV
Truck SUV
Minivan / Van
-41%
Since MY 1978
4
-41%
Since MY 1978
4
Pickup
-49%
Since MY 1978
4
27
-------
confidence in the precision of these data than that of formal compliance data. Beginning in
model year 2011, the first year when both car and truck CAFE standards were based on
footprint, automakers began to submit reports to EPA with footprint data at the end of the
model year, and these official footprint data are reflected in the final data through model
year 2020. EPA projects footprint data for the preliminary model year 2021 fleet based on
footprint values from the previous model year and, for new vehicle designs, publicly
available data.
Vehicle Footprint by Vehicle Type
Figure 3.12 shows overall new vehicle and vehicle type footprint data since model year
2008. Between model year 2008 and 2020, the overall average footprint increased 4.2%,
from 48.9 to 50.9 square feet. All five vehicle types have increased average footprint since
model year 2008, ranging from a small increase for car SUVs (up 0.4 square feet or 0.8%) to
a larger increase for pickup trucks (up 2.6 square feet, or 4.2%). The overall increase in
footprint is impacted by both the trends within each vehicle type and the changing mix of
vehicles over time, as the market has shifted towards larger vehicles.
Figure 3.12. Footprint by Vehicle Type for Model Year 2008-2021
Pickup
Minivan / Van
Truck SUV
Fleetwide^vg
CapSUV

Sedan/Wagon
2008 2010 2012 2014 2016
Model Year
2018
2020
28
-------
The distribution of footprints across all new vehicles, as shown in Figure 3.13, also shows a
slow reduction in the number of smaller vehicles with a footprint of less than 45 square
feet. Projected data for model year 2021 suggest that overall average footprint will increase
0.5 square feet to 51.4 square feet.
Figure 3.13. Footprint Distribution by Model Year
100%-
2008 2010 2012 2014 2016
Model Year
2018
2020
Footprint
>65
B60-65
55-60
¦ 50-55
45-50
40-45
<40
Vehicle Footprint and C02 Emissions
The relationship between vehicle footprint and CO2 emissions is shown in Figure 3.14.
Vehicles with a larger footprint are likely to weigh more and have more frontal area, which
leads to increased aerodynamic resistance. Increased weight and aerodynamic resistance
increase CO2 emissions and decreases fuel economy. The general trend of increasing
footprint and CO2 emissions holds true for vehicles from model year 2008 and model year
2020, although vehicles produced in model year 2020 are projected to produce roughly
20% less C02 emissions than model year 2008 vehicles of a comparable footprint. Electric
vehicles are shown in Figure 3.14 with zero tailpipe CO2 emissions, regardless of footprint.
As more electric vehicles enter the market, the relationship between footprint and tailpipe
CO2 emissions will become much flatter, or less sensitive to footprint.

-------
Figure 3.14. Relationship of Footprint and CO2 Emissions
750
IT
3
CM
O 500
0
33
S—
1
I
05
£ 250
0
E. Vehicle Type and Attribute Tradeoffs
The past 45+ years of data show striking changes in the mix of vehicle types, and the
attributes of those vehicles, produced for sale in the United States. In the two decades
prior to 2004, technology innovation and market trends generally resulted in increased
vehicle power and weight (due to increasing vehicle size and content) while average new
vehicle fuel economy steadily decreased and C02 emissions correspondingly increased.
Since model year 2004, the combination of technology innovation and market trends have
resulted in average new vehicle fuel economy increasing 32%, horsepower increasing 17%,
and weight increasing 1%. Footprint has increased 4% since EPA began tracking it in model
year 2008. These metrics are all at record highs, and horsepower, weight, and footprint are
projected to increase again in model year 2021, as shown in Figure 3.15.
Between model year 2008 and 2020, fuel economy and footprint increased within each of
the five vehicle types, and horsepower increased in four. Weight decreased within each of
the vehicle types. These trends within vehicle types are largely attributable to design and
technology changes over that time span. In addition to technology changes, the market
shifted towards car and truck SUVs, which are often larger, heavier, more powerful, and
less fuel efficient than sedan/wagons they replaced. These market changes further
Model Year
• 2008
1 2020
•
m
m
• s
•

•
• * ® ® •
• 1* it? g* ¦
i*-?""s 0
• - •»?
if' 1 ¦V'71*
Jr " . N •
• •
••

30 40 50 60 70
Footprint (sq ft)
30
-------
increased the overall horsepower and footprint of the average new vehicle, compared to
the technology-driven changes within each vehicle type alone. The trend towards larger,
heavier, and more powerful vehicles has also offset some of the fleetwide fuel economy
and CO2 emission benefits that otherwise would have been achieved through improving
technology. Market trends led to an increase in the weight of a new average vehicle, even
as weight fell within each vehicle type.
Figure 3.15. Relative Change in Fuel Economy, Weight, Horsepower, and
Footprint
100% -
75% -
r-
CD
CD 50%
o
c
(!)
0
cn
£= 25% H
CO
_c
O
0% -
-25% -
CO

CN
10%-
(1)

()
0%-
c
CD

0
-10%-
cn

•
•

Real-World Fuel Economy

HorseDower

•

Weigh

—1 1 1 1 1 1 1 1 1 1 1—

Footprint

1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025
Model Year
Vehicle fuel economy and CO2 emissions are clearly related to vehicle attributes
investigated in this section, namely weight, horsepower, and footprint. Future trends in fuel
economy and CO2 emissions will be dependent, at least in part, by design choices related to
these attributes.
31
-------
Table 3.1. Vehicle Attributes by Mod r
Model Year
Real-World
co2
(g/mi)
Real-World
FE
(mpg)
Weight
(lbs)
Horsepower
(HP)
0 to 60
(s)
Footprint
(ft2)
Car
Production
Share
Truck
Production
Share
1975
681
13.1
4,060
137
-
-
80.7%
19.3%
1980
466
19.2
3,228
104
15.6
-
83.5%
16.5%
1985
417
21.3
3,271
114
14.1
-
75.2%
24.8%
1990
420
21.2
3,426
135
11.5
-
70.4%
29.6%
1995
434
20.5
3,613
158
10.1
-
63.5%
36.5%
2000
450
19.8
3,821
181
9.8
-
58.8%
41.2%
2005
447
19.9
4,059
209
9.0
-
55.6%
44.4%
2006
442
20.1
4,067
213
8.9
-
57.9%
42.1%
2007
431
20.6
4,093
217
8.9
-
58.9%
41.1%
2008
424
21.0
4,085
219
8.9
48.9
59.3%
40.7%
2009
397
22.4
3,914
208
8.8
47.9
67.0%
33.0%
2010
394
22.6
4,001
214
8.8
48.5
62.8%
37.2%
2011
399
22.3
4,126
230
8.5
49.5
57.8%
42.2%
2012
377
23.6
3,979
222
8.5
48.8
64.4%
35.6%
2013
368
24.2
4,003
226
8.4
49.1
64.1%
35.9%
2014
369
24.1
4,060
230
8.3
49.7
59.3%
40.7%
2015
360
24.6
4,035
229
8.3
49.4
57.4%
42.6%
2016
359
24.7
4,035
230
8.3
49.5
55.3%
44.7%
2017
357
24.9
4,093
234
8.2
49.8
52.6%
47.4%
2018
353
25.1
4,137
241
8.0
50.4
48.0%
52.0%
2019
356
24.9
4,156
245
7.9
50.8
44.4%
55.6%
2020
349
25.4
4,166
246
7.8
50.9
43.9%
56.1%
2021 (prelim)
348
25.3
4,287
252
7.7
51.4
38.9%
61.1%
To explore this data in more depth, please see the report website at https://www.epa.gov/automotive-trends
-------
Table 3.2. Estimated Real-World Fuel Economy and CO2 by Vehicle Type
Sedan/Wagon Car SUV Truck SUV Minivan/Van Pickup

Real-
Real-

World
World

Prod
C02
FE
Prod
C02
FE
Prod
C02
FE
Prod
C02
FE
Prod
C02
FE
Model Year
Share
(g/mi)
(mpg)
Share
(g/mi)
(mpg)
Share
(g/mi)
(mpg)
Share
(g/mi)
(mpg)
Share
(g/mi)
(mpg)
1975
80.6%
660
13.5
0.1%
799
11.1
1.7%
806
11.0
4.5%
800
11.1
13.1%
746
11.9
1980
83.5%
446
20.0
0.0%
610
14.6
1.6%
676
13.2
2.1%
629
14.1
12.7%
541
16.5
1985
74.6%
387
23.0
0.6%
443
20.1
4.5%
538
16.5
5.9%
537
16.5
14.4%
489
18.2
1990
69.8%
381
23.3
0.5%
All
18.8
5.1%
541
16.4
10.0%
498
17.8
14.5%
511
17.4
1995
62.0%
379
23.4
1.5%
499
17.8
10.5%
555
16.0
11.0%
492
18.1
15.0%
526
16.9
2000
55.1%
388
22.9
3.7%
497
17.9
15.2%
555
16.0
10.2%
478
18.6
15.8%
534
16.7
2005
50.5%
379
23.5
5.1%
440
20.2
20.6%
531
16.7
9.3%
460
19.3
14.5%
561
15.8
2006
52.9%
382
23.3
5.0%
434
20.5
19.9%
518
17.2
7.7%
455
19.5
14.5%
551
16.1
2007
52.9%
369
24.1
6.0%
431
20.6
21.7%
503
17.7
5.5%
456
19.5
13.8%
550
16.2
2008
52.7%
366
24.3
6.6%
419
21.2
22.1%
489
18.2
5.7%
448
19.8
12.9%
539
16.5
2009
60.5%
351
25.3
6.5%
403
22.0
18.4%
461
19.3
4.0%
443
20.1
10.6%
526
16.9
2010
54.5%
340
26.2
8.2%
386
23.0
20.7%
452
19.7
5.0%
442
20.1
11.5%
527
16.9
2011
47.8%
344
25.8
10.0%
378
23.5
25.5%
449
19.8
4.3%
424
20.9
12.3%
516
17.2
2012
55.0%
322
27.6
9.4%
381
23.3
20.6%
445
20.0
4.9%
418
21.3
10.1%
516
17.2
2013
54.1%
313
28.4
10.0%
365
24.3
21.8%
427
20.8
3.8%
422
21.1
10.4%
509
17.5
2014
49.2%
313
28.4
10.1%
364
24.4
23.9%
412
21.6
4.3%
418
21.3
12.4%
493
18.0
2015
47.2%
306
29.0
10.2%
353
25.1
28.1%
406
21.9
3.9%
408
21.8
10.7%
474
18.8
2016
43.8%
303
29.2
11.5%
338
26.2
29.1%
400
22.2
3.9%
410
21.7
11.7%
471
18.9
2017
41.0%
293
30.2
11.6%
339
26.1
31.7%
398
22.3
3.6%
399
22.2
12.1%
470
18.9
2018
36.7%
286
30.8
11.3%
324
27.4
35.0%
384
23.1
3.1%
389
22.8
13.9%
466
19.1
2019
32.7%
285
30.9
11.7%
323
27.5
36.5%
378
23.5
3.4%
396
22.4
15.6%
467
19.0
2020
30.9%
277
31.7
13.0%
310
28.4
38.7%
374
23.8
2.9%
379
23.4
14.4%
465
19.2
2021 (prelim)
27.2%
275
31.7
11.7%
282
30.5
41.4%
370
24.0
2.6%
336
26.2
17.1%
461
19.4
To explore this data in more depth, please see the report website at https://www.epa.gov/automotive-trends
-------
Table 3.3. Model Year 2020 Vehicle Attributes by Manufacturer
Real-World Real-World
Manufacturer
C02
(g/mi)
FE
(mpg)
Weight
(lbs)
Horsepower
(HP)
0 to 60
(s)
Footprint
(ft2)
BMW
347
25.5
4324
304
6.6
49.7
Ford
386
23.0
4537
290
7.1
55.4
GM
386
23.0
4403
265
7.7
54.5
Honda
305
29.1
3711
207
8.0
47.7
Hyundai
312
28.4
3574
182
8.7
47.2
Kia
320
27.7
3622
193
8.5
47.5
Mazda
319
27.9
3837
192
8.9
46.3
Mercedes
379
23.4
4533
288
6.9
51.3
Nissan
317
27.9
3852
204
8.8
48.1
Stellantis
418
21.3
4661
295
7.3
54.9
Subaru
312
28.5
3973
197
9.1
46.1
Tesla
0
119.1
4394
444
3.6
50.6
Toyota
329
27.0
4037
223
7.8
49.3
VW
354
24.9
4144
241
7.6
48.1
Other
360
24.5
4316
263
8.0
48.6
All Manufacturers
349
25.4
4166
246
7.8
50.9
To explore this data in more depth, please see the report website at https://www.epa.gov/automotive-trends
-------
Table 3.4. Model Year 2020 Estimated Real-World Fuel Economy and CO2 by Manufacturer and Vehicle Type
Sedan/Wagon Car SUV Truck SUV Minivan/Van Pickup
Real- Real- Real- Real- Real- Real- Real- Real- Real- Real-
World World World World World World World World World World
Prod C02 FE Prod C02 FE Prod C02 FE Prod C02 FE Prod C02 FE
Manufacturer
Share
(g/mi)
(mpg)
Share
(g/mi)
(mpg)
Share
(g/mi)
(mpg)
Share
(g/mi)
(mpg)
Share (g/mi)
(mpg)
BMW
56.3%
327
27.1
5.5%
301
29.5
38.2%
383
23.1
-
-
-
-
-
-
Ford
14.2%
343
25.8
12.4%
311
28.6
39.9%
377
23.5
2.8%
368
24.2
30.8%
449
19.8
GM
13.4%
280
30.9
19.5%
315
28.2
38.7%
400
22.2
-
-
-
28.4%
465
19.3
Honda
48.0%
267
33.2
9.3%
301
29.5
34.6%
340
26.1
5.6%
376
23.6
2.5%
418
21.3
Hyundai
43.7%
256
34.6
45.6%
345
25.7
10.7%
398
22.3
-
-
-
-
-
-
Kia
53.1%
269
32.8
9.1%
331
26.8
35.9%
387
23.0
1.9%
420
21.1
-
-
-
Mazda
23.6%
293
30.3
26.6%
309
28.8
49.8%
336
26.4
-
-
-
-
-
-
Mercedes
32.6%
348
25.6
11.1%
338
26.3
50.6%
407
21.8
5.7%
396
22.4
-
-
-
Nissan
58.6%
284
31.1
9.9%
301
29.6
25.8%
377
23.6
1.8%
353
25.2
3.8%
449
19.8
Stellantis
9.2%
410
21.7
5.9%
335
26.5
46.3%
404
22.0
10.2%
376
23.5
28.3%
476
18.8
Subaru
18.0%
316
28.1
-
-
-
82.0%
311
28.6
-
-
-
-
-
-
Tesla
64.7%
0
124.2
32.3%
0
112.1
3.0%
0
97.1
-
-
-
-
-
-
Toyota
42.6%
255
34.7
9.9%
315
28.2
29.4%
356
25.0
3.3%
398
22.3
14.8%
483
18.4
VW
40.1%
300
29.3
1.8%
434
20.5
58.1%
389
22.7
-
-
-
-
-
-
Other
14.4%
296
30.0
10.2%
320
27.6
73.6%
379
23.3
1.8%
342
26.0
-
-
-
All Manufacturers
30.9%
277
31.7
13.0%
310
28.4
38.7%
374
23.8
2.9%
379
23.4
14.4%
465
19.2
To explore this data in more depth, please see the report website at https://www.epa.gov/automotive-trends
-------
Table 3.5. Footprint by Manufacturer for Model Year 2019-2021 (ft2)
Final MY 2019 Final MY 2020 Preliminary MY2021
Manufacturer
Car
Truck
All
Car
Truck
All
Car
Truck
All
BMW
47.7
52.3
49.3
47.8
52.8
49.7
47.6
52.2
49.4
Ford
46.9
59.1
55.3
47.8
58.1
55.4
47.8
58.5
57.2
GM
45.9
58.3
54.2
45.5
58.9
54.5
46.0
60.1
57.3
Honda
45.9
50.3
47.8
46.1
49.8
47.7
46.5
48.4
47.4
Hyundai
46.6
49.2
46.6
46.5
53.5
47.2
46.8
53.0
47.7
Kia
46.0
49.1
47.0
45.7
50.4
47.5
46.2
49.5
47.6
Mazda
44.9
47.7
46.3
45.5
47.1
46.3
45.7
46.7
46.2
Mercedes
48.6
51.0
49.5
48.8
53.2
51.3
49.7
51.9
50.8
Nissan
46.0
52.1
48.1
46.7
51.1
48.1
46.9
49.7
47.6
Stellantis
49.3
56.3
54.9
49.3
55.8
54.9
49.9
53.3
52.8
Subaru
44.9
46.1
45.9
44.9
46.4
46.1
44.9
45.9
45.7
Tesla
49.8
54.8
49.9
50.4
54.8
50.6
51.3
-
51.3
Toyota
46.5
52.0
49.5
46.3
52.6
49.3
46.6
52.5
50.1
VW
45.3
51.2
48.2
46.3
49.4
48.1
47.0
50.9
49.1
Other
44.5
49.5
48.0
44.8
49.9
48.6
47.1
50.6
49.6
All Manufacturers
46.5
54.2
50.8
46.6
54.3
50.9
47.1
54.1
51.4
To explore this data in more depth, please see the report website at https://www.epa.gov/automotive-trends
-------
4. Vehicle Technology
Since model year 1975, the technology used in vehicles has continually evolved. Today's
vehicles utilize an increasingly wide array of technological solutions developed by the
automotive industry to improve vehicle attributes discussed previously in this report,
including CO2 emissions, fuel economy, vehicle power, and acceleration. Automotive
engineers and designers are constantly creating and evaluating new technology and
deciding how, or if, it should be applied to their vehicles. This section of the report looks at
vehicle technology from two perspectives; first, how the industry has adopted specific
technologies over time, and second, how those technologies have impacted CO2 emissions
and fuel economy.
A. Technology Overview
All vehicles use some type of engine or motor to convert energy stored on the vehicle,
usually in a fuel or battery, into rotational energy to propel the vehicle forward. Internal
combustion engines, for example, typically combust gasoline or diesel fuel to rotate an
output shaft. Internal combustion engines are paired with a transmission to convert the
rotational energy from the relatively narrow range of speeds available at the engine to the
appropriate speed required for driving conditions. The transmission is connected to a
driveline that transfers the rotational energy from the transmission to the two or four
wheels being used to move the vehicle. Each of these components has energy losses, or
inefficiencies, which ultimately increase vehicle CO2 emissions and decrease fuel economy.
A basic illustration of the energy flow through a gasoline vehicle is shown in Figure 4.1.
Figure 4.1. Vehicle Energy Flow
I

Tires
Engine
1 O
Transmission .
1 O
Driveline

4
# *
¦±
Tires
37
-------
Manufacturers have been adopting many new technologies to improve gasoline internal
combustion engines. Figure 4.2 illustrates manufacturer-specific technology adoption for
model year 2020, where larger circles represent higher adoption rates. For gasoline
engines, technologies such as turbocharged engines (Turbo) and gasoline direct injection
(GDI) allow for more efficient engine design and operation. Cylinder deactivation (CD)
allows for only using part of the engine when less power is needed, and stop/start can turn
off the engine entirely when the vehicle is stopped to save fuel. Transmissions that have
seven or more speeds, and continuously variable transmissions (CVTs), allow an engine to
more frequently operate near its peak efficiency, providing more efficient average engine
operation and a reduction in fuel usage.
Manufacturers are also adopting hybrids, plug-in hybrid electric vehicles (PHEVs), electric
vehicles (EVs), and fuel cell vehicles (FCVs). Hybrid vehicles store some propulsion energy in
a battery, and often recapture braking energy, allowing for a smaller, more efficiently
operated engine. Plug-in hybrids operate similarly to hybrids but their batteries can be
charged from an external source of electricity, and generally have a longer electric only
operating range. Electric vehicles operate only on energy stored in a battery that is charged
from an external source of electricity, and rely exclusively on electric motors for propulsion
instead of an internal combustion engine. Fuel cell vehicles use a fuel cell stack to create
electricity from an onboard fuel source (usually hydrogen), which then powers an electric
motor or motors to propel the vehicle. These vehicle technologies offer fundamentally
different architectures than shown in Figure 4.1 and require different metrics9 and an
evolving analysis of vehicle technology. Hybrids, PHEVs and EVs are currently a small
portion of the fleet but are expected to grow rapidly, as most manufacturers have made
recent public announcements committing to billions of dollars in research towards
electrification, and in some cases, manufacturers have announced10 specific targets for
entirely phasing out internal combustion engines.
The technologies in Figure 4.2 are all being adopted by manufacturers to reduce CO2
emissions and increase fuel economy. Each of the fourteen largest manufacturers have
adopted several of these technologies into their vehicles, with many manufacturers
achieving high penetrations of several technologies as shown in Figure 4.2. It is also clear
that manufacturers' strategies to develop and adopt new technologies are unique and vary
significantly. Each manufacturer is choosing technologies that best meet the design
9 See Appendix E for a detailed discussion of EV and PHEV metrics.
10 Preston, B., Bartlett, J. "Automakers Are Adding Electric Vehicles to Their Lineups. Here's What's Coming."
Consumer Reports, https://www.consumerreports.org/hybrids-evs/why-electric-cars-may-soon-flood-the-us-
market-a9006292675/
ftfl 38
-------
requirements of their vehicles, and in many cases, that technology is changing quickly. The
rest of this section will explore how vehicle technology has changed since 1975, the impact
of those technology changes, and the rate at which technology is adopted by the industry.
Figure 4.2. Manufacturer Use of Emerging Technologies for Model Year 2020
Tesla -
Honda -
Subaru -
Hyundai -
Nissan -
Mazda -
Kia -
Toyota -
BMW -
VW-
Mercedes -
GM -
Ford -
Stellantis -
All Manufacturers -
Turbo GDI CVT 7+Gears Non-hybrid CD Hybrid PHEV/
StopStart EV/FC
50% 79% 64% 29% 11% 24% 5%
24% 98% 94% 63%
14% 65% 29% 44% 22%
4% 54% 89% 10%
17% 100%
5% 63% 35% 26% 33%
3% 1% 32% 42% 15%
99% 99%
92% 98%
99% 100%
98% 94%
64%
100%
0%
1%
2% 1%
1%
2% 1%
13% 1%
2%
91% 88% 2% 6% 3%
100% 81% 6% 13% 0%
43% 87% 9% 56% 70% 44% 1%
79% 63% 4% 81% 88% 10% 3% 0%
10% 9% 1% 92% 56% 23% 10% 1%
35% 57% 28% 52% 46% 15% 5% 2%
39
-------
B. Vehicle Propulsion
As discussed above, all vehicles use some type of engine or motor to convert stored energy
into rotational energy to propel the vehicle forward. Over the last 45+ years that EPA has
been collecting data, gasoline internal combustion engines have been the dominant
technology used as a power source in vehicles. Over that time, the technology used in
combustion engines has continually evolved. Modern gasoline combustion engines are
continuing that trend, employing technologies such as direct injection, turbocharging, and
cylinder deactivation to improve efficiency and performance.
A growing portion of new vehicles rely on partial or full electrification to achieve
operational improvements, reduce tailpipe CO2 emissions, and increase fuel economy.
Many new vehicles utilize stop-start technology, which turns off the engine during idle
conditions and uses the vehicle battery to restart the engine when needed. Mild hybrids
generally employ stop-start systems and have an electric motor that can assist the engine
with moving the vehicle forward. Full hybrids generally have larger batteries and motors
that can provide more power to move the vehicle or can directly drive the vehicle without
the engine. Plug-in hybrids (PHEVs) add the capability of charging the vehicle battery from
an external source, namely electricity from the power grid. Full electric vehicles (EVs) rely
on electric motors to provide propulsion and use energy stored onboard in a battery. EVs
are charged with electricity from the power grid, and do not have an internal combustion
engine. Most hybrids, PHEVs, and EVs also utilize regenerative braking to recapture braking
energy that otherwise would have been lost as heat, and further improve vehicle efficiency.
This "spectrum of electrification" is creating a wide range of technology implementation
strategies on modern vehicles, and offering numerous pathways to improve vehicle
efficiency, emissions, and performance.
The trend in vehicle propulsion technology since model year 1975 is shown in Figure 4.3.
Vehicles that use an engine that operates exclusively on gasoline (including hybrids, but not
plug-in hybrids which also use electricity) have held at least 95% of the light-duty vehicle
market in almost every year. Vehicles with diesel engines briefly captured almost 6% of the
market in model year 1981 but have been less than 1 % of the market in most other years
since 1985. PHEVs, EVs, and FCVs have added to the increasing array of technology
available in the automotive marketplace and have been capturing a small but growing
portion of the market. These vehicles captured 2.2% of the market in model year 2020 and
are projected to grow to 4.2% in model year 2021.
40
-------
Figure 4.3. Production Share by Engine Technology
100%
CD
i_
CO
CO
c
o
"G
"O
o
25%
Car
Truck
1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025
Model Year
Fuel Delivery
Valve Timing
Number of Valves
Key
Carbureted
Fixed
Two-N^lve
1

MultiA^lve
2
Throttle Body Injection
Fixed
Two-N^lve
3

MultiA^lve
4
Port Fuel Injection
Fixed
TwoA^lve
5

MultiA^lve
6

Variable
TwoA^lve
7

MultiA^lve
8
Gasoline Direct
Fixed
MultiA^lve
9
Injection (GDI)
Variable
MultiA^lve
10

Two-N^lve
KB
Diesel
—
—
B
EV/PHEV/FCV
—
—

41
-------
Engines that use only gasoline as a fuel (including hybrids) are further divided based on
three broad parameters for Figure 4.3: fuel delivery, valve timing, and number of valves per
cylinder. These parameters enable better control of the combustion process, which in turn
can allow for lower CO2 emissions, increased fuel economy, and/or more power. Fuel
delivery refers to the method of creating an air and fuel mixture for combustion. The
technology for fuel delivery has changed over time from carburetors to fuel injection
systems located in the intake system, and more recently to gasoline direct injection (GDI)
systems that spray gasoline directly into the engine cylinder.
The valves on each cylinder of the engine determine the amount and timing of air entering
and exhaust gases exiting the cylinder during the combustion process. Valve timing has
evolved from fixed timing to variable valve timing (WT), which can allow for much more
precise control. In addition, the number of valves per cylinder has generally increased,
again offering more control of air and exhaust flows. Combined, these changes have led to
modern engines with much more precise control of the combustion process.
Figure 4.3 shows many different engine designs as they have entered, and in many cases
exited, the automotive market. Some fleetwide changes occurred gradually, but in some
cases (for example trucks in the late 1980s), engine technology experienced widespread
change in only a few years. Evolving technology offers opportunities to improve fuel
economy, CO2 emissions, power, and other vehicle parameters. The following analysis will
look at technology trends within gasoline engines (including hybrids), PHEVs and EVs, and
diesel engines. Each of these categories of engine technologies has unique properties,
metrics, and trends over time.
Gasoline Engines
Since EPA began tracking vehicle data in 1975, over 650 million vehicles have been
produced for sale in the United States. For most of those years, vehicles relying on a
gasoline engine as the only source of power captured more than 99% of production. The
only exceptions were in the early 1980s when diesel engines peaked briefly at about 6% of
the market, and more recently as electric vehicle production has increased. For the
purposes of this report, hybrid vehicles are included with gasoline engines, as are "flex fuel"
vehicles that are capable of operating on gasoline or a blend of 85% ethanol and 15%
gasoline (E85).
-------
Engine Size and Displacement
Engine size is generally described in one of two ways, either the number of cylinders or the
total displacement of the engine (the total volume of the cylinders). Engine size is
important because larger engines strongly correlate with higher fuel use. Figure 4.4 shows
the trends in gasoline engine size over time, as measured by number of cylinders.
Figure 4.4. Gasoline Engine Production Share by Number of Cylinders
100%.
CD
s
ro
-C
cn
c
o
-4—'
o
T3
O
75%-
50% -
25%-
0%
Cylinders
Less than 4
4 Cylinder
¦ 5 Cylinder
H 6 Cylinder
8 Cylinder
I More than 8
i 1 1 1 1 1 1 1 1 1
1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025
Model Year
In the mid and late 1970s, the 8-cylinder engine was dominant, accounting for well over
half of all new vehicle production. Between model year 1979 and 1980 there was a
significant change in the market, as 8-cylinder engine production share dropped from 52%
to 23%, and those engines were replaced with smaller 4-cylinder and some 6-cylinder
engines. From model year 1987 through 2004, production moved back towards larger 6-
cylinder and 8-cylinder engines. This trend reversed again in 2005 as production began
trending back towards 4-cylinder engines. Four-cylinder engines are now the most popular
engine option, capturing just under 60% of the market in model year 2020.
43
-------
Overall engine size, as measured by the total volume of all the engine's cylinders, is directly
related to the number of cylinders. As vehicles have moved towards engines with a lower
number of cylinders, the total engine size, or displacement, is also at an all-time low. The
average new vehicle in model year 1975 had a displacement of nearly 300 cubic inches,
compared to an average of 170 cubic inches in model year 2020. Gasoline engine
displacement per cylinder has been relatively stable over the time of this report (around 34
cubic inches per cylinder since 1980), so the reduction in overall new vehicle engine
displacement is almost entirely due to the shift towards engines with fewer cylinders.
The contrasting trends in horsepower (at all-time high) and engine displacement (at an all-
time low) highlight the continuing improvement in engines. These improvements are due
to the development of new technologies and ongoing design improvements that allow for
more efficient use of fuel or reduce internal engine friction. One additional way to examine
the relationship between engine horsepower and displacement is to look at the trend in
specific power (HP/Displacement), which is a metric to compare the power output of an
engine relative to its size.
Specific power has increased nearly 200% between model year 2020 and model year 1975.
The rate at which specific power has increased has been remarkably steady, as shown in
Figure 4.5. The specific power of new vehicle gasoline engines has increased by about 0.02
horsepower per cubic inch every year for 45+ years. Considering the numerous and
significant changes to engines over this time span, changes in consumer preferences, and
the external pressures on vehicle purchases, the long-standing linearity of this trend is
noteworthy. The roughly linear increase in specific power does not appear to be slowing.
Turbocharged engines, direct injection, higher compression ratios, and many other engine
technologies are likely to continue increasing engine specific power.
Figure 4.5 also shows two other important engine metrics, the amount of fuel consumed
compared to the overall size of the engine (Fuel Consumption/Displacement), and the
amount of fuel consumed relative to the amount of power produced by an engine (Fuel
Consumption/HP). The amount of fuel consumed by a gasoline engine in model year 2020,
relative to the total displacement, is about 12% lower than in model year 1975. Fuel
consumption relative to engine horsepower has fallen more than 70% since model year
1975. Taken as a whole, the trend lines in Figure 4.5 clearly show that gasoline engine
improvements overtime have been steady and continual and have resulted in impressive
improvements to internal combustion engines.
44
-------
Figure 4.5. Percent Change for Specific Gasoline Engine Metrics
200% -
150%"
LO
o5 100%-
CD
O
c
cn
g, 50%-
£=
TO
sz
O
0%-
-50%"
1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025
Model Year
Fuel Delivery Systems and Valvetrains
All gasoline engines require a fuel delivery system that controls the flow of fuel delivered
into the engine. The process for controlling fuel flow has changed significantly over time,
allowing for much more control over the combustion process and thus more efficient
engines. In the 1970s and early 1980s, nearly all gasoline engines used carburetors to
meter fuel delivered to the engine. Carburetors were replaced over time with fuel injection
systems; first throttle body injection (TBI) systems, then port fuel injection (PFI) systems,
and more recently gasoline direct injection (GDI), as shown in Figure 4.3. TBI and PFI
systems use fuel injectors to electronically deliver fuel and mix it with air outside of the
engine cylinder; the resulting air and fuel mixture is then delivered to the engine cylinders
for combustion. Engines that utilize GDI spray fuel directly into the air in the engine
cylinder for better control of the combustion process. Engines using GDI were first
introduced into the market with very limited production in model year 2007. Ten years
later, GDI engines were installed in 57% of model year 2020 vehicles.
Another key aspect of engine design is the valvetrain. Each engine cylinder must have a set
of valves that allow for air (or an air/fuel mixture) to flow into the engine cylinder prior to
45
-------
combustion and for exhaust gases to exit the cylinder after combustion. The number of
valves per cylinder and the method of controlling the valves (i.e., the valvetrain) directly
impacts the overall efficiency of the engine. Generally, engines with four valves per cylinder
instead of two, and valvetrains that can alter valve timing during the combustion cycle can
provide more engine control and increase engine power and efficiency.
This report began tracking multi-valve engines (i.e., engines with more than two valves per
cylinder) for cars in model year 1986 and for trucks in model year 1994. Since that time
about 90% of the fleet has converted to multi-valve design. While some three- and five-
valve engines have been produced, the majority of multi-valve engines are based on four
valves per cylinder. Engines with four valves generally use two valves for air intake and two
valves for exhaust. In addition, this report began tracking variable valve timing (VVT)
technology for cars in model year 1990 and for trucks in model year 2000, and since then
nearly the entire fleet has adopted this technology. Figure 4.3 shows the evolution of
engine technology, including fuel delivery method and the introduction of VVT and multi-
valve engines.
As shown in Figure 4.3, fuel delivery and valvetrain technologies have often been
developed simultaneously. Nearly all carbureted engines relied on fixed valve timing and
had two valves per cylinder, as did early port-injected engines. Port-injected engines largely
developed into engines with both multi-valve and VVT technology. Engines with GDI are
almost exclusively using multi-valve and WT technology. These four engine groupings, or
packages, represent a large share of the engines produced over the timespan covered by
this report.
Figure 4.6 shows the changes in specific power and fuel consumption per horsepower for
each of these engine packages over time. There is a very clear increase in specific power of
each engine package as engines moved from carbureted engines, to engines with two
valves, fixed timing and port fuel injection, then to engines with multi-valve WT and port
fuel injection, and finally to GDI engines. Some of the increase for GDI engines may also be
due to the fact that GDI engines are often paired with turbochargers to further increase
power. Vehicles with fixed valve timing and two valves per cylinder have been limited in
recent years and are no longer included in Figure 4.6 after model year 2015 due to very
limited production.
46
-------
Figure 4.6. Engine Metrics for Different Gasoline Technology Packages
GDI Engines
Variable Timing,
Multi-Valve Engines
Fixed Timing,
Two-Valve Engines
Carbureted Engines
Carbureted Engines
Fixed Timing,
Two-Valve Engines
GD
Engi
nes
Variable Timing,
Multi-Valve Engines
1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025
Model Year

-------
Turbocliargieg
Turbochargers increase the power that an engine can produce by forcing more air, and
thus fuel, into the engine. An engine with a turbocharger can produce more power than an
identically sized engine that is naturally aspirated or does not have a turbocharger.
Turbochargers are powered using the pressure of the engine exhaust as it leaves the
engine. Superchargers operate the same way as turbochargers but are directly connected
to the engine for power, instead of using the engine exhaust. Alternate turbocharging and
supercharging methods, such as electric superchargers, are also beginning to emerge. A
limited number of new vehicles utilize both a turbocharger and supercharger in one engine
package.
Gasoline turbocharged engines have been increasing rapidly in the marketplace,
accounting for almost 35% of all production in model year 2020, as shown in Figure 4.7.
Many of these engines are applying turbochargers to create "turbo downsized" engine
packages that can combine the improved fuel economy of smaller engines during normal
operation but can provide the power of a larger engine by engaging the turbocharger when
necessary. As evidence of this turbo downsizing, almost 80% of gasoline turbocharged
engines are 4-cylinder engines in model year 2020, with most other turbochargers being
used in 6-cylinder engines. Model year 2021 is projected to be similar, with a small but
growing number of vehicles equipped with 3-cylinder turbocharged engines. This is shown
in Figure 4.8.
Most of the current gasoline turbocharged engines also use GDI and WT. This allows for
more efficient engine operation, helps increase the resistance to premature combustion
(engine knock), and reduces turbo lag (the amount of time it takes for a turbocharger to
engage). In model year 2020, almost 90% of new vehicles with gasoline turbocharged
engines also used GDI.
Figure 4.9 examines the distribution of engine displacement and power of gasoline
turbocharged engines over time. In model year 2011, turbochargers were used mostly in
cars, and were available on engines both above and below the average engine
displacement. The biggest increase in turbocharger use over the last few years has been in
cars with engine displacement well below the average displacement. The distribution of
horsepower for turbocharged engines is much closer to the average horsepower, even
though the displacement is smaller, reflecting the higher power per displacement of
turbocharged engines. This trend towards adding turbochargers to smaller, less powerful
engines is consistent with the turbo downsizing trend.
48
-------
Figure 4.7. Gasoline Turbo Engine Production Share by Vehicle Type
30% ¦
20% -
10%-
0%-
Vehicle Type
Sedan/Wagon
| Car SUV
Truck SUV
Minivan/Van
| Pickup
2003
2008
2013
Model Year
2018
2023
Figure 4.8. Gasoline Turbo Engine Production Share by Number of Cylinders
Cylinders
4 Cylinder
| 6 Cylinder
8 Cylinder
Other
2003
¦MHlril
2008
2013
Model Year
2018
2023

-------
Figure 4.9. Distribution of Gasoline Turbo Vehicles by Displacement and
Horsepower, Model Year 2011, 2014, and 2020
Horsepower
Displacement (cubic inches)
2,500
2,000
1,500
1,000
500
0
2,500
2,000
1,500
o
o
o,
c
o
8 1,000
TD
O
500
0
2,500
2,000
1,500
1,000
500
0
Mean
HP,

All Cars

! Mean HP,
1 All Trucks

Mean HP,
All Cars
Mean HP,
All Cars
Nk
Mean HP,
All Trucks
Mean Displacerrjent,
All Cars
. Mean Displacement,
All Trucks

Mean HP,
All Trucks

Mean Displacement,!
All Cars
¦

!— Mean Displacement
1 All Trucks

!¦

ro
o

Mean Displace
nent,

All Cars

. Mean Displacement,
All Trucks

—

¦¦
!

ro
o
Truck
Car
ro
o
ro
o
0 100 200 300 400 500 600 700 50 100 150 200 250 300 350 400 450

-------
Cylinder Deactivation
Cylinder deactivation is an engine management approach that turns off the flow of fuel to
one or more engine cylinders when driving conditions do not require full engine power.
This effectively allows a large engine to act as a smaller engine when the additional
cylinders are not needed, increasing engine efficiency and fuel economy. The use of
cylinder deactivation in gasoline vehicles has been steadily climbing, and in model year
2020 gasoline engines with cylinder deactivation were almost 15% of all vehicles. This trend
is expected to continue, especially as new improvements to cylinder deactivation
technology, such as dynamic cylinder deactivation, continue to enter the market.
Stop/Start
Engine stop/start technology allows the engine to be automatically turned off at idle and
very quickly restarted when the driver releases the brake pedal. By turning the engine off, a
vehicle can eliminate the fuel use and CO2 emissions that would have occurred if the
engine was left running. This report began tracking stop/start technology in model year
2012 at less than one percent. Since then, the use of stop/start has increased to almost
46% of all new vehicles in model year 2020, excluding hybrid vehicles. While non-hybrid
stop/start systems have been used in a wide range of applications, they are found more
often in larger vehicles and engines, as shown in Figure 4.10 and Figure 4.11.
Hybrids
Gasoline hybrid vehicles feature a battery pack that is larger than the battery found on a
typical gasoline vehicle, which allows these vehicles to store and strategically apply
electrical energy to supplement the gasoline engine. The result is that the engine can be
smaller than what would be needed in a non-hybrid vehicle, and the engine can be
operated near its peak efficiency more often. Hybrids also frequently utilize regenerative
braking, which uses a motor/generator to capture energy from braking instead of losing
that energy to friction and heat, as in traditional friction braking, and stop/start technology
to turn off the engine at idle. The combination of these strategies can result in significant
reductions in fuel use and CO2 emissions.
Hybrids were first introduced in the U.S. marketplace in model year 2000 with the Honda
Insight. As more models and options were introduced, hybrid production increased to 3.8%
of all vehicles in model year 2010, before declining somewhat over the next several years.
However, in model year 2020 hybrid production reached a new high at 4.9%, and is
projected to reach 8.9% in model year 2021, as shown in Figure 4.12 and Figure 4.13.
51
-------
Figure 4.10. Non-Hybrid Stop/Start Production Share by Vehicle Type
50%
40%
30%
g 20%
o
10%
0%
Vehicle Type
Sedan/Wagon
Car SUV
Truck SUV
Minivan/Van
I Pickup
2010
2015
2020
Model Year
Figure 4.11. Non-Hybrid Stop/Start Production Share by Number of Cylinders
50%
40%
-------
Figure 4.12. Gasoline Hybrid Engine Production Share by Vehicle Type
10%
9%
8%
7%
6%
5%
4%
3%
2%
1%
0%
Vehicle Type
Sedan/Wagon
Car SUV
Truck SUV
Minivan/Van
| Pickup
- ¦¦
I
2000
2005
2010
Model Year
2015
2020
Figure 4.13. Gasoline Hybrid Engine Production Share by Number of Cylinders

10%

7%
CD
6%
CO

¦C

W
5%
c

"S
4%
¦o

CL
3%

0%
Cylinders
4 Cylinder
| 6 Cylinder
8 Cylinder
I Other
III
2000
2005
2010
Model Year
2015
2020
53
-------
The growth in hybrid vehicles is largely attributable to growth outside of the sedan/wagon
vehicle type. In model year 2020 the production of hybrids in the truck SUV category
surpassed the production of sedan/wagon hybrids for the first time, and did so by more
than 50%. Sedan/wagon hybrids accounted for only about a third of all hybrid production
in model year 2020, as shown in Figure 4.12. The type of engine used in hybrid vehicles is
also changing, as an increasing number of 6 and 8-cylinder engines are being used in
hybrid system, as shown in and Figure 4.13.
The growth of hybrids in the pickup vehicle type is largely due to the introduction of "mild"
hybrid systems that are capable of regenerative braking and many of the same functions
as other hybrids but utilize a smaller battery and an electrical motor that cannot directly
drive the vehicle. These mild hybrids account for about a third of hybrid production in
model year 2020.
Plug-In Hybrid Electric, Electric, and Fuel Cell Vehicles
PHEVs and EVs are two types of vehicles that can store electricity from an external source
onboard the vehicle, utilizing that stored energy to propel the vehicle. PHEVs are similar to
gasoline hybrids discussed previously, but the battery packs in PHEVs can be charged from
an external electricity source; this cannot be done in gasoline hybrids. EVs operate using
only energy stored in a battery from external charging. Fuel cell vehicles use a fuel cell to
chemically convert a fuel (usually hydrogen) into electrical energy that is then used to
power the vehicle.
EVs rely on electricity stored in a battery for fuel. There is no combustion occuring on the
vehicle, and therefore there are no tailpipe emissions created by the vehicle. The electricity
used to charge EVs can create emissions at the power plant. The amount of emission varies
depending on the fuel source of the electricity, which can in turn vary based on location
and time of day. The electric grid in the US has also been changing over time, as natural gas
and renewable energy resources make up a growing portion of electricity generation
across the US. Depending on the source of electricity, EVs often result in lower CO2
emissions over their lifetime compared to gasoline vehicles.
Since EVs do not use gasoline, the familiar metric of miles per gallon cannot be applied to
EVs. Instead, EVs are rated in terms of miles per gallon-equivalent (mpge), which is the
number of miles that an EV travels on an amount of electrical energy equivalent to the
energy in a gallon of gasoline. This metric enables a direct comparison of energy efficiency
between EVs and gasoline vehicles. EVs generally have a much higher energy efficiency
54
-------
than gasoline vehicles because electric motors are much more efficient than gasoline
engines.
PHEVs can operate either on electricity stored in a battery, or gasoline, allowing for a wide
range of engine designs and strategies for the utilization of stored electrical energy during
typical driving. Most PHEVs will operate on electricity only, like an EV, for a limited range,
and then will operate like a more conventional hybrid until their battery is recharged from
an external source. The use of electricity to provide some or all of the energy required for
propulsion can significantly lower fuel consumption and tailpipe CO2 emissions. For a much
more detailed discussion of EV and PHEV metrics, as well as upstream emissions from
electricity, see Appendix E.
The production of EVs and PHEVs has increased rapidly in recent years. Prior to model year
2011, EVs were available, but generally only in small numbers for lease in California.11 In
model year 2011 the first PHEV, the Chevrolet Volt, was introduced along with the Nissan
Leaf EV. Many additional models have been introduced since, and in model year 2020
combined EV/PHEV production reached 2.2% of all new vehicles. Combined EV and PHEV
production is projected to reach a new high of 4.2% of all production in model year 2021.
The trend in EVs, PHEVs, and FCVs are shown in Figure 4.14.
11 At least over the timeframe covered by this report. EVs were initially produced more than 100 years ago.
55
-------
Figure 4.14. Production Share of EVs, PHEVs, and FCVs12
4% -
3% -
CD
ro
.c
CO
c
o
'¦g 2% -
=3
"O
0
1
Q_
1%-
0%-
The inclusion of model year 2020 EV and PHEV sales reduces the overall new vehicle
average CO2 emissions by 5 g/mi, and this impact will continue to grow if EV and PHEV
production increases. In model year 2020 there were three hydrogen FCVs produced, but
they were only available in the state of California and Hawaii and in very small numbers.
However there continues to be interest in FCVs as a future technology. Figure 4.15 and
Figure 4.16 show the production share by vehicle type for EVs and PHEVs. EV production is
mostly sedan/wagons and Car SUVs, while PHEVs have a much higher percentage of
production going to truck SUVs, and some minivan/vans.

Plug-In Hybrid EV

| Electric Vehicle
Fuel Cell Vehicle

.1
.
1
II

1995 2000 2005 2010 2015 2020
Model Year
12 EV production data were supplemented with data from Ward's and other publicly available production data
for model years prior to 2011. The data only include offerings from original equipment manufacturers and does
not include data on vehicles converted to alternative fuels in the aftermarket.
56
-------
Figure 4.15. Electric Vehicle Production Share by Vehicle Type
3%
e
j= 2%
w
c
O
o
VehicleType
Car
Car SUV
Truck SUV
¦a
o
1%
0%
2010 2015 2020
Model Year
Figure 4.16. Plug-In Hybrid Vehicle Production Share by Vehicle Type
CD
CD
-C
(D
c
o
'¦4—'
o
"O
o
3%
2%
1%
0%
VehicleType
Car
Car SUV
Truck SUV
Minivan/Van
2010
2015
2020
Model Year
57
-------
Figure 4.17 shows the range and fuel economy trends for EVs and PHEVs13. 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 almost 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. The range values shown for PHEVs are the charge-depleting range, where
the vehicle is operating on energy in the battery from an external source. This is generally
the electric range of the PHEV, although some vehicles also use the gasoline engine in small
amounts during charge depleting operation. The average charge depleting range for PHEVs
has remained largely unchanged since model year 2011.
Along with improving range, the fuel economy of electric vehicles has also improved as
measured in miles per gallon of gasoline equivalent (mpge). The fuel economy of electric
vehicles has increased by almost 20% between model years 2011 and 2020. The combined
fuel economy of PHEVs has been more variable but is about 15% lower in model year 2020
than in model year 2011. This may be attributable to the growth of truck SUV PHEVs, as
shown in Figure 4.16. For more information about EV and PHEV metrics, see Appendix E.
Figure 4.17. Charge Depleting Range and Fuel Economy for EVs and PHEVs
Range (mi) Fuel Economy (mpge)
Model Year
13 The range and fuel economy values in this figure are the combined values from the fuel economy label, which
weights city and highway driving 55% and 45%, as compared to the rest of the report, which uses a 43% city and
57% highway weighting. See Appendix C for more information.
58
-------
* -vsel Engines
Vehicles with diesel engines have been available in the U.S. at least as long as EPA has been
collecting data. However, sales of diesel vehicles have rarely broken more than 1 % of the
overall market. Diesel vehicle sales peaked at 5.9% of the market in model year 1981, but
quickly fell back to below 1 % of production per year. While the overall percentage of diesel
vehicles is low, there are still new vehicles entering the market.
Vehicles that rely on diesel fuel often achieve higher fuel economy than gasoline vehicles,
largely because the energy density of diesel fuel is about 15% higher than that of gasoline.
However, there is less of an advantage in terms of CO2 emissions because diesel fuel also
contains about 15% more carbon per gallon, and thus emits more CO2 per gallon burned
than gasoline.
Figure 4.18 shows the production share of diesel engines by vehicle type. Diesel engines
have historically been more prevalent in the sedan/wagon vehicle type, however, since
model year 2015 there have been very few sedan/wagons vehicles with diesel engines.
Light-duty diesel pickup trucks re-entered the market at about the same time and led to an
overall production share of 0.5% in model year 2020. This report does not include the
largest pickup trucks and work or vocational trucks, which have a higher penetration of
diesel engines. As shown in Figure 4.19, current production of diesel engines for light-duty
vehicles is almost exclusively six-cylinder engines.
Diesel engines, as with gasoline engines, have improved over time. Figure 4.20 shows the
same metrics and trends that are explored in Figure 4.5 for gasoline engines. The specific
power (HP/displacement) for diesel engines has increased more than 200% since model
year 1975. Fuel consumption per displacement dropped slightly in the 1980s but has
increased back to about the same level as in model year 1975. Finally, fuel consumption
per horsepower for diesel engines has declined about 70% since model year 1975.
59
-------
Figure 4.18. Diesel Engine Production Share by Vehicle Type
.¦¦¦¦llall-a
Vehicle Type
Sedan/Wagon
¦ Car SUV
Truck SUV
Minivan/Van
I Pickup
lbi.ll
~i 1 1 1 1 1 1 1 1—
1975 1980 1985 1990 1995 2000 2005 2010 2015 2020
Model Year
Figure 4.19. Diesel Engine Production Share by Number of Cylinders
6%
£ 4%
03
_C
co
c
o
H—'
o
"D
O
ft 2%H
0%
,ii
Cylinders
4 Cylinder
16 Cylinder
8 Cylinder
Other
IL
.¦¦¦HiiILi.II
—i—i—i—i—i—i 1 1—i—i—
1975 1980 1985 1990 1995 2000 2005 2010 2015 2020
Model Year
60
-------
Figure 4.20. Percent Change for Specific Diesel Engine Metrics
250%
200%
150%
LO
h-
CD
g 100% -
c
CO
CD
g> 50% -
03
-C
O
0% -
-50%
Other Engine Technologies
In addition to the engine technologies described above, there have been a small number of
other technologies available in the U.S. marketplace over the years. Vehicles that operate
on compressed natural gas (CNG) are one example, but there are currently no CNG
vehicles available from vehicle manufacturers (aftermarket conversions are not included
here). This report will continue to track all vehicles produced for sale in the U.S., and if CNG
or other technologies reach widespread availability they will be included in future versions
of this report.

HP/Displacement I
•

Fuel
Consumption/Displacement

Fi if*
P.nnQi imntinn/l-

•
1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025
Model Year
61
-------
t. Vehicle Drivetraie
A vehicle drivetrain includes all components responsible for transmitting rotational energy
from an engine or motor to the wheels. The design of the drivetrain impacts CO2 emissions
and fuel economy in two ways; first through direct energy losses or inefficiencies within the
drivetrain, and second by allowing a vehicle's engine, or motor, to operate in a more
efficient manner.
For non-hybrid vehicles with an internal combustion engine, the drivetrain includes a
transmission and the driveline (a driveshaft, differential, axle shafts and related
components), as shown in Figure 4.1. Mild hybrids generally use a conventional
transmission and drivetrain, but full hybrids often replace the transmission entirely with a
planetary gearset or some other configuration. PHEVs generally resemble full hybrids, but
can have numerous configurations that allow for complicated energy optimization. Electric
vehicles generally use a single speed transmission, and do not need the numerous gears
required by combustion engines. However, some electric vehicles are now being produced
with at least a 2-speed transmission (e.g. Porsche Taycan).
Trans missions
There are two important aspects of transmissions that impact overall vehicle efficiency and
fuel economy. First, as torque (rotational force) is transferred through the transmission, a
small amount is lost to friction, which reduces vehicle efficiency. Second, the design of the
transmission impacts how the engine is operated, and generally transmissions with more
speeds offer more opportunity to operate the engine in the most efficient way possible. For
example, a vehicle with an eight-speed transmission will have more flexibility in
determining engine operation than a vehicle with a five-speed transmission. This can lead
to reduced fuel consumption and CO2 emissions compared to a vehicle that is identical
except for the number of transmission gears.
Transmission designs have been rapidly evolving to increase the number of gears available
and allow for both better engine operation and improved efficiency. The number of gears
in new vehicles continues to increase, as does the use of continuously variable
transmissions (CVTs). Figure 4.21 shows the evolution of transmission production share for
cars and trucks since model year 1980.14 For this analysis, transmissions are separated into
manual transmissions, CVTs, and automatic transmissions. Automatic transmissions are
further separated into those with and without lockup mechanisms, which can lock up the
14 EPA has incomplete transmission data prior to model year 1980.
62
-------
torque converter in an automatic transmission under certain driving conditions and
improve efficiency. CVTs have also been split into hybrid and non-hybrid versions to reflect
the fact that hybrid CVTs are generally very different mechanically from traditional CVTs.
Dual clutch transmissions (DCTs) are essentially automatic transmissions that operate
internally much more like traditional manual transmissions. The two main advantages of
DCTs are that they can shift very quickly, and they can avoid some of the internal resistance
of a traditional automatic transmission by eliminating the torque converter. Currently,
automaker submissions to EPA do not explicitly identify DCTs as a separate transmission
category. Thus, the introduction of DCTs shows up in Figure 4.21 as a slight increase in
automatic transmissions without torque converters (although some DCTs may still be
reported as traditional automatic transmissions).
In the early 1980s, three-speed automatic transmissions, both with and without lockup
torque converters (shown as L3 and A3), were the most popular transmissions, but by
model year 1985, the four-speed automatic transmission with lockup (L4) became the most
popular transmission, a position it would hold for 25 years. Over 80% of all new vehicles
produced in model year 1999 were equipped with a four-speed transmission. After model
year 1999, the production share of four-speed transmissions slowly decreased as five and
six speed transmissions were introduced into the market. Six-speed transmissions peaked
in model year 2013 at 60% of new vehicle production, but then fell quickly, down to 17% in
model year 2020. Eight-speed transmissions became the most popular transmission in
model year 2019 and maintained that position for model year 2020, but CVTs and
transmissions with 9 or more speeds captured only slightly less share of production. These
trends are projected to continue in model year 2021, with CVTs (due to increasing hybrid
production), and transmissions with 9 or more speeds continuing to increase market share.
Another notable trend in Figure 4.21 is the decline in manual transmissions. Manual
transmissions were included in almost 35% of new vehicles in model year 1980, but have
been gradually declining since, and shrank to only 1.1% of all production in model year
2020. Today, manual transmissions are available only in a limited number of small vehicles,
sports cars, and a few pickups.
63
-------
Figure 4.21. Transmission Production Share
100%
75%
50%
<1)
&_
05
-C
CD
c
o
o
~C5
o
1980 1985 1990 1995 2000 2005 2010 2015 2020 2025
Model Year
Transmission
Lockup?
N umber of Gears
Key
Automatic
No
3
A3
Semi-Automatic

4
A4
Automated Manual

5
A5*

6
A6

7
A7

8
A8*

Yes
2
L2*

3
L3

4
L4

5
L5

6
L6

7
L7

8
L8

9
L9

10
L10
Manual
-
3
M3

4
M4

5
M5

6
M6

7
M7*
ContinuouslyVariable
(non-hybrid)
—
—
CVT(n-h)
ContinuouslyVariable
(hybrid)

CVT(h)
Other

-
Other
'Categories A5, A8, L2, and M7 are too small to depict in the area plot.
64
-------
Part of the reason for the decline in manual transmission is because modern automatic
transmissions now have more gears, are generally more efficient, and can offer better
performance than manual transmissions. Figure 4.22 shows the average number of gears
in new vehicle transmissions since model year 1980 for automatic and manual
transmissions. While both manual and automatic transmissions have added gears over
time, automatic transmission have added additional gears faster, and passed manual
transmissions in model year 2012. In model year 2020, the average number of gears for all
manual transmissions was 6 while the average for automatic transmissions was 7.8 gears.
Figure 4.22. Average Number of Transmission Gears

•

Manual

Automatic

1980 1985 1990 1995 2000 2005 2010 2015 2020 2025
Model Year
In the past, automatic transmissions have generally been less efficient than manual
transmissions, largely due to inefficiencies in the automatic transmission torque converter.
Figure 4.23 examines this trend over time by comparing the fuel economy of automatic and
manual transmission options where both transmissions were available in one model with
the same engine. Vehicles with a manual transmission were more efficient than their
automatic counterparts through about 2010, but modern automatic transmissions are now
more efficient. Two contributing factors to this trend are that automatic transmission
design has become more efficient (using earlier lockup and other strategies), and the
number of gears used in automatic transmissions has increased faster than in manual
65
-------
transmissions. The shrinking availability of manual transmissions does limit the relevance
of analyses comparing current manual transmissions to automatic transmissions.
Figure 4.23. Comparison of Manual and Automatic Transmission Real-World
Fuel Economy for Comparable Vehicles
1.05
1.00
0.95
0.90
Automatic transmissions
are more efficient
Manual transmissions
are more efficient
1980
1990
2000
Model Year
2010
2020
Drive Types
There has been a long and steady trend in new vehicle drive type away from rear-wheel
drive vehicles towards front-wheel drive and four-wheel drive (including all-wheel drive)
vehicles, as shown in Figure 4.24. In model year 1975, over 91 % of new vehicles were
produced with rear-wheel drive. Since then, production of rear-wheel drive vehicles has
steadily declined to less than 10% in model year 2020. Current production of rear-wheel
drive vehicles is mostly limited to pickup trucks and some performance vehicles.
Production of front-wheel drive vehicles increased from 5% of new vehicle production in
model year 1975 to 64% in model year 1990 and 63% in model year 2009. Since 2009
however, the production of front-wheel vehicles has also been declining, down to 41 % in
model year 2020. Four-wheel drive systems have steadily increased from 3.3% of new
vehicle production in model year 1975 to 50% of production in model year 2020.
66
-------
Figure 4.24. Front-, Rear-, and Four-Wheel Drive Production Share
100%
75%
05
-C
cn
o 50%
o
"a
o
CL
25%
0%
1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025
Model Year
D. Technology Adoption
One additional way to evaluate the evolution of technology in the automotive industry is to
focus on how technology has been adopted over time. Understanding how the industry has
adopted technology can lead to a better understanding of past changes in the industry,
and how emerging technology may be integrated in the future. The following analysis
provides more details about how manufacturers and the overall industry have adopted
new technology.
Industry-Wide Technology Adoption Since 1975
Figure 4.25 shows industry-wide adoption rates for seven technologies in passenger cars.
These technologies are fuel injection (including throttle body, port, and direct injection),
front-wheel drive, multi-valve engines (i.e., engines with more than two valves per cylinder),
engines with variable valve timing, lockup transmissions, advanced transmissions
(transmissions with six or more speeds, and CVTs), and gasoline direct injection engines. To
provide a common scale, the adoption rates are plotted in terms of the number of years
after the technology achieved first significant use in the industry. First significant use
67
-------
generally represents a production threshold of 1 %, though in some cases, where full data
are not available, first significant use represents a slightly higher production share.
The technology adoption pattern shown in Figure 4.25 is roughly similar for each of the
seven technologies, even though they vary widely in application, complexity, and when they
were initially introduced. It has taken, on average, approximately 15-20 years for new
technologies to reach maximum penetration across the industry. GDI is a newer technology
that has likely not reached maximum penetration across the industry but appears to be
following the adoption trend of other more mature technologies. While some of these
technologies may eventually be adopted in 100% of new vehicles, there may be reasons
that other technologies, like front-wheel drive, will likely never be adopted in all vehicles.
Adoption rates for these technologies in trucks are similar, with the exception of front-
wheel drive.
The analysis for Figure 4.25 focuses on technologies that have achieved widespread use by
multiple manufacturers and does not look at narrowly-adopted technologies which never
achieved widespread use. One limitation to the data in this report is that EPA does not
begin tracking technology production share data until after the technologies had achieved
some limited market share. For example, EPA did not begin to track multi-valve engine data
until model year 1986 for cars and model year 1994 for trucks, and in both cases multi-
valve engines had captured about 5% market share by that time. Likewise, turbochargers
were not tracked in this report until model year 1996 for cars and model year 2003 for
trucks, and while turbochargers had less than a 1% market share in both cases at that time,
it is likely that turbochargers had exceeded 1 % market share in the late 1980s. Cylinder
deactivation was utilized by at least one major manufacturer in the 1980s.
68
-------
Figure 4.25. Industry-Wide Car Technology Penetration after First Significant
Use
100%-
80% -
CD
s_
03
-C
w 60% -
c
O
o
o 40% -
CL
20% -
0% -
Technology Adoption by Manufacturers
The rate at which the overall industry adopts technology is determined by how quickly, and
at what point in time, individual manufacturers adopt the technology. While it is important
to understand the industry-wide adoption rates over time, the trends in Figure 4.25 mask
the fact that not all manufacturers introduced these technologies at the same time, or at
the same rate. The "sequencing" of manufacturers introducing new technologies is an
important aspect of understanding the overall industry trend of technology adoption.
Figure 4.26 begins to disaggregate the industry-wide trends to examine how individual
manufacturers have adopted new technologies.15 For each technology, Figure 4.26 shows
the amount of time it took specific manufacturers to move from initial introduction to 80%
penetration for each technology, as well as the same data for the overall industry. After
15 This figure is based on available data. Some technologies may have been introduced into the market before
this report began tracking them. Generally, these omissions are limited, with the exception of multi-valve
engine data for Honda. Honda had already achieved 70% penetration of multi-valve engines when this report
began tracking them in 1986, so this figure does not illustrate Honda's prior trends.

Fuel
Injection

Advanced /7
Transmission / / L

\ Lockup

fj GDI ]/
Multi-Valve

Front-Wheel

Drive

/ Variable-Valve
/ Timing

0 10 20 30 40 50
Years after First Significant Use
69
-------
80% penetration, the technology is assumed to be largely incorporated into the
manufacturer's fleet, and changes between 80% and 100% are not highlighted.
Of the seven technologies shown in Figure 4.26, five are now at or near full market
penetration for the included manufacturers, and two are still in the process of adoption by
manufacturers. The technologies shown in Figure 4.26 vary widely in terms of complexity,
application, and when they were introduced into the market. For each technology, there
are clearly variations between manufacturers, both in terms of when they began to adopt a
technology, and the rate with which they adopted the technology. The degree of variation
between the manufacturers also varies by technology.
The data for WT, for example, show that several manufacturers adopted the technology
much faster than the overall industry, which achieved 80% penetration in just over 20
years. It was not the rate of technology adoption alone, but rather the staggered
implementation timeframes among manufacturers that resulted in the longer industry-
wide average.
Fuel injection systems show the least amount of variation in initial adoption timing
between manufacturers, which resulted in a faster adoption by the industry overall than
technologies like WT. One important driver for adoption of fuel injection was increasingly
stringent emissions standards. Advanced transmissions, which have been available in small
numbers for some time, have very rapidly increased market penetration in recent years
and are now widely adopted. GDI engines appear to be following a similar path of quick
uptake in recent years. Turbocharged engines have long been available, but the focus on
turbo downsized engine packages is leading to much higher market penetration, although
it is too early to tell what level of penetration they will ultimately achieve industry-wide.
There are many factors outside the scope of this report that influence the rate and timing
of when technology is adopted by individual manufacturers (e.g., price, manufacturing
constraints, regulatory drivers, etc.) While no attempt is made here to identify the
underlying causes, it is important to recognize that variation between manufacturers for
given technologies can be masked when only the industry-wide trends are evaluated.
Technology adoption by individual manufacturers is often more rapid than the overall
industry trend would suggest.
70
-------
Figure 4.26. Manufacturer Specific Technology Adoption over Time for Key
Technologies
(D
!—
=3
O
&
=3
C
03
GM-
Toyota ¦
Ford ¦
Stellantis-
Honda-
Nissan ¦
Hyundai ¦
Ail Manufacturers-
GM-
Toyota ¦
Ford ¦
Stellantis ¦
Honda-
Nissan ¦
Hyundai ¦
Ail Manufacturers-
GM-
Toyota ¦
Ford ¦
Stellantis-
Honda-
Nissan ¦
Hyundai ¦
Ail Manufacturers-
GM-
Toyota ¦
Ford ¦
Stellantis ¦
Honda-
Nissan ¦
Hyundai ¦
Ail Manufacturers-
GM-
Toyota ¦
Ford ¦
Stellantis-
Honda-
Nissan ¦
Hyundai ¦
Ail Manufacturers-
GM-
Toyota ¦
Ford ¦
Stellantis-
Honda-
Nissan ¦
Hyundai ¦
Ail Manufacturers-
GM-
Toyota ¦
Ford ¦
Stellantis-
Honda-
Nissan ¦
Hyundai ¦
Ail Manufacturers-
Fuel Injection
1980
1990
2000
2010
2020
Lockup
1980
1990
2000
2010
2020
Multi-Valve
1980
1990
2000
2010
2020
Variable Valve
Timing
1980
1990
2000
2010
2020
1980
1990
2000
2010
Advanced
Transmissions
2020
1980
1990
2000
2010
Gasoline Direct
Injection
Turbocharged
1980 1990 2000
Model Year
2010
2020

Percent of Production

20% to 25%

10% to 15%

0% to 5% I
I I
25% to 50% 75% to 80%

I 15% to 20% 50% to 75%
80% to 100%
5% to 10%

71
-------
Table 4.1. Production Sha
Powertrain
Gasoline
Model Year
Gasoline
Hybrid
Diesel
EV
PHEV
FCV
Other

99.8%
-
0.2%
-
-
-
-
1980
95.7%
-
4.3%
-
-
-
-
1985
99.1%
-
0.9%
-
-
-
-
1990
99.9%
-
0.1%
-
-
-
-
1995
100.0%
-
0.0%
-
-
-
-
2000
99.8%
0.0%
0.1%
-
-
-
-
2001
99.7%
0.1%
0.1%
-
-
-
-
2002
99.6%
0.2%
0.2%
-
-
-
-
2003
99.5%
0.3%
0.2%
-
-
-
-
2004
99.4%
0.5%
0.1%
-
-
-
-
2005
98.6%
1.1%
0.3%
-
-
-
-
2006
98.1%
1.5%
0.4%
-
-
-
-
2007
97.7%
2.2%
0.1%
-
-
-
-
2008
97.4%
2.5%
0.1%
-
-
-
-
2009
97.2%
2.3%
0.5%
-
-
-
-
2010
95.5%
3.8%
0.7%
-
-
0.0%
-
2011
97.0%
2.2%
0.8%
0.1%
0.0%
0.0%
0.0%
2012
95.5%
3.1%
0.9%
0.1%
0.3%
0.0%
0.0%
2013
94.8%
3.6%
0.9%
0.3%
0.4%
-
0.0%
2014
95.7%
2.6%
1.0%
0.3%
0.4%
0.0%
0.0%
2015
95.9%
2.4%
0.9%
0.5%
0.3%
0.0%
0.0%
2016
96.9%
1.8%
0.5%
0.5%
0.3%
0.0%
0.0%
2017
96.1%
2.3%
0.3%
0.6%
0.8%
0.0%
-
2018
95.1%
2.3%
0.4%
1.4%
0.8%
0.0%
-
2019
94.4%
3.8%
0.1%
1.2%
0.5%
0.0%
-
2020
92.4%
4.9%
0.5%
1.8%
0.5%
0.0%
-
2021 (prelim)
85.7%
8.9%
1.2%
3.1%
1.0%
0.0%
-
To explore this data in more depth, please see the report website at https://www.epa.gov/automotive-trends.
-------
Table 4,2, Production Share by Engine Technologies
Fuel Delivery Method Avg. No. Non-hybrid
of Multi- Stop/
Model Year
Carb
TBI
Port
GDI
Other Cylinders
CID
HP
Valve
VVT
CD
Turbo
Start
1975
95.7%
0.0%
4.1%
-
0.2%
6.8
293
137
-
-
-
-
-
1980
89.7%
0.8%
5.2%
-
4.3%
5.6
198
104
-
-
-
-
-
1985
56.1%
24.8%
18.2%
-
0.9%
5.5
189
114
-
-
-
-
-
1990
2.1%
27.0%
70.8%
-
0.1%
5.4
185
135
23.1%
-
-
-
-
1995
-
8.4%
91.6%
-
0.0%
5.6
196
158
35.6%
-
-
-
-
2000
-
0.0%
99.8%
-
0.1%
5.7
200
181
44.8%
15.0%
-
1.3%
-
2001
-
-
99.9%
-
0.1%
5.8
201
187
49.0%
19.6%
-
2.0%
-
2002
-
-
99.8%
-
0.2%
5.8
203
195
53.3%
25.3%
-
2.2%
-
2003
-
-
99.8%
-
0.2%
5.8
204
199
55.5%
30.6%
-
1.2%
-
2004
-
-
99.9%
-
0.1%
5.9
212
211
62.3%
38.5%
-
2.3%
-
2005
-
-
99.7%
-
0.3%
5.8
205
209
65.6%
45.8%
0.8%
1.7%
-
2006
-
-
99.6%
-
0.4%
5.7
204
213
71.7%
55.4%
3.6%
2.1%
-
2007
-
-
99.8%
-
0.1%
5.6
203
217
71.7%
57.3%
7.3%
2.5%
-
2008
-
-
97.6%
2.3%
0.1%
5.6
199
219
76.4%
58.2%
6.7%
3.0%
-
2009
-
-
95.2%
4.2%
0.5%
5.2
183
208
83.8%
71.5%
7.3%
3.3%
-
2010
-
-
91.0%
8.3%
0.7%
5.3
188
214
85.5%
83.8%
6.4%
3.3%
-
2011
-
-
83.8%
15.4%
0.8%
5.4
192
230
86.4%
93.1%
9.5%
6.8%
-
2012
-
-
76.5%
22.5%
1.0%
5.1
181
222
91.8%
96.6%
8.1%
8.4%
0.6%
2013
-
-
68.3%
30.5%
1.2%
5.1
176
226
92.8%
97.4%
7.7%
13.9%
2.3%
2014
-
-
61.3%
37.4%
1.3%
5.1
180
230
89.2%
97.6%
10.6%
14.8%
5.1%
2015
-
-
56.7%
41.9%
1.4%
5.0
177
229
91.2%
97.2%
10.5%
15.7%
7.1%
2016
-
-
51.0%
48.0%
1.0%
5.0
174
230
92.3%
98.0%
10.4%
19.9%
9.6%
2017
-
-
49.4%
49.7%
0.9%
5.0
174
234
92.0%
98.1%
11.9%
23.4%
17.8%
2018
-
-
48.0%
50.2%
1.8%
5.0
172
241
91.0%
96.4%
12.5%
30.0%
29.8%
2019
-
-
45.7%
52.9%
1.4%
5.1
174
245
90.1%
97.2%
14.9%
30.0%
36.9%
2020
-
-
40.6%
57.1%
2.2%
5.0
170
246
90.7%
95.8%
14.7%
34.7%
45.8%
2021 (prelim)
-
-
40.0%
55.6%
4.4%
5.1
176
252
87.6%
93.9%
17.2%
34.4%
45.4%
-------
Table 43.
Production Share by Transmission Technologies
Model Year
1975
Manual
23.0%
Automatic
with
Lockup
0.2%
Automatic
without
Lockup
76.8%
CVT
(Hybrid)
CVT
(Non-
Hybrid)
Other
4 Gears
or
Fewer
99.0%
5
Gears
1.0%
6
Gears
7
Gears
8
Gears
9+
Gears
Average
No. of
Gears
1980
34.6%
18.1%
46.8%
-
-
0.5%
87.9%
12.1%
-
-
-
-
3.5
1985
26.5%
54.5%
19.1%
-
-
-
80.7%
19.3%
-
-
-
-
3.8
1990
22.2%
71.2%
6.5%
-
0.0%
0.0%
79.9%
20.0%
0.1%
-
-
-
4.0
1995
17.9%
80.7%
1.4%
-
-
-
82.0%
17.7%
0.2%
-
-
-
4.1
2000
9.7%
89.5%
0.7%
-
0.0%
-
83.7%
15.8%
0.5%
-
-
-
4.1
2001
9.0%
90.3%
0.6%
0.1%
0.0%
-
80.7%
18.5%
0.7%
-
-
-
4.2
2002
8.2%
91.4%
0.3%
0.1%
0.1%
-
77.1%
21.6%
1.1%
-
-
-
4.2
2003
8.0%
90.8%
0.1%
0.3%
0.8%
-
69.2%
28.1%
1.7%
-
-
-
4.3
2004
6.8%
91.8%
0.3%
0.4%
0.7%
-
63.9%
31.8%
3.0%
0.2%
-
-
4.4
2005
6.2%
91.5%
0.1%
1.0%
1.3%
-
56.0%
37.3%
4.1%
0.2%
-
-
4.5
2006
6.5%
90.6%
0.0%
1.5%
1.4%
-
47.7%
39.2%
8.8%
1.4%
-
-
4.6
2007
5.6%
87.1%
0.0%
2.1%
5.1%
-
40.5%
36.1%
14.4%
1.5%
0.2%
-
4.8
2008
5.2%
86.8%
0.2%
2.4%
5.5%
-
38.8%
31.9%
19.4%
1.8%
0.2%
-
4.8
2009
4.8%
85.6%
0.2%
2.1%
7.3%
-
31.2%
32.2%
24.5%
2.5%
0.1%
-
5.0
2010
3.8%
84.1%
1.2%
3.8%
7.2%
-
24.6%
23.5%
38.1%
2.7%
0.2%
-
5.2
2011
3.2%
86.5%
0.3%
2.0%
8.0%
-
14.2%
18.7%
52.3%
3.1%
1.7%
-
5.5
2012
3.6%
83.4%
1.1%
2.7%
9.2%
-
8.1%
18.2%
56.3%
2.8%
2.6%
-
5.5
2013
3.5%
80.4%
1.4%
2.9%
11.8%
-
5.4%
12.8%
60.1%
2.8%
4.1%
-
5.6
2014
2.8%
76.7%
1.6%
2.3%
16.6%
-
2.2%
7.8%
58.4%
3.3%
8.4%
1.1%
5.9
2015
2.6%
72.3%
1.4%
2.2%
21.5%
-
1.5%
4.5%
54.2%
3.1%
9.5%
3.5%
5.9
2016
2.2%
72.3%
2.6%
1.7%
21.2%
-
1.1%
3.0%
54.9%
2.9%
11.2%
4.1%
6.0
2017
2.1%
71.5%
2.6%
1.9%
21.8%
-
1.0%
2.4%
49.0%
3.4%
14.6%
5.9%
6.1
2018
1.6%
72.8%
3.2%
1.7%
20.6%
-
1.9%
2.0%
37.6%
3.7%
19.0%
13.5%
6.4
2019
1.4%
72.1%
2.4%
2.2%
21.9%
-
1.5%
1.6%
26.1%
2.6%
27.5%
16.5%
6.6
2020
1.1%
68.3%
2.7%
2.8%
25.0%
-
1.8%
0.8%
17.3%
2.1%
28.8%
21.2%
6.9
2021 (prelim)
1.0%
65.7%
5.1%
4.6%
23.7%
-
3.1%
1.5%
12.5%
2.0%
28.5%
24.2%
6.6
-------
Table 4.4. Production Share by Drive Technology
Car Truck
Model Year
Front
Wheel
Drive
Rear
Wheel
Drive
Four
Wheel
Drive
Front
Wheel
Drive
Rear
Wheel
Drive
1975
6.5%
93.5%
-
-
82.8%
1980
29.7%
69.4%
0.9%
1.4%
73.6%
1985
61.1%
36.8%
2.1%
7.3%
61.4%
1990
84.0%
15.0%
1.0%
15.8%
52.4%
1995
80.1%
18.8%
1.1%
18.4%
39.3%
2000
80.4%
17.7%
2.0%
20.0%
33.8%
2001
80.3%
16.7%
3.0%
16.3%
34.8%
2002
82.9%
13.5%
3.6%
15.4%
33.1%
2003
80.9%
15.9%
3.2%
15.4%
34.1%
2004
80.2%
14.5%
5.3%
12.5%
31.0%
2005
79.2%
14.2%
6.6%
20.1%
27.7%
2006
75.9%
18.0%
6.0%
18.9%
28.0%
2007
81.0%
13.4%
5.6%
16.1%
28.4%
2008
78.8%
14.1%
7.1%
18.4%
24.8%
2009
83.5%
10.2%
6.3%
21.0%
20.5%
2010
82.5%
11.2%
6.3%
20.9%
18.0%
2011
80.1%
11.3%
8.6%
17.7%
17.3%
2012
83.8%
8.8%
7.5%
20.9%
14.8%
2013
83.0%
9.3%
7.7%
18.1%
14.5%
2014
81.3%
10.6%
8.2%
17.5%
14.2%
2015
80.4%
9.7%
9.9%
16.0%
12.6%
2016
79.8%
9.1%
11.0%
15.9%
12.2%
2017
79.7%
8.3%
12.0%
16.1%
11.1%
2018
76.5%
9.4%
14.1%
13.4%
10.9%
2019
75.5%
10.1%
14.4%
14.4%
10.2%
2020
76.5%
8.8%
14.7%
12.5%
10.0%
2021 (prelim)
69.0%
11.8%
19.2%
10.0%
10.6%
All
Four
Wheel
Drive
17.2%
25.0%
31.3%
31.8%
42.3%
46.3%
48.8%
51.6%
50.4%
56.5%
52.2%
53.1%
55.5%
56.8%
58.5%
61.0%
65.0%
64.3%
67.5%
68.3%
71.4%
72.0%
72.8%
75.6%
75.4%
77.5%
79.4%
Front
Wheel
Drive
5.3%
25.0%
47.8%
63.8%
57.6%
55.5%
53.8%
52.7%
50.7%
47.7%
53.0%
51.9%
54.3%
54.2%
62.9%
59.6%
53.8%
61.4%
59.7%
55.3%
52.9%
51.2%
49.6%
43.7%
41.6%
40.7%
33.0%
Rear
Wheel
Drive
91.4%
70.1%
42.9%
26.1%
26.3%
24.3%
24.2%
22.3%
24.3%
22.4%
20.2%
22.3%
19.6%
18.5%
13.6%
13.7%
13.8%
10.9%
11.1%
12.1%
10.9%
10.5%
9.6%
10.2%
10.1%
9.4%
11.1%
Four
Wheel
Drive
3.3%
4.9%
9.3%
10.1%
16.2%
20.2%
22.0%
25.0%
25.0%
29.8%
26.8%
25.8%
26.1%
27.3%
23.5%
26.7%
32.4%
27.7%
29.1%
32.6%
36.1%
38.3%
40.8%
46.1%
48.3%
49.9%
56.0%
75
-------
5. Manufacturer GHG Compliance
Manufacturers that produce passenger cars, light-duty trucks, and medium-duty passenger
vehicles for sale in the United States are required to meet greenhouse gas (GHG) emissions
and fuel economy standards. The Environmental Protection Agency (EPA) regulates
greenhouse gas (GHG) emissions through the light-duty GHG program, and the National
Highway Traffic Safety Administration (NHTSA) regulates fuel economy through the
Corporate Average Fuel Economy (CAFE) program. The following analysis is designed to
provide as much information as possible about how manufacturers are performing under
EPA's GHG program, including final compliance data through model year 2020 and credit
trades reported to EPA as of October 31, 2021.
In August 2021, EPA and NHTSA proposed rules to revise the existing light-duty GHG and
fuel economy standards for model years 2023-2026 and 2024-2026, respectively. In
addition, EPA is reconsidering the withdrawal of California's waiver to enforce greenhouse
gas standards for cars and light trucks. Since these proposals have not been finalized, they
are not reflected in this report. Any applicable regulatory changes finalized by EPA and
NHTSA will be included in future versions of this report.
EPA's GHG program defines standards for each manufacturer's car and truck fleets based
on the average footprint of the vehicles produced for sale. Each manufacturer fleet
generates credits if the fleet average
emissions performance is below the
standards, or deficits if it is above the
standards. Credits, or deficits, that
manufacturers have accrued in previous
model years, credits earned as part of the
early credit program, credit trades, credit
forfeitures, and credit expirations are also
important components in determining the
final compliance status of each
manufacturer. Manufacturers that
maintain a positive, or zero, credit balance
are considered in compliance with the
GHG program. Manufacturers that end any
model year with a deficit have up to three
years to offset that deficit to avoid non-
Figure 5.1. The GHG Compliance Process
76
-------
compliance. The general compliance process that manufacturers follow at the end of each
model year is shown in Figure 5.1.
Averaging, banking, and trading (ABT) provisions have been an important part of many
mobile source programs under the Clean Air Act. These provisions help manufacturers in
planning and implementing a phase-in of emissions reduction technology in their
production that is consistent with their unique redesign schedules. As part of the GHG
program, ABT provisions allow manufacturers to average their car or truck fleet CO2
emissions (i.e., the standards do not apply to individual vehicles), to earn and "bank" credits
by reducing their car or truck fleet performance to below the applicable standards, and to
trade credits between manufacturers. EPA believes the net effect of the ABT provisions is
that they allow additional flexibility, encourage earlier introduction of emission reduction
technologies than might otherwise occur, and do so without reducing the overall
effectiveness of the program.
Manufacturer standards and model
year performance are discussed in
this report as per vehicle emission
rates, measured in grams of CO2 per
mile (g/mi). Any discussion of
manufacturer total credit balances,
credit transactions, and compliance
will be in terms of total mass of CO2
emissions, measured in Megagrams
of C02 (Mg). The use of a mass-based
metric enables the banking and
trading portions of the GHG program
by accounting for vehicle lifetime
emissions for all vehicles produced.
Converting from an emission rate to
total emissions is straightforward, as
shown in the box on the right.
Unlike the previous sections of this report, the tailpipe C02 emission data presented in this
section are compliance data, based on EPA's City and Highway test procedures (referred to
as the "2-cycle" tests). These values should not be compared to the estimated real-world
data throughout the rest of this report. For a detailed discussion of the difference between
real-world and compliance data, see Appendix C. To download the data presented in this
section please see the report website: https://www.epa.gov/automotive-trends.
How to Calculate Total Emissions
from an Emission Rate
Total emissions, or credits, are calculated by multiplying a
C02 emission rate, the production volume of applicable
vehicles, and the expected lifetime vehicle miles travelled
(VMT) of those vehicles. To calculate total emissions, or
credits, the following equation is used:
Credits = ( CO2 Emissions x VMT x Production ) /1,000,000
In the above equation, "Credits" are measured in megagrams
(Mg) of co2, "CO2 emissions" are measured in grams per
mile (g/mi), and "VMT" is in miles, and specified in the
regulations as 195,264 miles for cars and 225,865 for
trucks. To calculate g/mi from Mg:
CO2 Emissions = ( Credits x 1,000,000) / (VMT x Production )
When using these equations to calculate values for cars and
trucks in aggregate, use a production weighted average of
the car and truck VMT values. For the 2020 model year, the
industry wide weighted VMT is 212,417 miles.
77
-------
A. Footprint-Based CO2 Standards
At the end of each model year, manufacturers are required to calculate unique C02
standards for their car and truck fleets, based on the vehicles produced that model year.
The GHG program uses footprint, which is the area between the four tires, as a metric for
determining the specific standard for each manufacturer's car and truck fleets.
Manufacturers must calculate new standards each year as the regulations become more
stringent, and as their footprint distribution and production change. See Section 3 for a
discussion of footprint and vehicle production trends and the definitions of "car" and
"truck" under the regulations.
The regulations define footprint "curves" that provide a CO2 emissions target for every
vehicle footprint, as shown in Figure 5.2. For example, a car with a footprint of 46.6 square
feet in model year 2020 (the average car footprint) has a compliance CO2 target of 189
g/mi. This is a target and not a standard, as there are no footprint-based CO2 emissions
requirements for individual vehicles at the time of certification. The unique C02 standards
for each manufacturer's car and truck fleets are production-weighted averages of the CO2
target values, as determined from the curves, for all the unique footprint values of the
vehicles within that fleet. This is an element of the "averaging" approach of the ABT
provisions. Using one production-weighted average to define a single fleet standard allows
for some individual vehicles to be above that standard, while others are below.
Figure 5.2. 2012-2020 Model Year CO2 Footprint Target Curves
400
^ 350
E
3
CM
o
o 300
CD
O
c
as
Q.
o 250
O
200

2012 Truck

.X*

**
.X
X*
*
"X"
2020 Truck
2012 Car

¦1 ~ *
.X

2020 Car

40 50 60 70 80
Footprint (sq ft)
78
-------
The footprint curves for the 2012 and 2020 model years are shown in Figure 5.2. The
targets have gradually decreased (become more stringent) from 2012 to the current 2020
levels, as defined in the regulations. Larger vehicles have higher targets, although the
increases are capped beyond a certain footprint size (i.e., the curves become flat). Trucks
have higher targets than cars of the same footprint in the same model year.
In addition to the footprint-based standards, EPA established several alternative standards
for small to intermediate manufacturers. These provisions provide additional lead-time for
manufacturers that may not be able to take full advantage of averaging or other program
flexibilities due to the limited scope of the types of vehicles they sell.
The Temporary Lead-time Allowance Alternative Standards (TLAAS) provisions were
available to manufacturers with production of less than 400,000 vehicles in model year
2009. This provision allowed manufacturers to place vehicles in an alternative car or truck
TLAAS fleet each model year, with those vehicles subject to a less stringent standard. The
standard for a TLAAS fleet was 1.25 times the standard that would have applied to that
fleet based on the footprint-based approach applied to all other car and truck fleets. Each
manufacturer could apply the TLAAS standards to a maximum of 100,000 vehicles,
cumulative over model years 2012-2015. Mercedes, Jaguar Land Rover, Volvo, Porsche,
Ferrari, Aston Martin, Lotus, and McLaren participated in the TLAAS program. The overall
impact of the TLAAS program was less than 1 g/mi for all years it was available.
The intermediate volume provisions allowed intermediate volume manufacturers (those
that produced less than 50,000 vehicles in the 2009 model year) to use an alternative
compliance schedule in model years 2017-2020. Under these provisions, manufacturers
were required to meet the model year 2016 standards in the model years 2017 and 2018,
delay meeting the 2019-2020 standards by one model year, and finally align with the
primary standards and other manufacturers in the 2021 model year. Jaguar Land Rover
and Volvo are the two manufacturers utilizing these alternative compliance schedules.
Small volume manufacturers, with U.S. production of less than 5,000 vehicles per year,
have additional options under the GHG program. This includes the ability to petition EPA
for alternative standards for model year 2017 and later, and allowing these manufacturers
to meet an established alternative model year 2017 standard in model years 2015 and
2016. Aston Martin, Ferrari, Lotus, and McLaren applied for unique alternative standards
79
-------
for model years 2017-2021, and EPA established alternative standards for these
manufacturers in a July 2020 determination.16
Each manufacturer's standards for model year 2020 are shown in Table 5.1. In model year
2020, average car and truck footprints both increased slightly, with cars increasing about
0.2 square feet and trucks increasing by about 0.1 square feet. The more stringent model
year 2020 footprint targets, along with changes to footprint, resulted in a reduction of the
car standard by 9 g/mi, from 198 g/mi to 189 g/mi, and the truck standard by 7 g/mi, from
279 g/mi to 272 g/mi. While there is no combined car and truck standard for regulatory
purposes, this report will often calculate one to provide an overall view of the industry and
to allow comparison across manufacturers. Overall, the effective combined car and truck
standard decreased in model year 2020 by 7 g/mi, from 246 g/mi to 239 g/mi.
Table 5.1. Manufacturer Footprint and Standards for Model Year 2020

Footprint (ft2)

Standards (g/mi)
Manufacturer
Car
Truck
All
Car
Truck
All
Aston Martin
49.0
-
49.0
374
-
374
BMW
47.8
52.8
49.7
194
266
224
Ferrari
47.7
-
47.7
386
-
386
Ford
47.8
58.1
55.4
194
289
266
GM
45.5
58.9
54.5
186
290
259
Honda
46.1
49.8
47.7
187
252
217
Hyundai
46.5
53.5
47.2
188
269
198
Jaguar Land Rover
47.9
51.4
51.3
204
268
266
Kia
45.7
50.4
47.5
185
255
214
Mazda
45.5
47.1
46.3
186
240
215
McLaren
47.1
-
47.1
360
-
360
Mercedes
48.8
53.2
51.3
198
268
240
Mitsubishi
41.6
44.2
43.0
173
227
203
Nissan
46.7
51.1
48.1
189
258
213
Stellantis
49.3
55.8
54.9
200
279
268
Subaru
44.9
46.4
46.1
182
237
228
Tesla
50.4
54.8
50.6
204
275
206
Toyota
46.3
52.6
49.3
188
265
227
VW
46.3
49.4
48.1
188
250
226
Volvo
49.5
51.1
50.7
211
267
254
All Manufacturers
46.6
54.3
50.9
189
272
239
16 89 FR 39561, July 1, 2020.
80
-------
Since the beginning of the GHG program, two notable changes in manufacturer groupings
have occurred. Porsche was part of the program as an independent manufacturer for
model years 2012 and 2013, but has been included as part of Volkswagen for all following
model years. Beginning in model year 2020, Lotus and Volvo submitted data as one
manufacturer for compliance with the GHG program, since both companies are majority
owned by Zhejiang Geely Holding Group Co., Ltd (Geely). EPA determinations related to this
merger are subject to change and will be updated in future reports as necessary. Table 5.1
shows the manufacturers that produced vehicles in model year 2020 using current
manufacturer groupings, while later tables in this report show all manufacturers that were
regulated independently in any model year, to allow for complete credit accounting.
B. Model Year Performance
After determining car and truck fleet standards for the model year, manufacturers must
determine the performance value for their car and truck fleets. This is the average
production-weighted CO2 tailpipe emissions of each fleet, including the impact of several
optional performance credits and adjustments. These credits and adjustments allow
manufacturers to benefit from technologies that reduce emissions but are not wholly
captured in standard regulatory tests, provide incentives for manufacturers to adopt
advanced technologies, and provide flexibility in other areas of the program. The available
performance credits and adjustments include:
• Performance credits for producing alternative fuel vehicles
• Performance credits for improving air conditioning systems
• Performance credits for deploying "off-cycle" technologies that reduce emissions
but are not captured on EPA's regulatory test cycles
• Adjustments for utilizing alternate methane and nitrous oxide standards
The impact of these credits and adjustments are integral to the annual model year analysis.
Any performance credits generated must be included in the model year fleet calculations
before a manufacturer can bank or trade credits. In addition, the performance value,
including the impact of the performance credits and adjustments, is the most accurate way
to compare how manufacturers' car and truck fleets are performing in comparison to the
standards within a model year. The standards discussed previously were designed
assuming manufacturers would use these optional provisions; therefore, any comparison
that excludes them is incomplete. Manufacturer tailpipe emissions, and each of the
performance credits and adjustments, are examined in detail below.
81
-------
Tailpipe CO2 Emissions
The starting point for determining compliance for each manufacturer is its "2-cycle" tailpipe
GHG emissions value. All manufacturers are required to test their vehicles on the Federal
Test Procedure (known as the "City" test) and the Highway Fuel Economy Test (the
"Highway" test). Results from these two tests are combined by weighting the City test by
55% and the Highway test by 45%, to achieve a single combined CO2 value for each vehicle
model. Manufacturers then calculate a sales-weighted average of all the combined
city/highway values for each car and truck fleet. This represents the measured tailpipe CO2
emissions of a fleet without the application of any additional performance credits. As
discussed previously in this report, 2-cycle tailpipe C02 emissions should only be used in
the context of the compliance regulations and are not the same as and should not be
compared to the estimated real-world values reported in Sections 1-4.
As part of the GHG program, electric vehicles and fuel cell vehicles are included in the 2-
cycle tailpipe calculations with zero g/mi of tailpipe emissions. Plug-in hybrid vehicles
(PHEVs) are allowed to use a zero g/mi value for the portion of operation attributed to the
use of grid electricity (i.e., only emissions from the portion of operation attributed to the
gasoline engine are counted). Use of the zero g/mi option was limited to the first 200,000
qualified vehicles produced by a manufacturer in the 2012-2016 model years. No
manufacturer reached this limit. In the 2017-2026 model years, manufacturers may
continue to use zero g/mi for these vehicles, without any limits.
Figure 5.3 shows the 2-cycle tailpipe emissions reported by each manufacturer for the 2012
and 2020 model years, for all vehicles and for car and truck fleets. Companies that produce
solely electric vehicles (Tesla) are shown separately in the figure because they produce zero
tailpipe emissions on the 2-cycle tests. Figure 5.3 includes all manufacturers that reported
production in 2012 and 2020; there are additional manufacturers that produced vehicles in
that timespan that are not shown. The tailpipe values in Figure 5.3 should not be directly
compared to the manufacturer's standards presented in Table 5.1, as the standards were
created taking into consideration the optional performance credits and adjustments
available to manufacturers to reduce their performance values.
82
-------
Figure 5.3. Changes in "2-Cycle" Tailpipe CO2 Emissions by Manufacturer (g/mi)
Tesla <0
<0
100
Mitsubishi
Honda
Subaru
Hyundai
Mazda
Nissan
Kia
Toyota
BMW
VW
Volvo
Mercedes
Ford
GM
Jaguar Land Rover
Stellantis
Ferrari
All
Manufacturers
150
All
200
100
229<-2£
235<-2(
O/I
7
16
2!
9
13

1
244+24
247+21
248+—
255+2
258+:
273+
279'
282
3(
3(
3(

2
36
7:
_
K
<
1
3
14
1'

)

>
502
81
-311
<—343
<315
<-331

/IOC

32 <-357
4:

494

282
<
302

203
206
2i
23^
233
220
218-
203 <
2
23$
2
¦: !21
53
2
229
200
300
400
500
221
200
¦2:
5+
I63
150
Car
262
>7
257
>43
41
258
258
-277
¦274
—297
<—316
-261
—283
<
200
100
-376
B00- 307
>59
423 <-
-494
150
ruck
200
-28C
25II
250 «
269
240<-
312
260 <—
300-O
307
310
300<|
304-
289 <-
319
318
3:
316
33
300
400
500
200
311
300
320
6
318
¦324
-382
¦324
54
63
¦33q
-3
43
33
-393
-385
-397
-439
-384
369
400
500
83
-------
Every manufacturer that has been in the U.S. market since the GHG program was
implemented in 2012 has reduced fleetwide overall tailpipe GHG emissions, except for
those manufacturers that only produce electric vehicles. Compared to the first year of the
program, Jaguar Land Rover leads manufacturers in both the overall reduction in 2-cycle
CO2 emissions (111 g/mi) and the percentage reduction (26%). Seven manufacturers have
reduced tailpipe CO2 emissions by 10-16%, while the remainder have produced single digit
percentage reductions since the first year of the program. Overall, tailpipe CO2 emissions
of the entire fleet have been reduced by 27 g/mi, or about 9%, since the 2012 model year.
Compliance is assessed on a fleet-specific basis, and most manufacturers have reduced
emissions within their car and truck fleets, some considerably, leading to reductions of 38
and 58 g/mi in the car and truck fleets, respectively, since model year 2012. The overall
reduction in tailpipe C02 emissions is smaller than the reduction in either the car or truck
fleets because of the shifting fleet mix towards trucks.
Kn usance m t\M? nn uatix r hid Vdu» Irs
EPA's GHG program provides performance credits for dedicated and dual fuel alternative
fuel vehicles. Dedicated alternative fuel vehicles run exclusively on an alternative fuel while
dual fuel vehicles can run both on an alternative fuel and on conventional gasoline. This
section describes two pathways for manufacturers to benefit from the production of
alternative fuel vehicles. The first pathway is through a set of defined production
multipliers available for certain alternative fuel vehicles. The second pathway is based on
incentives for gasoline-ethanol flexible fuel vehicles (FFVs), which can run on E85 (85%
ethanol and 15% gasoline), or on conventional gasoline.
Performance Credits for Advanced Technology Vehicles
The GHG program created an incentive for advanced technology vehicles through the
introduction of vehicle "multipliers" for electric vehicles (EVs), plug-in hybrid electric
vehicles (PHEVs), fuel cell vehicles (FCVs), and compressed natural gas (CNG) vehicles.
Multipliers allow manufacturers to increase the amount of credits created by each vehicle
during the compliance process. For example, the 1.75 multiplier for 2020 model year EVs
allows a manufacturer to increase the credits created by each electric or fuel cell vehicle by
an additional 75%. The impact of the multipliers is calculated separately from the main car
or truck fleet of each manufacturer and is included in this report as an advanced
technology credit. The multipliers established by rulemaking are shown in Table 5.2.
84
-------
Table 5,2, Production Multipliers by Model Year
Model
Year
Electric Vehicles
and Fuel Cell Vehicles
Plug-In Hybrid Electric
Vehicles
Dedicated and Dual-
Fuel Natural Gas
Vehicles
2017
2018
2019
2020
2021
2.0
2.0
2.0
1.75
1.5
1.0
1.6
1.6
1.6
1.45
1.3
1.0
1.6
1.6
1.6
1.45
1.3
2.0
2022-2026
Figure 5.4 shows the model year 2020 production volume of vehicles qualifying for
multiplier incentives. More than 300,000 EVs, PHEVs, and FCVs were produced in the 2020
model year. Of those vehicles, about 78% were EVs, 22% were PHEVs, less than 1%were
FCVs. There were no CNG vehicles subject to the GHG standards in the 2020 model year,
and only a limited number of CNG vehicles in prior years. Figure 4.14 in the previous
section shows the overall growth in EVs, PHEVs, and FCVs.
The impacts of the advanced technology multiplier credit are shown in Figure 5.5. The
impact of this incentive is particularly evident for Tesla, because Tesla produces only
electric vehicles. Before including air condition and off-cycle credits, each Tesla vehicle on
average created 206 g/mi of credits, which is the difference between Tesla's standard and
their 0 g/mi tailpipe emissions. The 1.75 multiplier results in an additional 155 g/mi of
credit per vehicle, on average, for Tesla. Tesla is shown separately in Figure 5.5 due to the
scale of the credits generated by their vehicles.
After Tesla, Jaguar Land Rover had the highest g/mi effect on their fleet performance, at 3.9
g/mi. Of Jaguar Land Rover's model year 2020 production, 1.4% was EVs and another 1.2%
was PHEVs. VW had the third highest g/mi effect on their fleet performance, at 3.3 g/mi due
to 1.4% EVs and 1.3% PHEVs. Tesla, VW and Jaguar Land Rover had the highest percentage
of EVs, while Volvo. BMW, and Mitsubishi had the highest percentage of PHEVs.
85
-------
Figure 5.4. Model Year 2020 Production of EVs, PHEVs, and FCVs
Figure 5.5. Model Year 2020 Advanced Technology Credits by Manufacturer
-------
EPA finalized a technical amendment on March 31, 2020 that corrects the regulations
pertaining to how manufacturers calculate credits for the GHG program's advanced
technology incentives.17 Manufacturers that produced vehicles eligible for these incentives
have resubmitted 2-cycle data to EPA, and this report uses these updated data and
calculations.
Gasoliee-Etfaaeol Flexit ;1 Vehicles
For the 2012 to 2015 model years, FFVs could earn performance credits corresponding to
the fuel economy credits under CAFE. For both programs, it was assumed that FFVs
operated half of the time on each fuel. The GHG credits were based on the arithmetic
average of alternative fuel and conventional fuel CO2 emissions. Further, to fully align the
GHG credit with the CAFE program, the CO2 emissions measurement on the alternative fuel
was multiplied by 0.15. The 0.15 factor was used because, under the CAFE program's
implementing statutes, a gallon of alternative fuel is deemed to contain 0.15 gallons of
gasoline fuel, and the E85 fuel economy is divided by 0.15 before being averaged with the
gasoline fuel economy.
Starting in model year 2016, GHG compliance values for FFVs are based on the actual
emissions performance of the FFV on each fuel, weighted by EPA's assessment of the actual
use of these fuels in FFVs. In 2014, EPA issued a determination defining an "F factor" of 0.14
to use when weighting E85 and gasoline CO2 emissions for the 2016-2018 model years
FFVs; this reflects EPA's estimate that FFVs would be operating 14% of the time on E85.This
approach is comparable to the "utility factor" method used to weight gasoline and
electricity for PHEVs, which projects the percentage of miles that a PHEV will drive using
electricity based on how many miles a fully-charged PHEV can drive using grid electricity.
EPA also adopted an F-factor of 0.14 for model years 2019 and 2020, and in a separate
action has extended the use of 0.14 to model years 2021 and later.18 This value will
continue to apply until EPA issues a new determination.
FFVs can still represent a CO2 emissions benefit, and can help to lower the emissions of a
manufacturer's fleet, but the overall impact is significantly diminished. Because the FFV
values now incorporate the slightly lower C02 emissions when operating on E85 (typically
1 -3% lower than on gasoline), and a realistic rate of E85 fuel use, the benefit from FFVs is
no longer of the same magnitude that it was through the 2015 model year. Thus, we are no
17 85 FR 22609, April 23, 2020.
18 "E85 Flexible Fuel Vehicle Weighting Factor for Model Years 2020 and Later Vehicles," EPA Office of Air and
Radiation, CD-20-12.
87
-------
longer illustrating a g/mi benefit to manufacturers specific to producing FFVs. The impact of
E85, a lower-GHG fuel than gasoline, is inseparable from, and built into, the 2-cycle
emissions described earlier.
Km usance m Improved Air Conditioning Systems
Almost all new cars and light trucks in the United States are equipped with air conditioning
(A/C) systems. There are two mechanisms by which A/C systems contribute to the
emissions of greenhouse gases: through leakage of hydrofluorocarbon (HFC) refrigerants
(i.e., "direct" emissions) and through the combustion of fuel to provide mechanical power
to the A/C system (i.e., "indirect" emissions). The EPA 2-cycle compliance tests do not
measure either A/C refrigerant leakage or the increase in tailpipe emissions attributable to
the additional engine load of A/C systems. Thus, the GHG emission regulations include a
provision that allows manufacturers to earn optional credits for implementing technologies
that reduce either type of A/C-related emissions.
Air Conditioning Leakage Performance Credits
Refrigerants used in automotive air conditioning systems can have high global warming
potentials (GWP)19, such that leakage of a small amount of refrigerant can have a far
greater impact on global warming than emissions of a similar mass of CO2. The impacts of
refrigerant leakage can be reduced significantly by using systems with leak-tight
components, by using a refrigerant with a lower GWP, or by implementing both
approaches.
A manufacturer choosing to generate A/C leakage credits is required to calculate a leakage
"score" for the specific A/C system. This score is based on the number, performance, and
technology of the components, fittings, seals, and hoses of the A/C system and is calculated
as refrigerant emissions in grams per year, using the procedures specified by the SAE
Surface Vehicle Standard J2727. The score is then converted to a g/mi credit value based on
the GWP of the refrigerant. In model year 2012, all leakage credits were based on
improvements to the A/C system components (e.g., O-rings, seals, valves, and fittings).
In model year 2013, GM and Honda introduced vehicles using a refrigerant with a
significantly reduced GWP. This new refrigerant, HFO-1234yf, has a GWP of 4, compared to
a GWP of 1430 for the predominant refrigerant at the time, HFC-134a, as illustrated in
19 The global warming potential (GWP) represents how much a given mass of a chemical contributes to global
warming over a given time period compared to the same mass of C02. The GWP of C02 is 1.0.
88
-------
Figure 5.6. In the seven model years since, low GWP refrigerant use has expanded to 85%
of new vehicles. Of the remaining 15% of new vehicles that are not using HFO-1234yf, all
except a very small number achieved A/C leakage credits through improved performance
of the air conditioning system. All manufacturers reported some type of A/C leakage credits
in the 2020 model year, resulting in an overall performance credit of 13.9 g/mi for the
industry.
Figure 5.6. HFO-1234yf Adoption by Manufacturer
80% -
60% -
e
CO
¦C
W
.2 40% -
"S
¦o
o
CL
20% -
0%-
2013 2014 2015 2016 2017 2018 2019 2020
Model Year
Air Conditioning Efficiency Performance Credits
The A/C system also contributes to increased tailpipe CO2 emissions through the additional
work required by the engine to operate the compressor, fans, and blowers. This power
demand is ultimately met by using additional fuel, which is converted into C02 by the
engine during combustion and exhausted through the tailpipe. Increasing the overall
efficiency of an A/C system reduces the additional load on the engine from A/C operation,
and thereby leads to a reduction in fuel consumption and a commensurate reduction in
GHG emissions.
89
-------
Most of the additional load on the engine from A/C systems comes from the compressor,
which pressurizes the refrigerant and pumps it around the system loop. A significant
additional load may also come from electric or hydraulic fans, which move air across the
condenser, and from the electric blower, which moves air across the evaporator and into
the cabin. Manufacturers have several options for improving efficiency, including more
efficient compressors, fans, and motors, and system controls that avoid over-chilling the air
(and subsequently re-heating it to provide the desired air temperature). For vehicles
equipped with automatic climate-control systems, real-time adjustment of several aspects
of the overall system can result in improved efficiency.
The regulations provide manufacturers with a "menu" of A/C system technologies and
associated credit values (in g/mi of C02), some of which are described above. These credits
are capped at 5.7 g/mi for all vehicles in the 2012-2016 model years, and at 5.0 and 7.2
g/mi for cars and trucks, respectively, in the 2017 and later model years. Seventeen out of
twenty manufacturers reported A/C efficiency credits in 2020, resulting in an average credit
of 5.0 g/mi for the industry.
Air Conditioning Performance Credit Summary
A summary of the A/C leakage and efficiency performance credits reported by the industry
is shown in Figure 5.7. Leakage credits have been more prevalent than efficiency credits,
but both credit types are growing in use. Figure 5.8 shows the benefit of A/C credits, for
each manufacturer's fleet for the 2020 model year. All twenty manufacturers used the A/C
credit provisions—leakage reductions, efficiency improvements, or both—as part of their
compliance demonstration in the 2020 model year. Jaguar Land Rover had the highest
reported credit on a per vehicle g/mi basis, at 24.2 g/mi. Thus, A/C credits are the
equivalent of about an 8% reduction from tailpipe emissions for Jaguar Land Rover. Most
manufacturers reported total A/C credits of at least 12 g/mi. The overall industry reported
an average of 18.9 g/mi of total A/C credits.
90
-------
Figure 5.7. Fleetwide A/C Credits by Credit Type
2012 2013 2014 2015 2016 2017 2018 2019 2020
Model Year
Figure 5.8. Total A/C Credits by Manufacturer for Model Year 2020
25
20
E
s
(/>
-)—I
'u
-------
*Vi i .ance Credits for "Off-Cycle" Technology
In some cases, manufacturers employ technologies that result in CO2 emission reductions
that are not adequately captured on the 2-cycle test procedures. These benefits are
acknowledged in EPA's regulations by giving manufacturers three pathways by which to
accrue "off-cycle" performance credits. The first, and most widely used, pathway is a
predetermined list or "menu" of credit values for specific off-cycle technologies. The second
pathway is to use a broader array of emissions testing (5-cycle testing) to demonstrate the
CO2 emission reduction. The third pathway allows manufacturers to seek EPA approval to
use an alternative methodology to demonstrate C02 emission reductions.
Off Cycle Performance Credits Based on the Menu
The first pathway to generating off-cycle credits is for a manufacturer to install
technologies from a predetermined list or "menu" of technologies preapproved by EPA. The
off-cycle credit menu provides specific credit values, or the calculation method for such
values, for each technology.20 Technologies from the menu may be used beginning in
model year 2014. This pathway allows manufacturers to use conservative credit values
established by EPA for a wide range of off-cycle technologies, with minimal data submittal
or testing requirements.
The regulations clearly define each technology and any requirements that apply for the
technology to generate credits. Figure 5.9 shows the adoption of menu technologies, by
manufacturer. The amount of credit awarded varies for each technology and between cars
and trucks. The impact of credits from this pathway on a manufacturer's fleet is capped at
10 g/mi, meaning that any single vehicle might accumulate more than 10 g/mi, but the
cumulative effect on a single manufacturer's fleet may not exceed a credit of more than 10
g/mi. Off-cycle technology credits based on the menu were widely used in model year 2020,
with more than 90% of off-cycle credits generated via the menu pathway. Each of these
technologies is discussed below.
20 See 40CFR 86.1869-12(b).
-------
Figure 5.9. Off-Cycle Menu Technology Adoption by Manufacturer, Model Year
2020
All
Manufacturers"
46%
13%
15%
89%
78%
15%
42%
55%
50%
30%
85%
Volvo -
0%
0%
12%
100%
100%
27%
100%
0%
3%
0%
100%
VW-
31%
0%
19%
87%
0%
4%
98%
7%
85%
89%
98%
Toyota -
14%
0%
10%
61%
98%
19%
32%
54%
29%
0%
99%
Tesla -
100%
100%
no/
100%
0%
0%
0%
0%
0%
0%
100%
U /o
Subaru -
58%
0%
7%
90%
0%
0%
0%
94%
73%
94%
85%
Stellantis -
53%
15%
17%
99%
99%
3%
96%
51%
67%
0%
72%
Nissan -
54%
0%
4%
73%
31%
18%
27%
72%
0%
48%
77%
Mitsubishi -
0%
0%
0%
78%
0%
4%
0%
97%
2%
63%
87%
Mercedes -
0%
0%
29%
90%
0%
0%
0%
0%
0%
0%
94%
McLaren -
50%
0%
no/„
0%
0%
0%
0%
100%
100%
0%
99%
u /o
Mazda -
73%
0%
19%
82%
0%
0%
17%
96%
0%
17%
100%
Kia ¦
3%
62%
13%
100%
100%
22%
4%
73%
35%
97%
50%
Jaguar Land Rover-
98%
0%
44%
100%
100%
0%
0%
93%
99%
0%
100%
Hyundai -
5%
89%
17%
93%
100%
19%
11%
78%
3%
97%
62%
Honda -
33%
0%
6%
100%
100%
9%
9%
91%
50%
90%
100%
GM-
62%
0%
19%
99%
100%
23%
54%
7%
70%
0%
93%
Ford-
91%
0%
29%
100%
100%
25%
65%
88%
92%
0%
71%
Ferrari -
0%
0%
0%
0%
0%
0%
0%
100%
0%
0%
100%
BMW-
72%
100%
16%
54%
0%
0%
87%
0%
88%
0%
100%
Aston Martin -
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
Credit Type

• Active Aerodynamic Improvements
* Engine Idle Start Stop
Thermal Control Technologies
High Efficiency Alternator
Active Warmup
High Efficiency Lighting
93
-------
Active Aerodynamics
Active aerodynamics refers to technologies which are automatically activated to improve
the aerodynamics of a vehicle under certain conditions. These include grill shutters and
spoilers, which allow air to flow over and around the vehicle more efficiently, and
suspension systems that improve air flow at higher speeds by reducing the height of the
vehicle. Credits are variable and based on the measured improvement in the coefficient of
drag, a test metric that reflects the efficiency of airflow around a vehicle. Most
manufacturers implemented at least some level of active aerodynamics on their model
year 2020 vehicles. Tesla reported the highest implementation, at 100% of all new vehicles.
Overall, 46% of new vehicles qualified for these credits, reducing overall fleet C02
emissions by 0.4 g/mi.
Thermal Control Technologies
Thermal control systems help to maintain a comfortable air temperature of the vehicle
interior, without the use of the A/C system. These technologies lower the load on the A/C
system and thus the amount of fuel required to run the A/C system, subsequently lowering
GHG tailpipe emissions. The thermal control technologies included in the off-cycle menu
are:
• Active and passive cabin ventilation - Active systems use mechanical means to
vent the interior, while passive systems rely on ventilation through convective air
flow. Credits available for this technology range from 1.7 to 2.8 g/mi.
• Active seat ventilation - These systems move air through the seating surface,
transferring heat away from the vehicle occupants. Credits are 1.0 g/mi for cars
and 1.3 g/mi for trucks.
• Glass or glazing - Credits are available for glass or glazing technologies that
reduce the total solar transmittance through the glass, thus reducing the heat
from the sun that reaches the occupants. The credits are calculated based on
the measured solar transmittance through the glass and on the total area of
glass on the vehicle.
• Solar reflective surface coating - Credits are available for solar reflective surface
coating (e.g., paint) that reflects at least 65% of the infrared solar energy. Credits
are 0.4 g/mi for cars and 0.5 g/mi for trucks.
Active cabin ventilation was installed on 13% of all new vehicles in model year 2020, with
Tesla and BMW utilizing this technology on all of their vehicles. Hyundai, Kia, and Stellantis
94
-------
were the only other manufacturers utilizing active cabin ventilation. Passive cabin
ventilation technologies, however, were used much more widely, with nine manufacturers
reporting passive cabin ventilation on at least 98% of model year 2020 production, and a
78% adoption rate overall.
Active seat ventilation was used by many manufacturers and the rate of implementation
remained about the same at 15% in model year 2020. Jaguar Land Rover was the leader in
adopting active seat ventilation, with implementation on 44% of their vehicles. As was the
case in the previous model year, there was significant penetration of glass or glazing
technology with more than 89% of the model year 2020 vehicles equipped with glass or
glazing technologies. Solar reflective coatings have been used less widely, with a
penetration of 15% across new vehicles in model year 2020, and no manufacturer above
30%.
Due to the likelihood of synergistic effects among the various thermal technologies, the
total credit allowed from this technology group is capped at 3.0 g/mi for cars and 4.3 g/mi
for trucks. Overall, manufacturers widely adopted thermal control technologies, which
reduced model year 2020 overall new vehicle fleet CO2 emissions by 3.0 g/mi
Active Engine and Transmission Warmup
Active engine and transmission warmup systems use heat from the vehicle that would
typically be wasted (exhaust heat, for example) to warm up key elements of the engine,
allowing a faster transition to more efficient operation. An engine or transmission at its
optimal operating temperature minimizes internal friction, and thus operates more
efficiently and reduces tailpipe CO2 emissions. Systems that use a single heat-exchanging
loop that serves both transmission and engine warmup functions are eligible for either
engine or transmission warmup credits, but not both. Active engine and transmission
warmup technologies are each worth credit up to 1.5 g/mi for cars and 3.2 g/mi for trucks.
Most manufacturers adopted warmup technologies for their engines, transmissions, or
both. Active engine warmup was installed in 42% of all new vehicles, and active
transmission warmup in 55% of the fleet, resulting in a CO2 reduction of about 2.5 g/mi
across the 2020 model year fleet. Volvo, VW, and Stellantis led the industry in active engine
warmup, with more than 95% of their new vehicles employing the technology. Ferrari,
McLaren, Mitsubishi, and Mazda led the industry in active transmission warmup
technologies, with more than 95% of their new vehicles utilizing these technologies.
95
-------
Engine Idle Stop/Start
Engine idle stop/start systems allow the engine to turn off when the vehicle is at a stop,
automatically restarting the engine when the driver releases the brake and/or applies
pressure to the accelerator. If equipped with a switch to disable the system, EPA must
determine that the predominant operating mode of the system is the "on" setting
(defaulting to "on" every time the key is turned on is one basis for such a determination).
Thus, some vehicles with these systems are not eligible for credits. Credits range from 1.5
to 4.4 g/mi and depend on whether the system is equipped with an additional technology
that, at low ambient temperatures, allows heat to continue to be circulated to the vehicle
occupants when the engine is off during a stop-start event.
The implementation of stop/start has been increasing rapidly, as discussed in Section 4,
which aggregates and reports on these systems regardless of the regulatory eligibility for
credits. In model year 2020, 50% of new vehicles qualified for and claimed this credit,
resulting in a fleetwide CO2 reduction of about 1.9 g/mi. Nine manufacturers installed stop
start systems on at least half of all of their model year 2020 vehicles.
High Efficiency Exterior Lights
High efficiency lights (e.g., LEDs) reduce the total electric demand, and thus the fuel
consumption and related GHG emissions, of a lighting system in comparison to
conventional incandescent lighting. Credits are based on the specific lighting locations,
ranging from 0.06 g/mi for turn signals and parking lights to 0.38 g/mi for low beams. The
total of all lighting credits summed from all lighting locations may not exceed 1.0 g/mi.
Unlike some other off-cycle technologies, safety regulations require that all vehicles must
be equipped with lights, and the popularity of high efficiency lights across manufacturers
may reflect that lighting improvements are relatively straightforward to implement. All
manufacturers, except Aston Martin, reported high efficiency lighting implementation on at
least half of their new vehicles in model year 2020. Overall, in model year 2020, 85% of new
vehicles implemented high efficiency lighting in some form, reducing fleetwide CO2
emissions by 0.4 g/mi.
High Efficiency Alternators
Alternators convert mechanical energy from an engine into electrical energy, which is used
to power the vehicle's electrical system and accessories. High efficiency alternators reduce
the amount of mechanical energy needed to drive the alternator and provide the necessary
electrical requirements of the vehicle. High efficiency alternators were added as an off-
cycle menu option beginning in model year 2020. Eight manufacturers claimed menu
fit. 96
-------
credits for high-efficiency alternators, reducing fleetwide CO2 emissions by 0.2 g/mi. Four
additional manufacturers claimed credits for high-efficiency alternators in model year 2020
through the alternative methodology described below.
Solar Panels
Vehicles that use batteries for propulsion, such as electric, plug-in hybrid electric, and
hybrid vehicles may receive credits for solar panels that are used to charge the battery
directly or to provide power directly to essential vehicle systems (e.g., heating and cooling
systems). Credits are based on the rated power of the solar panels. Hyundai claimed this
credit in model year 2020 for a small number of vehicles.
Summary of Off-Cycle Menu-Based Performance Credits
As shown in Table 5.3, manufacturers are using a mix of off-cycle menu technologies,
though each uses and benefits from the individual technologies to differing degrees. In
model year 2020, the industry achieved 7.8 g/mi of credits from the menu, based on a
production weighted average of credits across all manufacturers. Ford, Jaguar Land Rover,
and Stellantis reached the 10 g/mi cap in 2020. For those manufacturers, the sum of the
credits from individual technologies in Table 5.3 will exceed the total allowable credits, and
only the 10 g/mi value will be used in subsequent calculations. The overall industry-wide
value of 7.8 g/mi reflects the capped credits for Ford, Jaguar Land Rover, and Stellantis.
Off-Cycle Performance Credits Based on 5-Cycle Testing
In cases where additional laboratory testing can demonstrate emission benefits, a second
pathway allows manufacturers to use a broader array of emission tests (known as "5-cycle"
testing because the methodology uses five different testing procedures) to demonstrate
and justify off-cycle CO2 credits.21 The additional emission tests allow emission benefits to
be demonstrated over elements of real-world driving not captured by the GHG compliance
tests, including high speeds, rapid accelerations, and cold temperatures. Credits
determined according to this methodology do not undergo additional public review.
21 See 40CFR 86.1869-12(c).
97
-------
Table 5.3. Model Year 2020 Off-Cycle Technology Credits from the Menu, by Manufacturer and Technology
(g/mi)
Manufacturer
Active
Aero-
dynamics
Thermal
Controls
Active
Engine
Warmup
Active
Trans
Warmup
Engine
Stop-Start
High
Efficiency
Alternator
High
Efficiency
Lighting
Solar
Panels
Total
Menu
Credits
Aston Martin
-
-
-
-
-
-
-
-
-
BMW
1.3
3.1
2.0
-
2.6
-
0.9
-
9.9
Ferrari
-
-
-
1,5
-
-
0.8
-
2.3
Ford
1.1
3.6
1.8
2.6
3.8
-
0.3
-
10.0
GM
0.7
3.7
1.4
0.2
2.7
-
0.6
-
9.2
Honda
0.1
2.9
0.2
2.1
1.8
0.5
0.4
-
7.9
Hyundai
0.0
2.7
0.2
1.4
0.1
0.6
0.1
0.0
5.1
Jaguar Land Rover
0.8
4.1
-
2.9
4.3
-
0.5
-
10.0
Kia
0.0
3.1
0.1
1.8
0.8
0.6
0.1
-
6.3
Mazda
0.3
1.1
0.5
2.4
-
0.1
0.6
-
5.0
McLaren
0.5
-
-
1.5
1.5
-
0.8
-
4.3
Mercedes
-
1.6
-
-
-
-
0.9
-
2.5
Mitsubishi
-
0.8
-
2.4
0.0
0.3
0.4
-
4.0
Nissan
0.3
1.0
0.5
1.5
-
0.3
0.3
-
3.9
Stellantis
0.6
3.8
2.9
1.6
2.9
-
0.2
-
10.0
Subaru
0.3
1.3
-
2.8
2.1
0.6
0.4
-
7.5
Tesla
1.1
3.0
-
-
-
-
0.7
-
4.8
Toyota
0.1
3.4
0.9
0.6
0.9
-
0.5
-
6.4
VW
0.1
1.2
2.5
0.2
2.9
1.6
0.8
-
9.3
Volvo
-
3.6
2.8
-
0.1
-
1.0
-
7.5
All Manufacturers
0.4
3.0
1.1
1.3
1.9
0.2
0.4
0.0
7.8
98
-------
GM is the only manufacturer to date to have claimed off-cycle credits based on 5-cycle
testing. These credits are for an auxiliary electric pump used on certain GM gasoline-
electric hybrid vehicles to keep engine coolant circulating in cold weather while the vehicle
is stopped and the engine is off. This enables the engine stop-start system to turn off the
engine more often during cold weather, while maintaining a comfortable temperature
inside the vehicle. GM received off-cycle credits during the early credits program for
equipping hybrid full size pick-up trucks with this technology and has since applied the
technology to several other vehicles through model year 2017. They did not claim credits
for this technology in model year 2020.
Off-Cycle Performance Credits Based on an Alternative Methodology
This third pathway for off-cycle technology performance credits allows manufacturers to
seek EPA approval to use an alternative methodology for determining off-cycle technology
C02 credits.22 This option is only available if the benefit of the technology cannot be
adequately demonstrated using the 5-cycle methodology. Manufacturers may also use this
option for model years prior to 2014 to demonstrate CO2 reductions for technologies that
are on the off-cycle menu, or reductions that exceed those available via use of the menu.
The regulations require that EPA seek public comment on and publish each manufacturer's
application for credits sought using this pathway. About half of the manufacturers have
petitioned for and been granted credits using this pathway in the 2020 model year.23
In the fall of 2013, Mercedes requested off-cycle credits for the following off-cycle
technologies in use or planned for implementation in the 2012-2016 model years: stop-
start systems, high-efficiency lighting, infrared glass glazing, and active seat ventilation. EPA
approved methodologies for Mercedes to determine these off-cycle credits in September
2014. Subsequently, Stellantis, Ford, and GM requested off-cycle credits under this
pathway, which EPA approved in September 2015. Stellantis and Ford submitted
applications for off-cycle credits from high efficiency exterior lighting, solar reflective
glass/glazing, solar reflective paint, and active seat ventilation. Ford's application also
demonstrated off-cycle benefits from active aerodynamic improvements (grill shutters),
active transmission warm-up, active engine warm-up technologies, and engine idle stop-
start. GM's application described the real-world benefits of an A/C compressor made by
Denso with variable crankcase suction valve technology. EPA approved the credits for
22 See 40CFR 86.1869-12(d).
23 EPA maintains a web page on which we publish the manufacturers' applications for these credits, the
relevant Federal Register notices, and the EPA decision documents. See https://www.epa.gov/ve-
certification/compliance-information-light-duty-greenhouse-gas-ghg-standards.
99
-------
Stellantis, Ford, and GM in September of 2015. EPA approved additional credits under this
pathway for the Denso compressor in 2017 for BMW, Ford, GM, Hyundai, Toyota, and VW.
In December 2016, EPA approved a methodology for determining credits from high-
efficiency alternators that Ford had applied for in 2016. EPA subsequently approved high-
efficiency alternator credits also for Stellantis, GM, and Toyota. High efficiency alternators
use new technologies that reduce the overall load on the engine while continuing to meet
the electrical demands of the vehicle systems, resulting in lower fuel consumption and
lower CO2 emissions. High efficiency alternators were added to the off-cycle menu credits
beginning in model year 2020; however, Ford, Stellantis, GM, and Toyota continue to
receive credits through the alternative methodology.
In September of 2017 GM applied for credits under this pathway for "active climate-
controlled seats," which provide cooled air directly to the occupants through the seats, thus
reducing the overall load on the air conditioning system. GM reported credits for this
technology in the 2018 and 2019 model year. The most recent addition to the list of
technologies receiving credits through the alternative methodology pathway came in
October 2019 with the approval of a pulse width modulated brushless motor power
controller. This "brushless motor" technology is used to improve the efficiency of the HVAC
system. Toyota applied for and received the brushless motor credit for the 2013 through
2019 model years. Subaru also received approval for brushless motors and received credit
for this technology in model year 2020.
In October 2020, Honda applied for credits for a "cold storage evaporator." Air-conditioning
systems employing this technology essentially freeze a mass of material during normal
operation, such that the material can provide cabin cooling when the engine is off. This
allows stop-start systems to leave the engine off longer, resulting in reductions in
emissions and fuel usage. Honda has received credits for this technology for model years
2017-2020.
Most of the approved credits have been for previous model years. Credit balances have
been updated to include retroactive credits that have been reported to EPA, and any
relevant tables that include data from previous model years will reflect the addition of
these credits. Table 5.4shows the impact of the credits submitted for brushless motors,
improved air conditioning systems, high-efficiency alternators, and cold storage
evaporators. On a total fleetwide basis, the aggregated credit is 0.5 g/mi for model year
2020.
100
-------
Table 5.4. Model Year 2020 Off-Cycle Technology Credits from an Alternative
Methodology, by Manufacturer and Technology (g/mi)
Manufacturer
Brushless
Motors
Improved
A/C
Systems
High-
Efficiency
Alternator
Cold
Storage
Evaporator
Total
Alternative
Methodology
Credits
Ford
-
-
0.7
-
0.7
GM
-
0.7
0.8
-
1.4
Honda
-
-
-
0.5
0.5
Hyundai
-
0.0
-
-
0.0
Kia
-
0.0
-
-
0.0
Nissan
-
0.1
-
-
0.1
Stellantis
-
-
0.5
-
0.5
Subaru
0.0
-
-
-
0.0
Toyota
-
-
0.3
-
0.3
All Manufacturers
0.0
0.1
0.3
0.0
0.5
Off-Cycle Performance Credit Summary
In total, the industry achieved 8.3 g/mi of off-cycle performance credits in model year 2020.
More than 90% of those credits were claimed using technologies, and credit definitions, on
the off-cycle menu. The remaining credits were due almost entirely to manufacturer
submitted alternative methodologies. Figure 5.10 shows the average credit, in g/mi, that
each manufacturer achieved in model year 2020. Ford and GM achieved the highest gram
per mile benefit from off-cycle credits at 10.7 g/mi, followed closely by Stellantis at 10.5
g/mi and Jaguar Land Rover at 10.0 g/mi. Most manufacturers achieved at least some off-
cycle credits; Aston Martin was the only manufacturer to not report any off-cycle credits for
model year 2020.
1
-------
Figure 5.10. Total Off-Cycle Credits by Manufacturer for Model Year 2020
¦ Non-Menu Credits
Menu Credits
Alternative Standards for Methane and Nitrous Oxide
As part of the GHG Program, EPA set emission standards for methane (CH4) and nitrous
oxide (N2O) at 0.030 g/mi for CbUand 0.010 g/mi for N2O. Current levels of CH4 and N2O
emissions are generally well below these established standards, however the caps were set
to prevent future increases in emissions.
There are three different ways for a manufacturer to demonstrate compliance with these
standards. First, manufacturers may submit test data as they do for all other non-GHG
emission standards; this option is used by most manufacturers. Because there are no
credits or deficits involved with this approach, and there are no consequences with respect
to the CO2 fleet average calculation, the manufacturers are not required to submit this data
as part of their GHG reporting. Hence, this GHG compliance report does not include
information from manufacturers using this option.
102
-------
The second option for manufacturers is to include CH4 and N2O, on a C02-equivalent basis,
when calculating their fleet average performance values, in lieu of demonstrating
compliance with the regulatory caps. This method directly accounts for CH4 and N20,
increasing the performance value of a manufacturer's fleets, while the standards remain
unchanged. Analyses of emissions data have shown that use of this option may add
approximately 3 g/mi to a manufacturer's fleet average. Only Subaru chose to use this
approach in the 2020 model year.
The third option for complying with the CH4 and N2O standards allows manufacturers to
propose an alternative, less stringent CH4 and/or N2O standard for any vehicle that may
have difficulty meeting the specific standards. However, manufacturers that use this
approach must also calculate the increased emissions due to the less stringent standards
and the production volumes of the vehicles to which those standards apply, and then add
that impact from their overall fleet performance. Nine manufacturers made use of the
flexibility offered by this approach in the 2020 model year. In aggregate, the impact of the
methane and nitrous oxide flexibilities resulted in an increase in the industry-wide
performance of about 0.2 g/mi.
Summary of Manufactui « 1 ma nee
Each of the performance credits and adjustments described here have been used by
manufacturers as part of their compliance strategies under the GHG program. As
described above, the availability of these provisions, and the magnitude of their impact,
has varied both by manufacturer and model year. Table 5.5 through Table 5.10 below
detail the impact of these provisions by manufacturer for model year 2020, and for the
aggregated industry over the course of the GHG Program. The performance values in these
tables can be derived by subtracting the credits and adjustment from the 2-Cycle Tailpipe
value.
103
-------
Table 5.5. Manufacturer Performanc lodel Year 2020, All (g/mi)
Performance Credits and Adjustments
2-Cycle Adv. Off- CH4 & Performance
Manufacturer
Tailpipe
Tech
FFV
A/C
Cycle
n2o
Value
Aston Martin
346
-
-
4.6
-
-
341
BMW
273
1.7
-
20.9
9.9
-
240
Ferrari
423
-
-
13.8
2.3
-
407
Ford
303
0.3
-
22.7
10.7
-0.2
269
GM
304
1.9
-
21.2
10.7
-0.6
271
Honda
235
0.2
-
19.8
8.4
-
207
Hyundai
244
1.8
-
12.7
5.1
-0.0
225
Jaguar Land Rover
314
3.9
-
24.2
10.0
-0.1
276
Kia
255
1.5
-
18.3
6.3
-0.1
229
Mazda
247
-
-
4.5
5.0
-0.5
238
McLaren
395
-
-
2.3
4.3
-
388
Mercedes
301
0.3
-
13.6
2.5
-
284
Mitsubishi
229
0.4
-
20.3
4.0
-
205
Nissan
248
1.5
-
13.4
4.0
-
229
Stellantis
332
0.6
-
22.9
10.5
-0.2
298
Subaru
241
0.4
-
20.0
7.5
-
213
Tesla
0
154.9
-
18.7
4.8
-
-178
Toyota
258
0.6
-
15.1
6.7
-0.3
236
VW
279
3.3
-
21.4
9.3
-0.0
245
Volvo
282
2.2
-
22.1
7.5
-
250
All Manufacturers
275
2.9
-
18.9
8.3
-0.2
245
Table Sa\. industry Performance by Model Year, All (g/mi)
Performance Credits and Adjustments
2-Cycle Adv. Off- CH4 & Performance
Model Year
Tailpipe
Tech
FFV
A/C
Cycle
n2o
Value
2012
302
-
8.1
6.1
1.0
-0.2
287
2013
294
-
7.8
6.9
1.1
-0.3
278
2014
294
-
8.9
8.5
3.3
-0.2
273
2015
286
-
6.4
9.4
3.4
-0.2
267
2016
285
-
-
10.3
3.6
-0.1
271
2017
284
2.2
-
13.8
5.5
-0.2
262
2018
280
3.7
-
16.3
6.9
-0.1
253
2019
282
3.0
-
18.0
7.6
-0.1
253
2020
275
2.9
-
18.9
8.3
-0.2
245
104
-------
Table 5.7. Manufacturer Performance ^ WHodel Year 2020, Car (g/mi)
Performance Credits and Adjustments
2-Cycle Adv. Off- CH4 & Performance
Manufacturer
Tailpipe
Tech
FFV
A/C
Cycle
n2o
Value
Aston Martin
346
-
-
4.6
-
-
341
BMW
253
2.2
-
18.5
7.6
-
225
Ferrari
423
-
-
13.8
2.3
-
407
Ford
255
0.9
-
18.2
7.6
-0.0
228
GM
229
6.3
-
16.1
7.2
-0.0
199
Honda
206
0.3
-
16.9
4.8
-
184
Hyundai
234
2.0
-
13.3
4.6
-0.0
214
Jaguar Land Rover
263
-
-
18.8
6.9
-
237
Kia
218
2.5
-
16.7
4.6
-0.1
194
Mazda
233
-
-
3.0
2.9
-0.1
227
McLaren
395
-
-
2.3
4.3
-
388
Mercedes
273
0.0
-
11.2
1.7
-
260
Mitsubishi
203
1.0
-
18.0
2.5
-
181
Nissan
220
2.3
-
14.7
2.9
-
200
Stellantis
307
-
-
18.5
5.7
-0.2
283
Subaru
245
-
-
11.6
2.6
-
231
Tesla
0
153.1
-
18.6
4.7
-
-176
Toyota
203
1.3
-
14.0
5.1
-0.1
183
VW
239
5.2
-
17.5
5.8
-0.0
211
Volvo
256
1.1
-
18.0
4.8
-
232
All Manufacturers
221
6.5
_
15.4
5.1
-0.0
194
Table 5.8. Industry Performance by Model Year, Car (g/mi)
Performance Credits and Adjustments
2-Cycle Adv. Off- CH4& Performance
Model Year
Tailpipe
Tech
FFV
A/C
Cycle
n2o
Value
2012
259
-
4.0
5.4
0.6
-0.1
249
2013
251
-
4.0
6.3
0.7
-0.3
240
2014
250
-
4.6
7.5
2.2
-0.3
236
2015
243
-
3.1
8.1
2.3
-0.1
230
2016
240
-
-
8.8
2.3
-0.1
229
2017
235
4.3
-
10.2
3.5
-0.0
217
2018
228
7.6
-
13.0
4.1
-0.0
204
2019
228
6.3
-
14.8
4.4
-0.1
203
2020
221
6.5
-
15.4
5.1
-0.0
194
105
-------
Table 5.9. Manufacturer Performance ^ WHodel Year 2020, Truck (g/mi)
Performance Credits and Adjustments
2-Cycle Adv. Off- CH4 & Performance
Manufacturer
Tailpipe
Tech
FFV
A/C
Cycle
n2o
Value
Aston Martin
-
-
-
-
-
-
-
BMW
300
1.0
-
24.2
13.2
-
262
Ferrari
-
-
-
-
-
-
-
Ford
318
0.1
-
24.1
11.7
-0.2
282
GM
336
-
-
23.4
12.2
-0.8
301
Honda
269
-
-
23.1
12.6
-
233
Hyundai
318
-
-
8.3
8.9
-
301
Jaguar Land Rover
316
4.1
-
24.4
10.1
-0.1
277
Kia
307
-
-
20.6
8.8
-
278
Mazda
260
-
-
5.9
6.8
-0.9
248
McLaren
-
-
-
-
-
-
-
Mercedes
319
0.6
-
15.2
3.0
-
300
Mitsubishi
250
-
-
22.0
5.2
-
223
Nissan
300
-
-
10.9
6.1
-
283
Stellantis
336
0.7
-
23.5
11.2
-0.2
301
Subaru
240
0.4
-
21.6
8.5
-
210
Tesla
0
206.3
-
22.6
8.3
-
-237
Toyota
310
-
-
16.2
8.4
-0.5
286
VW
304
2.2
-
23.8
11.5
-0.0
267
Volvo
289
2.5
-
23.3
8.3
-
255
All Manufacturers
311
0.5
_
21.3
10.4
-0.3
279
Table 5.10. Industry Performance by Model Year, Truck (g/mi)
Performance Credits and Adjustments
2-Cycle Adv. Off- CH4 & Performance
Model Year
Tailpipe
Tech
FFV
A/C
Cycle
n2o
Value
2012
369
-
14.5
7.3
1.6
-0.3
346
2013
360
-
13.8
7.9
1.7
-0.3
337
2014
349
-
14.3
9.7
4.6
-0.1
321
2015
336
-
10.3
11.0
4.6
-0.2
310
2016
332
-
-
11.8
5.1
-0.2
316
2017
330
0.2
-
17.3
7.5
-0.3
305
2018
320
0.6
-
19.0
9.1
-0.2
292
2019
318
0.7
-
20.2
9.8
-0.1
288
2020
311
0.5
-
21.3
10.4
-0.3
279
106
-------
C. GHG Program Credits and Deficits
The previous two sections outlined how to determine manufacturer standards and
manufacturer performance values for the current model year. The next step in the
compliance process it to compare the car and truck standards to the corresponding
performance values to determine if each fleet was above or below the standards. This
process then allows manufacturers to determine if each fleet will create GHG program
credits or deficits. These program credits are the credits available to manufacturers to
bank, trade, and ultimately show compliance with the overall GHG program.
Program credits are always expressed as mass-based credits in megagrams of C02. A mass-
based credit metric captures the performance of each manufacturer's fleets relative to the
standards, the total number of vehicles produced in each fleet, and the expected lifetime
vehicle miles travelled for those vehicles. This conversion is necessary to enable the
banking and trading of credits across manufacturer fleets, model years, and between
manufacturers. To convert g/mi emission rates to total emission reductions in Mg, see the
insert "How to Calculate Total Emissions from an Emission Rate" at the beginning of this
section.
Manufacturers also had a limited, and voluntary, option to generate program credits in
model years 2009 through 2011 from early technology adoption before the standards went
into effect. Credit expirations, credit forfeitures, and credit trades between manufacturers,
are also important in determining the overall program credits available to manufacturers.
This section will detail these components of the GHG program, which are essential in
determining manufacturer overall credit balances and manufacturer compliance with the
GHG program.
Generating Credits and Deficits from M.x - I W h *ViMu>iance
Manufacturers can calculate the credits or deficits created within a model year by
comparing their car and truck fleet standards to their respective performance values and
converting from a gram per mile emission rate to a mass-based total. When a car or truck
fleet is below the applicable standard, that fleet generates credits for the manufacturer.
Conversely, when a car or truck fleet is above the applicable standard, that fleet generates
deficits.
The GHG program evaluates car and truck fleets separately, which means that there is no
single, overall standard for manufacturers. However, it is possible to calculate an effective
overall manufacturer standard, and performance value, from the underlying passenger car
107
-------
and truck data. Figure 5.11 illustrates the performance of all manufacturers in model year
2020, compared to their effective overall standards.
Of the 20 manufacturers that produced vehicles in model year 2020, five were below their
overall effective standards and one manufacturer was even with the standards. Tesla,
Honda, Subaru, Volvo, and Aston Martin were all below their standards, and generated net
credits (accounting for credits and deficits from each manufacturer's car and truck fleets).
Fifteen manufacturers were above their standards and generated net deficits in model year
2020. The fact that manufacturers were above their standards in Figure 5.11 does not
mean that these manufacturers were out of compliance with the GHG program, as
discussed later in this report.
Figure 5.11. Performance and Standards by Manufacturer, Model Year 2020
Mercedes
Stellantis
McLaren
Hyundai
Mazda
Ferrari
VW
Nissan
BMW
Kia
GM
Jaguar Land Rover
Toyota
Ford
Mitsubishi
Volvo
Honda
Subaru
Aston Martin
Tesla
| Standard
k Performance
-200
¦3601388
|198|225

¦215|238

M386|407
0 200
Compliance GHG (g/mi)
400

-------
In model year 2020, seven manufacturers generated credits from their truck fleets, while
ten generated deficits. Four manufacturers generated credits with their car fleets,
compared to 16 that generated deficits. Table 5.11 through Table 5.16 provide a summary
of the standards, manufacturer performance, and the credits and deficits generated by
each manufacturer's car and truck fleets for model year 2020 and for the aggregated
industry for model years 2009-2020 (including early credits). These tables show only credits
generated within a model year, and do not account for credits used to offset deficits in
other model years, credits that are traded between manufacturers, or credits that have
expired or been forfeited. The tables showing combined car and truck, or overall industry
values, are aggregated from the underlying car and truck data.
109
-------
Table 5.11. Credits Earned by Manufacturers in Model Year 2020, All

Performance

Standard

Credits

Value
Standard
Exceedance

Generated
Manufacturer
(g/mi)
(g/mi)
(g/mi)
Production
(Mg)
Aston Martin
341
374
-33
1,732
11,026
BMW
240
224
16
307,186
-1,021,626
Ferrari
407
386
21
2,244
-9,174
Ford
269
266
3
1,689,365
-1,123,744
GM
271
259
12
2,268,518
-5,847,712
Honda
207
217
-10
1,339,880
2,868,950
Hyundai
225
198
27
742,556
-3,958,850
Jaguar Land Rover
276
266
10
137,831
-318,453
Kia
229
214
15
644,884
-1,970,803
Mazda
238
215
23
234,218
-1,164,296
McLaren
388
360
28
1,691
-9,405
Mercedes
284
240
44
323,862
-3,044,545
Mitsubishi
205
203
2
102,582
-30,441
Nissan
229
213
16
1,030,249
-3,346,608
Stellantis
298
268
30
1,547,816
-10,218,644
Subaru
213
228
-15
606,124
2,042,007
Tesla
-178
206
-384
182,179
13,755,004
Toyota
236
227
9
2,022,067
-3,453,989
VW
245
226
19
430,426
-1,732,145
Volvo
250
254
-4
105,379
107,050
All Manufacturers
245
239
6
13,720,789
-18,466,398
Table 5.12. Total Credits Earned in Mod Yt\ rs 2009-2020, All

Performance

Standard

Credits

Model
Value
Standard
Exceedance

Generated
Credit
Year
(g/mi)
(g/mi)
(g/mi)
Production
(Mg)
Expiration
2009
-
-
-
-
98,522,058
2014
2010
-
-
-
-
96,891,340
2021
2011
-
-
-
-
38,770,273
2021
2012
287
299
-12
13,446,550
33,033,097
2021
2013
278
292
-14
15,200,118
42,234,774
2021
2014
273
287
-13
15,514,338
43,292,494
2021
2015
267
274
-7
16,740,264
25,218,704
2021
2016
271
263
8
16,279,911
-27,615,344
2021
2017
262
258
4
17,015,504
-15,761,664
2022
2018
253
252
1
16,259,244
-3,682,197
2023
2019
253
246
7
16,139,324
-23,502,913
2024
2020
245
239
6
13,720,789
-18,466,398
2025
D
-------
Table 5.13. Credits Earned by Manufacturers in Model Year 2020, Car

Performance

Standard

Credits

Value
Standard
Exceedance

Generated
Manufacturer
(g/mi)
(g/mi)

(g/mi) Production
(Mg)
Aston Martin
341
374

-33
1,732
11,026
BMW
225
194

31
189,955
-1,139,530
Ferrari
407
386

21
2,244
-9,174
Ford
228
194

34
448,765
-3,001,264
GM
199
186

13
745,477
-1,960,354
Honda
184
187

-3
768,077
449,007
Hyundai
214
188

26
663,210
-3,388,070
Jaguar Land Rover
237
204

33
5,230
-34,014
Kia
194
185

9
400,909
-722,157
Mazda
227
186

41
117,548
-948,842
McLaren
388
360

28
1,691
-9,405
Mercedes
260
198

62
141,500
-1,715,472
Mitsubishi
181
173

8
49,028
-81,290
Nissan
200
189

11
706,451
-1,521,629
Stellantis
283
200

83
233,540
-3,784,139
Subaru
231
182

49
109,343
-1,041,743
Tesla
-176
204

-380
176,802
13,133,010
Toyota
183
188

-5
1,061,549
1,084,957
VW
211
188

23
180,177
-794,323
Volvo
232
211

21
26,464
-108,848
All Manufacturers
194
189

5
6,029,692
-5,582,254
Table 5,14, Total Credits Earned in Model Years 2009-2020, Car
Performance
Standard

Credits

Model
Value Standard
Exceedance

Generated
Credit
Year
(g/mi) (g/mi)
(g/mi)
Production
(Mg)
Expiration
2009
-

-
-
58,018,752
2014
2010
-

-
-
50,856,700
2021
2011
-

-
-
8,831,637
2021
2012
249 267

-18
8,657,393
30,484,967
2021
2013
240 261

-21
9,747,624
39,249,608
2021
2014
236 253

-17
9,209,352
30,407,996
2021
2015
230 241

-12
9,602,215
22,043,043
2021
2016
229 231

-2
9,012,178
3,411,251
2021
2017
217 219

-2
8,954,269
2,905,609
2022
2018
204 209

-6
7,800,108
8,548,771
2023
2019
203 198

4
7,170,547
-5,860,191
2024
2020
194 189

5
6,029,692
-5,582,254
2025
1
-------
Table 5.15. Credits Earned by Manufacturers in Model Year 2020, Truck
Manufacturer
Aston Martin
BMW
Ferrari
Ford
GM
Honda
Hyundai
Jaguar Land Rover
Kia
Mazda
McLaren
Mercedes
Mitsubishi
Nissan
Stellantis
Subaru
Tesla
Toyota
VW
Volvo
All Manufacturers
Performance Standard
Value Standard Exceedance
(g/mi) (g/mi) (g/mi)
262
282
301
233
301
277
278
248
300
223
283
301
210
-237
286
267
255
279
266
289
290
252
269
268
255
240
268
227
258
279
237
275
265
250
267
272
-7
11
-19
32
9
23
8
32
-4
25
22
-27
-512
21
17
-12
7
Production
117,231
1,240,600
1,523,041
571,803
79,346
132,601
243,975
116,670
182,362
53,554
323,798
1,314,276
496,781
5,377
960,518
250,249
78,915
7,691,097
Credits
Generated (Mg)
117,904
1,877,520
-3,887,358
2,419,943
-570,780
-284,439
-1,248,646
-215,454
-1,329,073
50,849
-1,824,979
-6,434,505
3,083,750
621,994
-4,538,946
-937,822
215,898
-12,884,144
Table 5.16. Total Credits Earned in Model Wvrs 2009-2020, Truck

Performance

Standard

Credits

Model
Value
Standard
Exceedance

Generated
Credit
Year
(g/mi)
(g/mi)
(g/mi)
Production
(Mg)
Expiration
2009
-
-
-
-
40,503,306
2014
2010
-
-
-
-
46,034,640
2021
2011
-
-
-
-
29,938,636
2021
2012
346
349
-2
4,789,157
2,548,130
2021
2013
337
339
-3
5,452,494
2,985,166
2021
2014
321
330
-9
6,304,986
12,884,498
2021
2015
310
312
-2
7,138,049
3,175,661
2021
2016
316
297
19
7,267,733
-31,026,595
2021
2017
305
295
10
8,061,235
-18,667,273
2022
2018
292
286
6
8,459,136
-12,230,968
2023
2019
288
279
9
8,968,777
-17,642,722
2024
2020
279
272
7
7,691,097
-12,884,144
2025
112
-------
Program Credits for Early Adoption of Technology
The GHG program included an optional provision that allowed manufacturers to generate
credits in the 2009-2011 model years, prior to the implementation of regulatory standards
in model year 2012. This flexibility allowed manufacturers to generate credits for achieving
tailpipe C02 emissions targets or introducing emission-reducing technology before model
year 2012. Sixteen manufacturers participated in the early credits program, generating a
large bank of credits for the industry before the standards took effect in model year 2012.
The pathways for earning credits under the early credit program mirrored those built into
the annual GHG requirements, including improved tailpipe C02 performance and A/C
systems, off-cycle credits for other technologies that reduced CO2 emissions, and credits
for manufacturing electric, plug-in hybrid, and fuel cell vehicles.
Of the 234 Tg of early credits, 85% of those credits were generated from performing better
than the tailpipe CO2 emissions targets established in the regulations. To earn credits
based on tailpipe C02 performance, manufacturers could demonstrate tailpipe emissions
levels below either California or national standards, dependent on the state the car was
sold in. California developed GHG standards prior to the adoption of the EPA GHG
program, and some states had adopted these standards. In all other states, CO2 levels were
calculated based on the national CAFE standards. Of the remaining early credits, about 10%
were created through improving A/C system leakage, 4% were due to A/C efficiency
improvements, and just over 1 % were due to off-cycle credits for other technologies.
The model year 2009 credits could not be traded between companies and were limited to a
5-year credit life. Thus, all credits earned in model year 2009, or about a third of the early
credits generated, expired at the end of the 2014 model year if not already used. The
remaining 2010-2011 model year credits were banked and may be used until the 2021
model year. Manufacturers can no longer generate early credits. More details of the early
credit program can be found in the "Early Credits Report," which was released by EPA in
2013.24
24 Greenhouse Gas Emission Standards for Light-Duty Automobiles: Status of Early Credit Program for Model
Years 2009-2011, Compliance Division, Office of Transportation and Air Quality, U.S. Environmental Protection
Agency, Report No. EPA-420-R-13-005, March 2013.
113
-------
Figure 5.12. Early Credits by Manufacturer
Expiration and Forfeiture of Credits
All credits earned within the GHG program have expiration dates. However, the only credits
that have expired so far were credits earned under the early credit program in model year
2009. All credits earned from model years 2010 to 2016, which make up a little more than
half of all credits currently held by manufacturers, will expire at the end of model year
2021. Beginning in model year 2017, all credits have a 5-year lifetime; for example, credits
earned in model year 2020 will expire at the end of model year 2025.
A limited number of credits have been forfeited by several manufacturers. Although
forfeiture and expiration both have fundamentally the same effect - a loss or removal of
credits - forfeiture is considered a different and less common mechanism, brought about
by unique circumstances. Hyundai and Kia forfeited a specified quantity of 2013 model
year credits after an investigation into their testing methods that concluded with a
settlement announced on November 3, 2014.
114
-------
VW similarly forfeited some credits, deducted from their 2017 model year balance. In the
course of the investigation concerning defeat devices in VW's diesel vehicles, the EPA
discovered that the company employed software to manage vehicle transmissions in
gasoline vehicles. This software causes the transmission to shift gears during the EPA-
prescribed emissions test in a manner that sometimes optimizes fuel economy and
greenhouse gas (GHG) emissions during the test, but not under normal driving conditions.
This resulted in inflated fuel economy values for some vehicles. VW forfeited credits to
account for the higher C02 emissions of these vehicles in actual use.
Additional manufacturers forfeited credits because of their participation in the Temporary
Lead Time Alternative Allowance Standards (TLAAS). Opting into these less stringent
standards, which are no longer available, came with some restrictions, including the
requirement that any credits accumulated by using the TLAAS standards may not be used
by or transferred to a fleet meeting the primary standard. This impacted Porsche, which
was bought by VW in 2012. Porsche held some credits earned against the TLAAS standards
at the time they were merged with VW, and VW was not participating in the TLAAS
program. Thus, those credits could not carry over to the merged company and were lost.
Similarly, Mercedes and Volvo reached the end of the TLAAS program, which applied
through the 2015 model year, with credits in their TLAAS bank that could not be
transferred to their post-2015 bank and thus were forfeited.
Credit Transactions
Credit trading among manufacturers has been an important part of the program for many
manufacturers. An active credit market is enabling manufacturers to purchase credits to
demonstrate compliance, with nine manufacturers selling credits, ten manufacturers
purchasing credits, and 80 credit transactions occurring since the inception of the program.
Credits may be traded among manufacturers with a great deal of flexibility, however there
are several limitations, including:
1) Manufacturers must offset any existing deficits before selling credits.
2) Manufacturers may not sell credits they do not have.
3) Manufacturers are the only parties that may engage in credit transactions and hold
credits (although a third party may facilitate transactions).
4) Manufacturers may not sell early credits created in model year 2009.
5) Manufacturers may not sell credits generated under an alternative standard
(including TLAAS and small volume manufacturer standards).
115
-------
As of October 31, 2021, about 142 Tg of credits have been traded between manufacturers.
Figure 5.13 shows the total quantity of credits that have been bought or sold by
manufacturers since the beginning of the GHG program. Credits that have been sold are
shown as negative credits, since the sale of credits will reduce the selling manufacturer's
credit balance. Conversely, credits that have been purchased are shown as positive credits,
since they will increase the purchasing manufacturer's credit balance.
Manufacturers can purchase or sell credits generated in any model year. The model year
the credits were generated in remains important, as those credits must be used (and will
expire) according to the model year in which they were originally created. Figure 5.13 also
shows the distribution of credits sold and acquired by the model year after which the
credits will expire. One additional credit transaction occurred in 2021, as Volvo used
banked credits to offset the small deficit Lotus held prior to their merger into one
manufacturer under the GHG program.
Figure 5.13. Total Credits Transactions
Sold Purchased
80-
60-
CM
O
o
O 40
O)
T3
CD
^ 20
to
2 0
o
03
,2
^ -20 H
-40-
-60
Expiring 2025
Expiring 2024
Expiring 2023
Expiring 2022
Expiring 2021

^ # # Cf

-------
D. End of Year GHG Program Credit Balances
The final GHG program credit balance at the end of each model year, and compliance
status, for each manufacturer relies on all the components outlined to this point in the
report. Manufacturer car and truck standards and performance within each model year,
early credits, credit trades, credit forfeitures, and credit expirations are all required to
determine final model year credit balances for each manufacturer. If a manufacturer ends
the model year with a positive credit balance, they are in compliance with the GHG
program and the accrued credits will be carried forward to the next model year. If a
manufacturer ends the model year with a deficit, that manufacturer must offset the deficit
within three years to avoid non-compliance. For example, a manufacturer with a deficit
remaining from model year 2017 after the 2020 model year would be considered out of
compliance with the 2017 standards. Manufacturers may not carry forward any credits
unless all deficits have been offset.
Using Credits to Offset Deficits
If a manufacturer generates a deficit from either their car or truck fleets, that deficit must
be offset from existing credits, if they are available. When applying credits, the oldest
available credits are applied to the current deficit by default. Credits earned in past model
years may be applied to car or truck deficits, regardless of how they were generated. Table
5.17 shows a simple example. In this case, a manufacturer generated 300,000 Mg of credits
from its car fleets in model years 2018, 2019, and 2020. The manufacturer's truck fleets did
not generate any credits or deficits in model years 2018 or 2019 but generated a deficit of
500,000 Mg in 2020. Because the oldest credits are applied first, credits generated in model
year 2018 are the first credits applied towards the 2020 truck deficit, then 2019 and 2020
credits would be applied until the deficit is offset. After offsetting the example truck deficit
in Table 5.17, this manufacturer would be left with 100,000 Mg of credits from model year
2019, and 300,000 Mg of credits from model year 2020 to bank for future use.
Table 5.17. E> v e of a Deficit Offset with Credits from Previous Model Years
Generated Truck Credits
Generated Car Credits
Model Model Model
Year 2018 Year 2019 Year 2020
0 0 -500,000
300,000 300,000 300,000
Applied to 2020 Deficits
-300,000 -200,000
Remaining Credits
0 100,000 300,000
117
-------
The complete credit and deficit accounting for each manufacturer also includes the impact
of credits earned as part of the early credit program, credit trades, credit forfeitures, and
credit expirations over the full span of the GHG program. The detailed deficit offset
calculations for each manufacturer are not published in this report, since some of the
credit trade information is considered confidential business information and is not
published in detail by EPA. However, most of the underlying data for all manufacturers and
model years is available on the Automotive Trends website at
https://www.epa.gov/automotive-trends.
Compliance Status After the 2020 Mc-« I W u
EPA determines the compliance status of each manufacturer based on their credit balance
at the end of the model year, after offsetting all deficits. Because credits may not be carried
forward unless deficits from all prior model years have been resolved, a positive credit
balance means compliance with the current and all previous model years of the program. If
a manufacturer ends the model year with a any deficits, that manufacturer must offset the
deficit within three years to avoid non-compliance. For model year 2020, deficits from
model year 2017 or prior would be considered non-compliant
Figure 5.14 shows the credit balance of all manufacturers after model year 2020 including
the breakdown of expiration dates, and the distribution of deficits, by age of the deficit. All
manufacturers, except one, ended the 2020 model year with a positive credit balance and
are thus in compliance with model year 2020 and all previous years of the GHG program.
McLaren, the only manufacturer carrying a deficit into the 2021 model year, has deficits
remaining from model year 2019 and 2020, but those deficits are within the allowable time
span and will not result in non-compliance or enforcement actions from EPA. However,
McLaren will have to offset the existing deficits in future model years either by producing
efficient vehicles that exceed future standards, or by purchasing credits from other
manufacturers.
The breakdown of each manufacturer's final model year 2020 credit balance, based on the
source of the credits or deficits, is shown in Table 5.18. Each manufacturer has pursued a
unique combination of early credits generated in model years 2009-2011, credits or deficits
created in model years 2012-2020, and credit expirations, forfeitures, and trades to
achieve their current credit balance. The "net" credits earned in Table 5.18 are a sum of all
credits and deficits earned by a manufacturer, and may not be the amount of credits
remaining due to the use of banked credits across model years. The actual distribution of
credits, by expiration date, and deficits, by the age of the deficit, are shown in Table 5.19.
118
-------
Figure 5.14. Manufacturer Credit Balance After Model Year 2020
Stellantis
Honda
Toyota
GM
Subaru
Nissan
Ford
Hyundai
Mazda
BMW
Mercedes
VW
Volvo
Mitsubishi
Kia
Jaguar Land Rover
Tesla
Ferrari
Karma Automotive
Aston Martin
BYD Motors
McLaren
I
I
Credits
| Expiring 2025
Expiring 2024
Expiring 2023
Expiring 2022
Expiring 2021
McLaren ended Model Year 2020 with a deficit of 0.0142 Tg.
I 10 20 30 40
GHG Credits (Tg of COz)
fjM]
-------
Table 5.18. Final Credit Balance by Manufacturer for Model Year 2020 (Mg)
Manufacturer
Early Credits
Earned
2009-2011
Net Credits
Earned
2012-2019
Net Credits
Earned
2020
Credits
Expired
Credits
Forfeited
Credits
Purchased
or Sold
Final 2020
Credit
Balance
Aston Martin
3,332
-22,334
11,026
-
-
35,844
27,868
BMW
1,251,522
-603,236
-1,021,626
-134,791
-
5,500,000
4,991,869
BYD Motors
-
5,568
-
-
-
-
5,568
Coda
-
7,251
-
-
-
-7,251
-
Ferrari
-
-153,006
-9,174
-
-
265,000
102,820
Ford
16,116,453
-461,183
-1,123,744
-5,882,011
-
-
8,649,515
GM
25,788,547
-10,003,223
-5,847,712
-6,998,699
-
20,777,251
23,716,164
Honda
35,842,334
59,948,101
2,868,950
-14,133,353
-
-45,015,245
39,510,787
Hyundai
14,007,495
2,937,205
-3,958,850
-4,482,649
-169,775
-
8,333,426
Jaguar Land Rover
-
-2,874,258
-318,453
-
-
3,572,736
380,025
Karma Automotive
-
58,852
-
-
-
-2,841
56,011
Kia
10,444,192
-5,449,785
-1,970,803
-2,362,882
-123,956
-
536,766
Lotus
-
-3,147
-
-
-
3,147
-
Mazda
5,482,642
4,851,951
-1,164,296
-1,390,883
-
-
7,779,414
McLaren
-
-16,969
-9,405
-
-
12,193
-14,181
Mercedes
378,272
-12,273,308
-3,044,545
-
-28,416
17,427,713
2,459,716
Mitsubishi
1,449,336
1,373,190
-30,441
-583,146
-
-850,000
1,358,939
Nissan
18,131,200
13,055,704
-3,346,608
-8,190,124
-
-3,545,570
16,104,602
Porsche
-
426,439
-
-
-426,439
-
-
Stellantis
10,827,083
-45,886,550
-10,218,644
-
-
90,572,768
45,294,657
Subaru
5,755,171
15,438,093
2,042,007
-491,789
-
-
22,743,482
Suzuki
876,650
-183,097
-
-265,311
-
-428,242
-
Tesla
49,772
39,810,154
13,755,004
-
-
-53,354,766
260,164
Toyota
80,435,498
18,294,380
-3,453,989
-29,526,679
-
-38,962,431
26,786,779
VW
6,613,985
-5,698,224
-1,732,145
-1,442,571
-219,419
4,000,000
1,521,626
Volvo
730,187
638,383
107,050
-
-85,163
-306
1,390,151
All Manufacturers
234,183,671
73,216,951
-18,466,398
-75,884,888
-1,053,168
-
211,996,168
120
-------
Table 5.19. Distribution of Credits by Expiration Date (Mg)
Model Model Model
Final 2020 Credits Credits Credits Credits Credits Year Year Year Non-
Credit Expiring in Expiring in Expiring in Expiring in Expiring in 2020 2019 2018 Compliant
Manufacturer Balance 2021 2022 2023 2024 2025 Deficits Deficits Deficits Deficits
Aston Martin
27,868
-
-
1,672
15,170
11,026
-
-
BMW
4,991,869
800,746
3,652,752
138,811
281,656
117,904
-
-
BYD Motors
5,568
4,871
529
168
-
-
-
-
Coda
-
-
-
-
-
-
-
-
Ferrari
102,820
90,448
8,180
4,192
-
-
-
-
Ford
8,649,515
6,771,995
-
-
-
1,877,520
-
-
GM
23,716,164
10,953,638
2,127,946
5,534,580
-
5,100,000
-
-
Honda
39,510,787
17,044,774
4,938,779
9,292,781
5,365,503
2,868,950
-
-
Hyundai
8,333,426
8,333,426
-
-
-
-
-
-
Jaguar Land Rover
380,025
333,235
-
-
46,790
-
-
-
Karma Automotive
56,011
56,011
-
-
-
-
-
-
Kia
536,766
536,766
-
-
-
-
-
-
Lotus
-
-
-
-
-
-
-
-
Mazda
7,779,414
7,443,421
171,051
164,942
-
-
-
-
McLaren
(14,181)
-
-
-
-
-
-9,405
-4,776
Mercedes
2,459,716
2,459,716
-
-
-
-
-
-
Mitsubishi
1,358,939
880,387
171,946
202,975
52,782
50,849
-
-
Nissan
16,104,602
15,452,917
651,685
-
-
-
-
-
Porsche
-
-
-
-
-
-
-
-
Stellantis
45,294,657
9,129,522
4,731,544
11,915,822
11,073,882
8,443,887
-
-
Subaru
22,743,482
10,551,290
3,215,610
2,599,744
3,293,088
3,083,750
-
-
Suzuki
-
-
-
-
-
-
-
-
Tesla
260,164
-
-
52,161
-
208,003
-
-
Toyota
26,786,779
20,111,181
1,911,327
2,106,312
1,573,002
1,084,957
-
-
VW
1,521,626
-
-
1,009,990
511,636
-
-
-
Volvo
1,390,151
-
78,996
778,606
316,651
215,898
-
-
All Manufacturers
211,996,168
110,954,344
21,660,345
33,802,756
22,530,160
23,062,744
-9,405
-4,776
1
-------
Figure 5.15 shows the overall industry performance, standards, and credit bank for all
years of the GHG program. The industry created a large bank of credits using the early
credits provision in model year 2009 through 2012. For the next four years, manufacturers
continued to generate credits, as the industry GHG performance was below the industry-
wide average standard. In the last five years, the industry GHG performance has been
above the standard, resulting in net withdrawals from the bank of credits to maintain
compliance. In model year 2020, the industry achieved overall GHG performance at 245
g/mi, while the standard fell from 246 g/mi to 239 g/mi. The gap between the standard and
GHG performance decreased from 7 g/mi in model year 2019 to 6 g/mi in model year 2020.
To maintain compliance, the industry drew down their industry-wide total credit bank by
about 18 teragrams (Tg), which was less than 10% of the total available credit balance. The
overall industry emerged from model year 2020 with a bank of 212 Tg of GHG credits
available for future use, as seen in Figure 5.15.
In addition to the balance of the industry-wide credit bank, the expiration date and
distribution of credits are also important factors. Credits earned in model year 2017 or
beyond have a five-year life, while all prior credits (a little over half of the current bank) will
expire at the end of model year 2021. An active credit market has resulted in
approximately 80 credit trades since 2012, with nine manufacturers selling credits and nine
manufacturers purchasing credits.
After accounting for the use of credits, and the ability to carry forward a deficit in the case
of McLaren, the industry overall does not face any non-compliance issues as of the end of
the 2020 model year.
122
-------
Figure 5.15. Industry Performance and Standards, Credit Generation and Use
-------
Appendices: Methods and Additional Data
A. Sources of Input Data
Nearly all of the data for this report are based on automakers' direct submissions to EPA.
EPA has required manufacturers to provide vehicle fuel economy to consumers since 1977
and has collected data on every new light-duty vehicle model sold in the United States
since 1975. The data are obtained either from testing performed by EPA at the National
Vehicle and Fuel Emissions Laboratory in Ann Arbor, Michigan, or directly from
manufacturers using official EPA test procedures.
National fuel economy standards have been in place in the United States for cars and light
trucks since 1978. The Department of Transportation, through the National Highway Traffic
Safety Administration (NHTSA), has the responsibility for setting and enforcing fuel
economy standards through the Corporate Average Fuel Economy (CAFE) program. Since
the inception of CAFE, EPA has been responsible for establishing test procedures and
calculation methods, and for collecting data used to determine vehicle fuel economy levels.
EPA calculates the CAFE value for each manufacturer and provides it to NHTSA. NHTSA
publishes the final CAFE values in its annual "Summary of Fuel Economy Performance"
reports at www.nhtsa.gov/Laws-&-Regulations/CAFE—Fuel-Economy. Since model year
2012, NHTSA and EPA have maintained coordinated fuel economy and greenhouse gas
standards that apply to model year 2012 through model year 2026 vehicles. EPA's light-
duty GHG program is described in detail in Section 5 of this report.
The data that EPA collects for this report comprise the most comprehensive database of its
kind. For recent model years, the vast majority of data in this report comes from the
Engines and Vehicles Compliance Information System (EV-CIS) database maintained by EPA.
This database contains a broad amount of data associated with CO2 emissions and fuel
economy, vehicle and engine technology, and other vehicle performance metrics. This
report extracts only a portion of the data from the EV-CIS database.
In some cases, the data submitted by automakers are supplemented by data that were
obtained through independent research by EPA. For example, EPA relied on published data
from external sources for certain parameters of pre-model year 2011 vehicles: (1) engines
with variable valve timing (WT), (2) engines with cylinder deactivation, and (3) vehicle
footprint, as automakers did not submit this data until model year 2011. EPA projects
footprint data for the preliminary model year 2021 fleet based on footprint values for
existing models from previous years and footprint values for new vehicle designs available
1
-------
through public sources. In addition, vehicle 0-to-60 acceleration values are not provided by
automakers, but are either calculated from other Trends data, as discussed in Section 3, or
taken from external sources.
The website for this report has been expanded with an emphasis on allowing users to
access and evaluate more of the data behind this report. EPA plans to continue to add
content and tools on the web to allow transparent access to public data. To explore the
data using EPA's interactive data tools, visit the report webpage at
https://www.epa.gov/automotive-trends.
The full database used for the analysis in this report is not publicly available. The detailed
production data necessary for demonstrating compliance is considered confidential
business information by the manufacturers and cannot be shared by EPA. However, EPA
will continue to provide as much information as possible to the public.
Preliminary vs Final Data
For each model year, automakers submit two phases of data: preliminary data provided
to EPA for vehicle certification and labeling prior to the model year sales, and final data
submitted after the completion of the model year for compliance with EPA's light-duty GHG
regulations and NHTSA's CAFE program.
Preliminary data are collected prior to the beginning of each model year and are not used
for manufacturer GHG compliance. Automakers submit "General Label" information
required to support the generation of thejoint EPA/NHTSA Fuel Economy and Environment
Labels that appear on all new personal vehicles. As part of these submissions, automakers
report pre-model year vehicle production projections for individual models and
configurations to EPA.
Final data are submitted a few months after the end of each model year and include
detailed final production volumes. EPA and NHTSA use this final data to determine
compliance with GHG emissions and CAFE standards. These end-of-the-year submissions
include detailed final production volumes. All data in this report for model years 1975
through 2020 are considered final. However, manufacturers can submit requests for
compliance credits for previous model years, so it is possible that additional credits under
the GHG program could be awarded to manufacturers.
Since the preliminary fuel economy values provided by automakers are based on projected
vehicle production volumes, they usually vary slightly from the final fuel economy values
that reflect the actual sales at the end of the model year. With each publication of this
A-2
-------
report, the preliminary values from the previous year are updated to reflect the final
values. This allows a comparison to gauge the accuracy of preliminary projections.
Table A.1 compares the preliminary and final fleetwide real-world fuel economy values for
recent years (note that the differences for CO2 emissions data would be similar, on a
percentage basis). Since model year 2011, the final real-world fuel economy values have
generally been close to the preliminary fuel economy values. In eight out of the last nine
years, manufacturer projections have led to preliminary estimates that were higher than
final data.
It is important to note that there is no perfect apples-to-apples comparison for model years
2011-2014 due to several small differences in data, such as inclusion of alternative fuel
vehicle (AFV) data. The preliminary values in Table A.1 through model year 2014 did not
integrate AFV data, while the final values in Table A.1 are the values reported elsewhere in
this report and do include AFV data. The differences due to this would be small, on the
order of 0.1 mpgorless.
Table A.1. Comparison of Preliminary and Final Real-World Fuel Economy
Values (mpg)
Model Year
Preliminary

Final Minus

Value
Final Value
Preliminary
2011
22.8
22.3
-0.5
2012
23.8
23.6
-0.2
2013
24.0
24.2
+0.2
2014
24.2
24.1
-0.1
2015
24.7
24.6
-0.2
2016
25.6
24.7
-0.9
2017
25.2
24.9
-0.3
2018
25.4
25.1
-0.3
2019
25.5
24.9
-0.6
2020
25.7
25.4
-0.3
2021 (prelim)
25.3
-
-
A-3
-------
B. Harmonic Averaging of Fuel Economy Values
Averaging multiple fuel economy values must be done harmonically in order to obtain a
correct mathematical result. Since fuel economy is expressed in miles per gallon (mpg), one
critical assumption with any harmonic averaging of multiple fuel economy values is
whether the distance term (miles, in the numerator of mpg) is fixed or variable. This report
makes the assumption that the distance term in all mpg values is fixed, i.e., that for
purposes of calculating a harmonically averaged fuel economy value, it is assumed that the
distance term (representing miles traveled) is equivalent across various vehicle fuel
economies. This assumption is the standard practice with harmonic averaging of multiple
fuel economy values (including, for example, in calculations for CAFE standards
compliance), and simplifies the calculations involved.
Mathematically, when assuming a fixed distance term as discussed above, harmonic
averaging of multiple fuel economy values can be defined as the inverse of the average of
the reciprocals of the individual fuel economy values. It is best illustrated by a simple
example.
Consider a round trip of 600 miles. For the first 300-mile leg, the driver is alone with no
other passengers or cargo, and, aided by a tailwind, uses 10 gallons of gasoline, for a fuel
economy of 30 mpg. On the return 300-mile trip, with several passengers, some luggage,
and a headwind, the driver uses 15 gallons of gasoline, for a fuel economy of 20 mpg. Many
people will assume that the average fuel economy for the entire 600-mile trip is 25 mpg,
the arithmetic (or simple) average of 30 mpg and 20 mpg. But, since the driver consumed
10+ 15 = 25 gallons of fuel during the trip, the actual fuel economy is 600 miles divided by
25 gallons, or 24 mpg.
Why is the actual 24 mpg less than the simple average of 25 mpg? Because the driver used
more gallons while (s)he was getting 20 mpg than when (s)he was getting 30 mpg.
This same principle is often demonstrated in elementary school mathematics when an
airplane makes a round trip, with a speed of 400 mph one way and 500 mph the other way.
The average speed of 444 mph is less than 450 mph because the airplane spent more time
going 400 mph than it did going 500 mph.
As in both of the examples above, a harmonic average will typically yield a result that is
slightly lower than the arithmetic average.
The following equation illustrates the use of harmonic averaging to obtain the correct
mathematical result for the fuel economy example above:
B-1
-------
2
Average mpg
I hough the above example was for a single vehicle with two different fuel economies over
two legs of a single round trip, the same mathematical principle holds for averaging the fuel
economies of any number of vehicles. For example, the average fuel economy for a set of 10
vehicles, with three 30 mpg vehicles, four 25 mpg vehicles, and three 20 mpg vehicles would
be (note that, in order to maintain the concept of averaging, the total number of vehicles in
the numerator of the equation must equal the sum of the individual numerators in the
denominator of the equation):
Arithmetic averaging, not harmonic averaging, provides the correct mathematical result for
averaging fuel consumption values (in gallons per mile, the inverse of fuel economy) and C02
emissions (in grams per mile). In the first, round trip, example above, the first leg had a fuel
consumption rate of 10 gallons over 300 miles, or 0.033 gallons per mile. The second leg had
a fuel consumption of 15 gallons over 300 miles, or 0.05 gallons per mile. Arithmetically
averaging the two fuel consumption values, i.e., adding them up and dividing by two, yields
0.04167 gallons per mile, and the inverse of this is the correct fuel economy average of 24
mpg. Arithmetic averaging also works for CO2 emissions values, i.e., the average of 200 g/mi
and 400 g/mi is 300 g/mi CO2 emissions.
In summary, fuel economy values must be harmonically averaged to maintain mathematical
integrity, while fuel consumption values (in gallons per mile) and C02 emissions values (in
grams per mile) can be arithmetically averaged.
10
Average mpg = —- : — = 24.4 mpg
-------
C. Fuel Economy and CO2 Metrics
The C02 emissions and fuel economy data in this report fall into one of two categories:
compliance data and estimated real-world data. These categories are based on the
purpose of the data, and the subsequent required emissions test procedures. The
following sections discuss the differences between compliance and real-world data and
how they relate to raw vehicle emissions test results.
2-Cycle Test Data
In 1975 when the Corporate Average Fuel Economy (CAFE) regulation was put into place,
EPA tested vehicles using two dynamometer-based test cycles, one based on city driving
and one based on highway driving. CAFE was—and continues to be—required by law to use
these "2-cycle tests". For consistency, EPA also adopted this approach for the GHG
regulations.
Originally, the fuel economy values generated from the "2-cycle" test procedure were used
both to determine compliance with CAFE requirements and to inform consumers of their
expected fuel economy via the fuel economy label. Today, the raw 2-cycle test data are
used primarily in a regulatory context as the basis for determining the final compliance
values for CAFE and GHG regulations.
The 2-cycle testing methodology has remained largely unchanged25 since the early 1970s.
Because of this, the 2-cycle fuel economy and CO2 values can serve as a useful comparison
of long-term trends. Previous versions of this report included 2-cycle fuel economy and CO2
data, referred to as "unadjusted" or "laboratory" values. These 2-cycle fuel economy values
are still available on the report website for reference. It is important to note that these 2-
cycle fuel economy values do not exactly correlate to the 2-cycle tailpipe CO2 emissions
values provided in Section 5 for the GHG regulations. There are three methodological
reasons for this:
25 There were some relatively minor test procedure changes made in the late 1970s that, in the aggregate,
made the city and highway tests slightly more demanding, i.e., the unadjusted fuel economy values for a given
car after these test procedure changes were made are slightly lower relative to prior to the changes. EPA has
long provided CAFE "test procedure adjustments" (TPAs) for passenger cars in recognition of the fact that the
original CAFE standards were based on the EPA test procedures in place in 1975 (there are no TPAs for light
trucks). The resulting impacts on the long-term unadjusted fuel economy trends are very small. The TPAs for
cars vary but are typically in the range of 0.2-0.5 mpg for cars, or 0.1-0.3 mpg when the car TPAs are averaged
over the combined car/truck fleet.
-------
1. The GHG regulations require a car and truck weighting based on a slightly higher
lifetime vehicle miles traveled (VMT) for trucks. The 2-cycle fuel economy values do
not account for this difference.
2. The GHG regulations allow manufacturers to use an optional compliance approach
which adds nitrous oxide and methane emissions to their 2-cycle CO2 emissions.
3. The GHG regulations and CAFE regulations result in very slightly different annual
production values. Prior to model year 2017, the 2-cycle fuel economy values rely on
CAFE production values (see Appendix D).
GHG Compliance Data
Compliance data in this report are used to determine how the manufacturers are
performing under EPA's GHG program. These data are reported in the Executive Summary
and Section 5. The 2-cycle CO2 test values form the basis for the compliance data, but there
are some important differences due to provisions in the standards. Manufacturers' model
year performance is calculated based on the measured 2-cycle CO2 tailpipe emissions as
well as optional performance credits and adjustments that manufacturers may qualify for
and use.
Compliance data also includes the overall credit balances held by each manufacturer, and
may incorporate credit averaging, banking, and trading by manufacturers. The compliance
process is explained in detail in Section 5. Compliance C02 data is not comparable to
estimated real-world C02 data, as described below.
Estimated Real-World Fuel Economy and CO2 Data
Estimated real-world (previously called "adjusted") data is EPA's best estimate of real-world
fuel economy and C02 emissions, as reported in Sections 1-4 of this report. The real-world
values are the best data for researchers to evaluate new vehicle C02 and fuel economy
performance. Unlike compliance data, the method for calculating real-world data has
evolved over time, along with technology and driving habits. These changes in
methodology are detailed in Appendix D.
Calculating estimated real-world fuel economy
Estimated real-world fuel economy data are currently measured based on the "5-cycle" test
procedure that utilizes high-speed, cold start, and air conditioning tests in addition to the 2-
cycle tests to provide data more representative of real-world driving. These additional
laboratory tests capture a wider range of operating conditions (including hot/cold weather
C-2
-------
and higher acceleration) that an average driver will encounter. City and highway results are
weighted 43% / 57%, consistent with fleetwide driver activity data.
Calculating estimated real-world CO2 emissions
The estimated real-world CO2 emissions shown in Sections 1 -4 are not based directly on
the 2-cycle tested values, but rather they are based on calculated values that convert
estimated real-world fuel economy values to CO2 using emission factors. This approach is
taken because: 1) test data are not available for most historic years of data, and 2) some
manufacturers choose to use an optional compliance approach which adds nitrous oxide
(N20) and methane (CH4) emissions to their C02 emissions (also referred to as Carbon
Related Exhaust Emissions, or CREE), leading to slightly different test results.
The estimated real-world C02 emissions from gasoline vehicles are calculated by dividing
8,887 g/gal by the fuel economy of the vehicle. The 8,887 g/gal emission factor is a typical
value for the grams of C02 per gallon of gasoline test fuel and assumes all the carbon is
converted to C02. For example, 8,887 g/gal divided by a gasoline vehicle fuel economy of 30
mpg would yield an equivalent C02 emissions value of 296 grams per mile.
The estimated real-world C02 emissions for diesel vehicles are calculated by dividing
10,180 g/gal by the diesel vehiclefuel economy value. The 10,180 g/gal diesel emission
factor is higher than for a gasoline vehicle because diesel fuel has a 14.5% higher carbon
content per gallon than gasoline. Accordingly, a 30 mpg diesel vehicle would have a C02
equivalent value of 339 grams per mile. Emissions for vehicles other than gasoline and
diesel are also calculated using appropriate emissions factors.
Example Comparison of Fuel Economy Metrics
The multiple ways of measuring fuel economy and GHG emissions can understandably
lead to confusion. As an illustration to help the reader understand the various fuel
economy values that can be associated with an individual vehicle, Table 1.2 shows three
different fuel economy metrics for the model year 2021 Toyota Prius Eco. The 2-cycle city
and highway fuel economy values are direct fuel economy measurements from the 2-cycle
tests and are harmonically averaged with a 55% city / 45% highway weighting to generate a
combined value. The 2-cycle laboratory tested city fuel economy of the Prius Eco is 84 mpg,
the highway fuel economy is 78 mpg, and the combined 2-cycle value is 81 mpg.
Using the 5-cycle methodology, the Toyota Prius Eco has a vehicle fuel economy label value
of 58 mpg city and 53 mpg highway. On the vehicle label, these values are harmonically
averaged using a 55% city / 45% highway weighting to determine a combined value of 56
C-3
-------
mpg. The estimated real-world fuel economy for the Prius Eco, which is the set of values
used in calculations for this report, has the same city and highway fuel economy as the
label, but the 43% city and 57% highway weighting leads to a combined value of 55 mpg,
which is one mpg less than the values found on the label.
Table C.l. Fuel Economy Metrics for the Model Wvr 2021 Toyota Prius Eco
City/Highway Test
Purpose Weighting Basis
Fuel
Economy
Metric
Basis for manufacturer
2-cycleTest compliance with
(unadjusted) standards 55%/ 45% 2-cycle
Consumer information
to compare individual
Label vehicles 55%/45% 5-cycle
Estimated Best estimate of real-
Real-World world performance 43%/57% 5-cycle
Fuel Economy Value
(MPG)
Combined
City/Hwy City Hwy
81
56
55
84
58
58
78
53
53
Greenhouse Gases other than COz
In addition to tailpipe CO2 emissions, vehicles may create greenhouse gas emissions in
several other ways. The combustion process can result in emissions of N2O, and CH4, and
leaks in vehicle air conditioning systems can release refrigerants, which are also
greenhouse gases, into the environment. N2O, CH4, and air conditioning greenhouse gases
are discussed as part of the GHG regulatory program in Section 5. Estimated real-world CO2
emissions in Sections 1 -4 only account for tailpipe C02 emissions.
The life cycle of the vehicle (including manufacturing and vehicle disposal) and the life cycle
of the fuels (including production and distribution) can also create significant greenhouse
gases. Life cycle implications of vehicles and fuels can vary widely based on the vehicle
technology and fuel and are outside the scope of this report. However, there is academic
research, both published and ongoing, in this area for interested readers.
C-4
-------
D. Historical Changes in the Database and
Methodology
Over the course of this report's publication, there have been some instances where
relevant methodologies and definitions have been updated. Since the goal of this report is
to provide the most accurate data and science available, updates are generally propagated
back to through the historical database. The current version of this report supersedes all
previous reports.
Changes in Estimated Real-world Fuel Economy and CO2
The estimated real-world fuel economy values in this report are closely related to the label
fuel economy values. Over the course of this report, there have been three updates to the
fuel economy label methodology (for model years 1985, 2008, and 2017), and these
updates were propagated through the Trends database. However, there are some
important differences in how the label methodology updates have been applied in this
report. This section discusses how these methodologies have been applied, partially or in
full, to the appropriate model years based on the authors' technical judgement. The
changes are intended to provide accurate real-world values for vehicles at the time they
were produced to better reflect available technologies, changes in driving patterns, and
composition of the fleet. These changes are also applicable to real-world C02 values, which
are converted from fuel economy values using emissions factors.
Model year 1975-1985; Universal Multipliers
The first change to the label methodology occurred when EPA recognized that changing
technology and driving habits led to real-world fuel economy results that over time were
diverging from the fuel economy values measured using the 2-cycle tests. To address this
issue, EPA introduced an alternative calculation methodology in 1985 that applied a
multiplication factor to the 2-cycle test data of 0.9 for city and 0.78 for highway. The
estimated real-world fuel economy values from model year 1975-1985 in this report were
calculated using the same multiplication factors that were required for the model year
1985 label update. The authors believe that these correction factors were appropriate for
new vehicles from model year 1975 through 1985. The combined fuel economy and CO2
values are based on a 55% city / 45% highway weighting factor, consistent with the CAFE
and label fuel economy calculations.
1
-------
Model year 1986-2010; The 2006 5-cycle methodology and 43% City / 57%
Highway Weighting
In 2006, EPA established a major change to the fuel economy label calculations by
introducing the 5-cycle methodology24 In addition to the city and highway tests required for
2-cycle fuel economy, the 5-cycle methodology introduces tests for high speeds (US06), air-
conditioning (SC03), and a cold temperature test. It also indirectly accounts for a number of
other factors that are not reflected in EPA laboratory test data (e.g., changing fuel
composition, wind, road conditions) through the use of a 9.5% universal downward
adjustment factor. The change from the universal adjustment factors to the 2006 5-cycle
method lowered estimated real-world fuel economy values, particularly for high fuel
economy vehicles. In the 2006 rulemaking, EPA projected an overall average fleetwide
adjustment of 11% lower for city fuel economy and 8% lower for highway fuel economy.
For model year 1986-2004, the authors implemented the 2006 5-cycle methodology by
assuming the changes in technology and driver behavior that led to lower real-world fuel
economy occurred in a gradual, linear manner over 20 years. We did not attempt to
perform a year-by-year analysis to determine the extent to which the many relevant factors
(including higher highway speed limits, more aggressive driving, increasing vehicle
horsepower-to-weight ratios, suburbanization, congestion, greater use of air conditioning,
gasoline composition, etc.) that have affected real-world fuel economy since 1985 have
changed over time.
Under the 5-cycle methodology, manufacturers could either: 1) perform all five tests on
each vehicle (the "full 5-cycle" method), 2) use an alternative analytical "derived 5-cycle"
method based on 2-cycle testing if certain conditions were met, or 3) voluntarily use lower
fuel economy label estimates than those resulting from the full 5-cycle or derived 5-cycle. If
manufacturers are required to perform all five tests, the results are weighted according to
composite 5-cycle equations.25 To use the derived 5-cycle method, manufacturers are
required to evaluate whether fuel economy estimates using the full 5-cycle tests are
comparable to results using the derived 5-cycle method. In recent years, the derived 5-cycle
approach has been used to generate approximately 85% of all vehicle label fuel economy
values.
For vehicles that were eligible to use the 2006 derived 5-cycle methodology, the following
equations were used to convert 2-cycle city and highway fuel economy values to label
24 See 71 Federal Register 77872, December 27, 2006.
25 See 71 Federal Register 77883-77886, December 27, 2006.
D-2
-------
economy values. These equations were based on the relationship between 2-cycle and 5-
cycle fuel economy data for the industry as a whole.
1
Label CITY =
(o.003259
1.1805
Label HWY =
2CYCLE CITY;
1
(0.001376 +
1.3466
2 CYCLE HWYJ
Over the same timeframe, EPA phased in a change in the city and highway weightings used
to determine a single combined fuel economy or CO2 value. EPA's analysis of real-world
driving activity underlying the 5-cycle fuel economy methodology assumed a "speed
cutpoint" of 45 miles per hour to differentiate between (and "bin" the amount of) city and
highway driving.26 Based on this speed cutpoint, the correct weighting for correlating the
new city and highway fuel economy values with real-world driving activity data from on-
road vehicle studies, on a miles driven basis, is 43% city and 57% highway; this updated
weighting is necessary to maintain the integrity of fleetwide fuel economy performance
based on Trends data. The 55% city and 45% highway weighting is still used for both Fuel
Economy and Environment Labels and the CAFE and GHG emissions compliance programs.
The authors used the same gradual, linear approach to phase in the change in city and
highway weightings along with the phase-in of the 2006 5-cycle methodology.
From model year 2005 to model year 2010, the 2006 5-cycle methodology and the 43% city
and 57% highway weightings were used to determine the real-world fuel economy values
for this report. This required using the derived 5-cycle equations and the 43% city and 57%
highway weightings to recalculate real-world fuel economy values for model year 2005 to
2007, because the 2006 5-cycle methodology was not required until 2008. Model year 2008
to model year 2010 real-world fuel economy values were the same as the label fuel
economy values, except for the city and highway weightings.
Model year 2011-present: Implementing the 2017 derived 5-cycle updates
In 2015, EPA released a minor update to the derived 5-cycle equations that modified the
coefficients used to calculate derived 5-cycle fuel economy from 2-cycle test data.27 This
26 See 71 Federal Register 77904, December 27, 2006.
27 See https://www.epa.gov/fueleconomy/basic-information-fuel-economy-labeling and
http://iaspub.epa.gov/otaqpub/display file.jsp?docid=35113&flag=1
D-3
-------
update was required under existing regulations and applies to fuel economy label
calculations for all model year 2017 and later vehicles. The following equations are used to
convert 2-cycIe test data values for city and highway to label fuel economy values:
1
Label CITY =
(°-
004091 +
1.1601 >
2 CYCLE CITY;
1
Label HWY =
(0.003191 +
1.2945
2 CYCLE HWYJ
The updated 5-cycle calculations introduced for model year 2017 and later labels were
based on test data from model year 2011 to model year 2016 vehicles. Therefore, the
authors chose to retroactively apply the updated 5-cycle methodology to model years 2011
to 2016. This required recalculating the real-world fuel economy of vehicles from model
year 2011 to 2016 using the new derived 5-cycle equations. Vehicles that conducted full 5-
cycle testing or voluntarily lowered fuel economy values were unchanged. The 43% city and
57% highway weightings were maintained. The changes for model years 2011 -2016 due to
the 5-cycle update were relatively small (0.1 to 0.2 mpg overall) and did not noticeably alter
the general data trends, therefore the authors determined that a phase-in period was not
required for this update.
Figure D.1 below summarizes the impact of the changes in real-world data methodology
relative to the 2-cycle test data, which has had a consistent methodology since 1975 (See
Appendix C for more information). Over time, the estimated real-world fuel economy of
new vehicles has continued to slowly diverge from 2-cycle test data, due largely to changing
technology, driving patterns, and vehicle design.
D-4
-------
Figure D.1. Estimated Real-World versus 2-Cycle Fuel Economy since Model
Year 1975
Phase I
1975-1985
Universal
adjustment factors
55/45% weighting
! \Ratio ol
Real-World
Estimated
to 2-cycle:
y 85.2% ^
Phase II
1986-2006
2006 5-cycle is phase-in
43/57% weighting phase-in
Phase III
2007-2010
5-cycle
43/57%
weighting 43/57% weighting
Phase IV
2011-present
Updated 5-cycle
Ratio of\
Real-World
Estimated
to 2-cycle:
v 76.3% J
1975 1980 1985 1990 1995 2000 2005 2010 2015 2020
Model Year
2-cycle method
unchanged
since 1975
Estimated
Real-World
Phases I - IV
Other Database Changes
Addition of Medium-Duty Passenger Vehicles
Beginning in 2011 medium-duty passenger vehicles (MDPVs), those SUVs and passenger
vans (but not pickup trucks) with gross vehicle weight ratings between 8,500 and 10,000
pounds, are included in the light-duty truck category. This coincided with new regulations
by NHTSA to treat these vehicles as light-duty, rather than heavy-duty, vehicles beginning in
model year 2011. This represents a minor change to the database, since the number of
MDPVs is much smaller than it once was (e.g., only 6,500 MDPVs were sold in model year
2012). It should be noted that this is one change to the database that has not been
propagated back through the historic database, as we do not have MDPV data prior to
model year 2011. Accordingly, this represents a small inflection point for the database for
the overall car and truck fleet in model year 2011; the inclusion of MDPVs decreased
average real-world fuel economy by 0,01 mpg and increased average real-world C02
emissions by 0,3 g/mi, compared to the fleet without MDPVs. The impacts on the truck fleet
D-5
-------
only were about twice as high, but still very small in absolute terms. Pickup trucks above
8,500 pounds are not included in this report.
Addition of Alternative Fuel Vehicles
Data from alternative fuel vehicles are integrated into the overall database, beginning with
MY 2011 data. These vehicles include electric vehicles, plug-in hybrid vehicles, fuel cell
vehicles, and compressed natural gas vehicles. CO2 emissions from alternative fuel vehicles
represent tailpipe emissions, and fuel economy for these vehicles is reported as mpge
(miles per gallon of gasoline equivalent), or the miles an alternative fuel vehicle can travel
on an amount of energy equivalent to that in a gallon of gasoline. Sales data prior to MY
2011 are included in some cases based on available industry reports (e.g., Ward's
Automotive data).
Changes in Vehicle Classification Definitions
The car-truck classifications in this report follow the current regulatory definitions used by
EPA and NHTSAfor compliance with GHG emissions and CAFE standards (see definitions
for passenger automobiles (cars) and non-passenger automobiles (trucks) in 49 CFR 523).
These current definitions differ from those used in the 2010 and older versions of the Light-
Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy Trends report, and
reflect a decision by NHTSA to reclassify many small, 2-wheel drive sport utility vehicles
(SUVs) from the truck category to the car category, beginning with model year 2011. When
this re-classification was initiated in the 2011 report, the absolute truck share decreased by
approximately 10%.
The current car-truck definitions have been propagated back throughout the entire
historical Trends database to maintain the integrity of long-term trends of car and truck
production share. Since the authors did not have all of the requisite technical information
on which to make retroactive car-truck classifications, we used engineering judgment to
classify past models.
This report previously presented data on more vehicle types, but recent vehicle design has
led to far less distinction between vehicle types and reporting on more disaggregated
vehicle types was no longer useful.
Manufacturer Definitions
When a manufacturer grouping changes under the GHG and CAFE programs, the current
manufacturer definitions are generally applied to all prior model years. This maintains
D-6
-------
consistent manufacturer and make definitions over time, which enables better
identification of long-term trends. However, some of the compliance data maintain the
previous manufacturer definitions where necessary to preserve the integrity of compliance
data as they were accrued.
Differences in Production Data Between CAFE and GHG Regulations
The data used to discuss real-world trends in Sections 1 through 4 of this report are based
on production volumes reported under CAFE prior to model year 2017, not the GHG
standards. The production volume levels automakers provide in their final CAFE reports
may differ slightly from their final GHG reports (typically less than 0.1 %) because of
different reporting requirements. The EPA regulations require emission compliance in the
50 states, the District of Columbia, Puerto Rico, the Virgin Islands, Guam, American Samoa,
and the Commonwealth of the Northern Mariana Islands, whereas the CAFE program
requires data from the 50 states, the District of Columbia, and Puerto Rico only. All
compliance data detailed in Section 5, for all years, are based on production volumes
reported under the GHG standards. Starting with model year 2017 and forward, the real-
world data are also based on production volumes reported under EPA's GHG standards. As
described above, the difference in production volumes is very small and does not impact
the long-term trends or analysis.
D-7
-------
E. Electric Vehicle and Plug-In Hybrid Metrics
Electric Vehicles (EVs) and Plug-in Hybrid Vehicles (PHEVs) have continued to gain market
share. While overall market penetration of these vehicles is still relatively low, their
production share is projected to reach 4% in model year 2021. This section addresses some
of the technical metrics used both to quantify EV and PHEV operation and to integrate data
from these vehicles with gasoline and diesel vehicle data.
EVs operate using only energy stored in a battery from external charging. PHEVs blend EV
technology with more familiar powertrain technology from petroleum-fueled vehicles.
Current PHEVs feature both an electric drive system designed to be charged from an
electricity source external to the vehicle (like an EV) and a gasoline internal combustion
engine. There are generally three ways that a PHEV can operate:
• Charge-depleting electric-only mode - In electric-only mode the vehicle operates
like an EV, using only energy stored in the battery to propel the vehicle.
• Charge-depleting blended mode - In blended mode the vehicle uses both energy
stored in the battery and energy from the gasoline tank to propel the vehicle.
Depending on the vehicle design and driving conditions, blended operation can
include substantial all-electric driving.
• Charge-sustaining mode - In charge-sustaining mode, the PHEV has exhausted the
external energy from the electric grid that is stored in the battery and relies on the
gasoline internal combustion engine. In charge-sustaining mode, the vehicle will
operate much like a traditional hybrid.
The presence of both electric drive and an internal combustion engine results in a complex
system that can be used in many different combinations, and manufacturers are choosing
to operate PHEV systems in different ways. This complicates direct comparisons among
PHEV models.
This section discusses EV and PHEV metrics for several example model year 2021 vehicles.
For consistency and clarity for the reader, the data for specific vehicles discussed in this
section reflect values from the EPA/DOT Fuel Economy and Environment Labels, which use
a 55% city and 45% highway weighting for combined fuel economy and CO2 values. When
data for these vehicles are integrated into the data for the rest of the report, the real-world
highway and city values are combined using a 43% city and 57% highway weighting.
Additionally, some PHEV calculations are also adjusted, as explained at the end of this
section.
E-1
-------
Table E.1 shows the label driving range for several EVs and PHEVs when operating only on
electricity, as well as the total electricity plus gasoline range for PHEVs. The average range
of new EVs is increasing, as shown in Section 4, and many EVs are approaching the range of
an average gasoline vehicle.28 PHEVs generally have a much smaller all electric range,
however the combined electric and gasoline range for PHEVs often exceeds gasoline-only
vehicles. Several PHEVs now exceed 500 miles of total range.
Table E.1. Model Year 2021 Example EV and PHEV Powertrain and Range

Electric
Total

Fuel or
Range
Range
Utility
Manufacturer
Model
Powertrain
(miles)
(miles)
Factor
GM
Bolt
EV
259
259
-
Nissan
Leaf 62 kWh
EV
226
226
-
Tesla
Model 3 LR
EV
353
353
-
Ford
Escape
PHEV
37
520
0.66
Honda
Clarity
PHEV
48
340
0.73
Stellantis
Pacifica
PHEV
32
520
0.61
Toyota
Prius Prime
PHEV
25
640
0.53
Volvo
XC90
PHEV
18
520
0.43
Determining the electric range of PHEVs is complicated if the vehicle can operate in
blended modes. For PHEVs like the Ford Escape, which cannot operate in blended mode,
the electric range represents the estimated range operating in electric only mode.
However, for PHEVs that operate in a blended mode, the electric range represents the
estimated range of the vehicle operating in either electric only or blended mode, due to the
design of the vehicle. For example, the Volvo XC90 uses electricity stored in its battery and
a small amount of gasoline to achieve an alternative fuel range of 18 miles. Some PHEVs
did not use any gasoline to achieve their electric range value on EPA test cycles; however,
certain driving conditions (e.g., more aggressive accelerations, higher speeds, and air
conditioning or heater operation) would likely cause these vehicles to operate in a blended
mode instead of an all-electric mode.
Table E.1 also introduces the concept of a utility factor. The utility factor is directly related
to the electric range for PHEVs, and is a projection, on average, of the percentage of miles
that will be driven using electricity (in electric-only and blended modes) by an average
28 In addition to growing EV range, the number of public electric vehicle charging stations is growing rapidly. For
more information, see the U.S. Department of Energy's Alternative Fuels Data Center at
https://www.afdc.energy.gov/.
-------
driver. The model year 2021 Escape, for example, has a utility factor of 0.66, i.e., it is
expected that, on average, the Escape will operate 66% of the time on electricity and 34%
of the time on gasoline. Utility factor calculations are based on an SAE methodology that
EPA has adopted for regulatory compliance (SAE 2010).
Table E.2 shows five energy-related metrics for model year 2021 example EVs and PHEVs
that are included on the EPA/NHTSA Fuel Economy and Environment labels. Comparing the
energy or fuel efficiency performance from alternative fuel vehicles raises complex issues
of how to compare different fuels. Consumers and OEMs are familiar and comfortable with
evaluating gasoline and diesel vehicle fuel economy in terms of miles per gallon, and it is
the primary efficiency metric in this report. To enable this comparison for alternative fuel
vehicles, the overall energy efficiency of vehicles operating on electricity, hydrogen, and
CNG are evaluated in terms of miles per gallon of gasoline equivalent (an energy metric
described in more detail below).
Table E.2. Model Year 2021 Example EV and PHEV Fuel Economy Label Metrics
Charge
Manufacturer
Model
Fuel
or
Power
-train
Charge Depleting
Electricity Gasoline Fuel
(kW-hrs/ (gallons/ Economy
100 miles) 100 miles) (mpge)
Sustaining
Fuel
Economy
(mpg)
Overall
Fuel
Economy
(mpge)
GM
Bolt
EV
29
-
118
N/A
118
Nissan
Leaf 62 kWh
EV
31
-
108
N/A
108
Tesla
Model 3 LR
EV
25
-
134
N/A
134
Ford
Escape
PHEV
32
0.0
105
41
67
Honda
Clarity
PHEV
31
0.0
110
42
76
Stella ntis
Pacifica
PHEV
41
0.0
82
30
48
Toyota
Prius Prime
PHEV
25
0.0
133
54
78
Volvo
XC90
PHEV
58
0.1
55
27
34
The fourth column in Table E.2 gives electricity consumption rates for EVs and PHEVs
during charge depleting operation in units of kilowatt-hours per 100 miles (kW-hrs/100
miles). As shown on the vehicle label, the electricity consumption rate is based on the
amount of electricity required from an electric outlet to charge the vehicle and includes
wall-to-vehicle charging losses. The values for all of the EVs and PHEVs reflect the electricity
consumption rate required to operate the vehicle in either electric-only or blended mode
operation. PHEVs that are capable of operating in a blended mode may also consume
some gasoline in addition to electricity. Any additional gasoline used is shown in the fifth
column. For example, the Volvo XC90 PHEV consumes 58 kW-hrs and 0.1 gallons of
gasoline per 100 miles during this combination of electric-only and blended modes.
}j§ E-3^
-------
The sixth column converts the electricity consumption data in the fourth column and the
gasoline consumption data in the fifth column into a combined miles per gallon of
gasoline-equivalent (mpge) metric. The mpge metric is a measure of the miles the vehicle
can travel on an amount of energy that is equal to the amount of energy stored in a gallon
of gasoline. For a vehicle operating on electricity, mpge is calculated as 33.705 kW-
hrs/gallon divided by the vehicle electricity consumption in kW-hrs/mile. For example, for
the Leaf, 33.705 kW-hrs/gallon divided by 0.31 kW-hrs/mile (equivalent to 31 kW-hrs/100
miles) is 108 mpge.29 Because the Volvo XC90 consumes both electricity and gasoline over
the alternative fuel range of 18 miles, the charge depleting fuel economy of 55 mpge
includes both the electricity and gasoline consumption, at a rate of 58 kW-hrs/100 miles of
electricity and 0.1 gal/100 miles of gasoline.
The seventh column gives label fuel economy values for vehicles operating on gasoline
only, which is relevant here only for the PHEVs operating in charge sustaining mode. For
PHEVs, the EPA/NHTSA label shows both electricity consumption in kW-hrs/100 miles and
mpge, when the vehicle operates exclusively on electricity or in a blended mode, and
gasoline fuel economy in mpg, when the vehicle operates exclusively on gasoline.
The final column gives the overall mpge values reflecting the overall energy efficiency of
the vehicle for all of the fuels on which the vehicle can operate, and provide a common
metric to compare vehicles that operate on different fuels. In addition to the energy
metrics in the previous columns, the one key additional parameter necessary to calculate a
combined electricity/gasoline mpge value for a PHEV is the utility factor that was
introduced in Table E.1. For EVs the overall fuel economy in the last column is equal to the
charge depleting fuel economy, as EVs can only operate in a charge depleting mode.
Table E.3 gives vehicle tailpipe C02 emissions values that are included on the EPA/DOT Fuel
Economy and Environment labels (and reflected in the label's Greenhouse Gas Rating).
These label values reflect EPA's best estimate of the C02 tailpipe emissions that these
vehicles will produce, on average, in real-world city and highway operation. EVs, of course,
have no tailpipe emissions. For the PHEVs, the label CO2 emissions values utilize the same
utility factors discussed above to weight the CO2 emissions on electric and gasoline
operation.
29 The actual calculations were done with unrounded numbers. Using the rounded numbers provided here may
result in a slightly different number due to rounding error.
-------
Table E.3. Model Year 2021 Example EV and PHEV Label Tailpipe C02 Emissions
Metrics

Fuel or
Tailpipe C02
Manufacturer
Model
Powertrain
(g/mile)
GM
Bolt
EV
0
Nissan
Leaf 62 kWh
EV
0
Tesla
Model 3 LR
EV
0
Ford
Escape
PHEV
77
Honda
Clarity
PHEV
57
Stellantis
Pacifica
PHEV
119
Toyota
Prius Prime
PHEV
78
Volvo
XC90
PHEV
197
Table E.4 accounts for the "upstream" C02 emissions associated with the production and
distribution of electricity used in EVs and PHEVs. Gasoline and diesel fuels also have C02
emissions associated with their production and distribution, but these upstream emissions
are not reflected in the tailpipe C02 emissions values discussed elsewhere in this report.
Combining vehicle tailpipe and fuel production/distribution sources, gasoline vehicles emit
about 80 percent of total C02 emissions at the vehicle tailpipe with the remaining 20
percent of total CO2 emissions associated with upstream fuel production and distribution.
Diesel fuel has a similar approximate relationship between tailpipe and upstream CO2
emissions. On the other hand, vehicles powered by grid electricity emit no CO2 (or other
emissions) at the vehicle tailpipe; therefore, all CO2 emissions associated with an EV are
due to fuel production and distribution. Depending on how the electricity is produced,
these fuels can have very high fuel production/distribution CO2 emissions (for example, if
coal is used with no C02 emissions control) or very low C02 emissions (for example, if
renewable processes with minimal fossil energy inputs are used).
Electricity production in the United States varies significantly from region to region and has
been changing over time. Hydroelectric plants provide a large percentage of electricity in
the Northwest, while coal-fired power plants produce the majority of electricity in the
Midwest. Natural gas, wind, and solar have increased their electricity market share in many
regions of the country. Nuclear power plants make up most of the balance of U.S.
electricity production. In order to bracket the possible GHG emissions impact, Table E.4
provides ranges with the low end of the range corresponding to the California power plant
GHG emissions factor, the middle of the range represented by the national average power
E-5
-------
plant GHG emissions factor, and the upper end of the range corresponding to the power
plant GHG emissions factor for part of the Midwest (Illinois and Missouri).
Table E.4. Model Year 2021 Example EV and PHEV Upstream C02 Emission
Metrics (g/mi)
Tailpipe + Total Tailpipe + Net

Fuel or
Upstream C02

Upstream C02

Manufacturer
Model
Powertrain
Low
Avg
High
Low
Avg
High
GM
Bolt
EV
67
128
222
15
76
170
Nissan
Leaf 62 kWh
EV
74
139
243
19
84
188
Tesla
Model 3 LR
EV
59
112
195
-2
51
134
Ford
Escape
PHEV
146
191
261
90
135
205
Honda
Clarity
PHEV
124
171
244
68
115
189
Stella ntis
Pacifica
PHEV
208
261
344
124
177
261
Toyota
Prius Prime
PHEV
129
157
202
81
109
153
Volvo
XC90
PHEV
305
359
440
220
273
355
Average Sedan/Wagon

344
344
344
275
275
275
Based on data from EPA's eGRID power plant database,30 and accounting for additional
greenhouse gas emissions impacts for feedstock processing upstream of the power
plant,31 EPA estimates that the electricity CO2 emission factors for various regions of the
country vary from 235 g CCVkW-hr in California to 775 g CCVkW-hr in the Midwest, with a
national average of 445 g CCVkW-hr. Emission rates for small regions in upstate New York
and Alaska have lower electricity upstream CO2 emission rates than California. However,
California is a good surrogate for the "low" end of the range because California is a leading
market for current EVs and PHEVs. Initial sales of electric vehicles have been largely, though
not exclusively, focused in regions of the country with power plant C02 emissions factors
lower than the national average, such as California, New York, and other coastal areas.
Accordingly, in terms of C02 emissions, EPA believes that the current "sales-weighted
average" vehicle operating on electricity in the near term will likely fall somewhere between
the low end of this range and the national average.32
30 United States Environmental Protection Agency (EPA). 2021. "Emissions & Generation Resource Integrated
Database (eGRID), 2019" Washington, DC: Office of Atmospheric Programs, Clean Air Markets Division. Available
from EPA's eGRID web site: https://www.epa.gov/egrid.
31 Argonne National Laboratory 2019. GREET_1_2020 Model, greet.es.anl.gov.
32 To estimate the upstream greenhouse gas emissions associated with operating an EV or PHEV in a specific
geographical area, use the emissions calculator at www.fueleconomy.gov/feg/Find.do?action=bt2.
E-6
-------
The fourth through sixth columns in Table E.4 provide the range of tailpipe plus total
upstream C02 emissions for EVs and PHEVs based on regional electricity emission rates.
For comparison, the average model year 2021 car is also included in the last row of Table
E.4. The methodology used to calculate the range of tailpipe plus total upstream C02
emissions for EVs is shown in the following example for the model year 2021 Nissan Leaf
(62 kWh battery):
• Start with the label (5-cycle values weighted 55% city / 45% highway) vehicle
electricity consumption in kW-hr/mile, which for the Leaf is 31 kW-hr/100 miles, or
0.31 kW-hr/mile
• Determine the regional powerplant emission rate, regional losses during electricity
distribution, and the additional regional emissions due to fuel production upstream
of the powerplant (for California, these numbers are 206 g/kW-hr, 5.1%, and 8.6%,
respectively).
• Determine the regional upstream emission factor (for California 206 g/kW-hr / (1 -
0.051) * (1 +0.086) = 235 g C02/kW-hr)33
• Multiply by the range of Low (California = 235g C02/kW-hr), Average (National
Average = 445 g C02/kW-hr), and High (Midwest = 775 g C02/kW-hr) electricity
upstream C02 emission rates, which yields a range for the Leaf of 74-243 grams
C02/mile.
The tailpipe plus total upstream C02 emissions values for PHEVs include the upstream C02
emissions due to electricity operation and both the tailpipe and upstream C02 emissions
due to gasoline operation, using the utility factor discussed above to weight the values for
electricity and gasoline operation. The tailpipe plus total upstream C02 emissions values
for the average car are the average projected real-world model year 2021 car tailpipe C02
emissions multiplied by 1.25 to account for upstream emissions due to gasoline
production.
The values in columns four through six are tailpipe plus total upstream C02 emissions. As
mentioned, all of the gasoline and diesel vehicle C02 emissions data in the rest of this
report refer only to tailpipe emissions and do not reflect the upstream emissions
associated with gasoline or diesel production and distribution. Accordingly, in order to
equitably compare the overall relative impact of EVs and PHEVs with tailpipe emissions of
petroleum-fueled vehicles, EPA uses the metric "tailpipe plus net upstream emissions" for
33The actual calculations were done with unrounded numbers. Using the rounded numbers provided here may
result in a slightly different number due to rounding error.
E-7
-------
EVs and PHEVs. The net upstream emissions for an EV is equal to the total upstream
emissions for the EV minus the upstream emissions that would be expected from a
comparably sized gasoline vehicle; size is a good first-order measure for utility, and
footprint is the size-based metric used for standards compliance. The net upstream
emissions for PHEVs are equal to the net upstream emissions of the PHEV due to electricity
consumption in electric or blended mode multiplied by the utility factor. The net upstream
emissions for a gasoline vehicle are zero. This approach was adopted for EV and PHEV
regulatory compliance with the 2012-2016 light-duty vehicle GHG emissions standards for
the production of EVs and PHEVs beyond a threshold; however, those thresholds were
never exceeded.
For each EV or PHEV, the upstream emissions for a comparable gasoline vehicle are
determined by first using the footprint-based compliance curves to determine the C02
compliance target for a vehicle with the same footprint. Since upstream emissions account
for approximately 20% of total C02 emissions for gasoline vehicles, the upstream emissions
for the comparable gasoline vehicle are equal to one-fourth of the tailpipe-only compliance
target.
The final three columns of Table E.4 give the tailpipe plus net upstream CO2 values for EVs
and PHEVs using the same Low, Average, and High electricity upstream CO2 emissions rates
discussed above. These values bracket the possible real-world net CO2 emissions that
would be associated with consumer use of these vehicles. For the Leaf, these values are
simply the values in columns four through six minus the upstream GHG emissions of a
comparably sized gasoline vehicle. Based on the model year 2021 C02 footprint curve, the
5-cycle tailpipe GHG emissions for a Leaf-sized gasoline vehicle meeting its compliance
target would be close to 220 grams/mi, with upstream emissions of one-fourth of this
value, or 55 g/mi. The net upstream emision for a Leaf (with the 62 kWh battery) are
determined by subtracting this value, 55 g/mi, from the total (tailpipe + total upstream).
The result is a range for the tailpipe plus net upstream value of 19-188 g/mile as shown in
Table E.4, with a more likely sales-weighted value in the 19-84 g/mi range.
For PHEVs, the tailpipe plus net upstream emissions values use the utility factor values
discussed above to weight the individual values for electric operation and gasoline
operation.
Alternative Metrics for EVs and PHEVs
Determining metrics for EVs and PHEVs that are meaningful and accurate is challenging. In
particular, vehicles capable of using dual fuels, such as PHEVs, can have complicated
E-8
-------
modes of operation that make it difficult to determine meaningful metrics. Here we've
discussed several metrics that are used on the EPA/DOT Fuel Economy and Environment
Labels and in a regulatory context, namely mpge, tailpipe C02 emissions, and net upstream
GHG emissions. There are, however, other ways that alternative fuel vehicle operation can
be quantified.
Other energy metric options that could be considered include: (1) mpge plus net fuel life
cycle energy, which would also reflect differences in upstream energy consumption in
producing the alternative fuel relative to gasoline-from-oil; and (2) miles per gallon of
gasoline, which would only count gasoline use and not other forms of energy. Compared to
mpge, using the mpge plus net fuel life-cycle energy metric would generally result in lower
fuel economy values, and using the miles per gallon of gasoline metric would yield higher
fuel economy values.
Additional Note on PHEV Calculations
Calculating fuel economy and C02 emission values for PHEVs is a complicated process, as
discussed in this section. The examples given for individual vehicles were based on
calculations behind the EPA/DOT Fuel Economy and Environment Labels. In addition to the
approach used for the labels, there are multiple methods for determining utility factors
depending on the intended use of the value. The standardized utility factor calculations are
defined in the Society of Automobile Engineers (SAE) document SAE J2841.
The utility factors that are used for fleetwide calculations are somewhat different than
those used to create label values. For label values, multi-day individual utility factors
(MDIUF) are used to incorporate "a driver's day to day variation into the utility calculation."
For fleetwide calculations, fleet utility factors (FUF) are applied to "calculate the expected
fuel and electric consumption of an entire fleet of vehicles." Since the Trends report is
generally a fleetwide analysis, the FUF utility factors were applied, instead of the MDIUF
utility factors, when the data were integrated with the rest of the fleet data. Additionally,
since Trends uses a 43% city / 57% highway weighting for combining real-world fuel
economy and C02 data, the FUF utility factors created for Trends were based on that
weighting, not on 55% city / 45% highway weighting used on the fuel economy label.
i
-------
F. Authors and Acknowledgments
The authors of this year's Trends report are Aaron Hula, Andrea Maguire, Amy Bunker,
Tristan Rojeck, and Sarah Harrison, all of whom work for the EPA Office of Transportation
and Air Quality (OTAQ) at the National Vehicle and Fuel Emissions Laboratory in Ann
Arbor, Michigan. OTAQ colleagues including Karen Danzeisen, Robert Peavyhouse, and
Ching-Shih Yang provided critical access and expertise pertaining to the EV-CIS data that
comprise the Trends database. The authors also want to thank Gwen Dietrich and Eloise
Anagnost of OTAQ for greatly improving the design and layout of the report. General
Dynamics Information Technology (GDIT) under contract to OTAQ (contract number
EP-C-16-012), provided key support for database maintenance, and table and figure
generation. DOT/NHTSA staff reviewed the report and provided helpful comments. Of
course, the EPA authors take full responsibility for the content and any errors.
The authors also want to acknowledge those OTAQ staff that played key roles in creating
and maintaining the Trends database and report since its inception in the early 1970s.
Karl Hellman, who conceived of and developed the initial Trends reports with Thomas
Austin in the early 1970s, was the guiding force behind the Trends report for over 30 years.
Dill Murrell made significant contributions from the late 1970s through the early 1990s,
and Robert Heavenrich was a lead author from the early 1980s through 2006. Jeff Alson
oversaw the continued transformation and modernization of this report from 2007
through 2018. The compliance portion of this report (now section 5), was developed by
Roberts French, and he remained the lead author through the 2019 report. This report has
benefitted immensely from the wealth of insight, creativity, and dedication from each of
these outstanding emeritus authors.
F-1
-------