Draft Joint Technical Support Document
Proposed Rulemaking to Establish Light-
Duty Vehicle Greenhouse Gas Emission
Standards and Corporate Average Fuel
Economy Standards
United States
Environmental Protection
Agency
NHTSA
-------
Draft Joint Technical Support Document
Proposed Rulemaking to Establish Light-
Duty Vehicle Greenhouse Gas Emission
Standards and Corporate Average Fuel
Economy Standards
Assessment and Standards Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
and
Office of International Policy, Fuel Exonomy, and Consumer Programs
National Highway Traffic Safety Administration
U.S. Department of Transportation
S EDA Uniled sua> tk.tt-.rrf n.
f> trA s1proleclion NHTSA
vvww. iihtsa.ov
EPA-420-D-09-901
September 2009
-------
Table of Contents
CHAPTER 1: THE BASELINE VEHICLE FLEET
1.1 Why do the agencies establish a baseline and reference vehicle fleet? 1-1
1.2 The 2008 baseline vehicle fleet 1-2
1.2.1 Why did the agencies choose 2008 as the baseline model year? 1-2
1.2.2 On what data is the baseline vehicle fleet based? 1-2
1.3 The MY 2011-2016 Reference Fleet 1-13
1.3.1 On what data is the baseline vehicle fleet based? 1-13
1.3.2 How do the agencies develop the reference vehicle fleet? 1-15
1.3.3 What are the sales volumes and characteristics of the reference fleet? 1-24
1.3.4 How is the development of the baseline fleet for this proposal different from
NHTSA's historical approach and why is this approach preferable? 1-26
1.3.5 How does manufacturer product plan data factor into the baseline used in this
proposal? 1-31
CHAPTER 2: WHAT ARE THE ATTRIBUTE-BASED CURVES THE AGENCIES ARE
USING, AND HOW WERE THEY DEVELOPED?
2.1 Standards are attribute-based and defined by a mathematical function 2-1
2.2 What attribute do the agencies use, and why? 2-4
2.3 What mathematical function do the agencies use, and why? 2-5
CHAPTER 3: TECHNOLOGIES CONSIDERED IN THE AGENCIES'ANALYSIS
3.1 How do the agencies decide which technologies to include in the analysis? 3-1
3.1.1 Reports and papers in the literature 3-1
3.1.2 Fuel economy certification data 3-2
3.2 Which technologies will be applicable in the rulemaking time frame? 3-3
3.3 What technology assumptions have the agencies used for this proposal? 3-7
3.3.1 How are the technologies applied in the agencies' respective models? 3-7
3.3.2 How did the agencies develop technology cost and effectiveness estimates for the
proposal? 3-9
3.4 Specific technologies considered and estimates of costs and effectiveness 3-20
3.4.1 What data sources did the agencies evaluate? 3-20
3.4.2 Individual technology descriptions and cost/effectiveness estimates 3-20
3.5 Cost and effectiveness tables 3-81
3.5.1 NHTSA cost and effectiveness tables 3-82
3.5.2 EPA cost and effectiveness tables 3-89
Appendix 3.A 3-94
CHAPTER 4: EONOMIC ASSUMPTIONS USED IN THE AGENCIES' ANALYSES
4.1 How the Agencies Use the Economic Assumptions in their Analyses 4.1
4.2 What Economic Assumptions Do the Agencies Use? 4-2
4.2.1 Potential Opportunity Costs of Improved Fuel Economy and Reduced GHG
Emissions 4-2
4.2.2 The On-Road Fuel Economy "Gap" 4-3
4.2.3 Fuel prices and the Value of Saving Fuel 4-5
-------
4.2.4 Vehicle Survival and Use Assumptions 4-7
4.2.5 Accounting for the Fuel Economy Rebound Effect 4-16
4.2.6 Benefits from Increased Vehicle Use 4-22
4.2.7 Added Costs from Increased Vehicle Use 4-22
4.2.8 Petroleum and Energy Security Impacts 4-23
4.2.9 Air Pollutant Emissions 4-33
4.2.10 Reductions in Emissions of Greenhouse Gases 4-43
4.2.11 The Value of Increased Driving Range 4-44
4.2.12 Discounting future Benefits and Costs 4-46
-------
List of Acronyms
CAFE: Corporate Average Fuel Economy
CO2: carbon dioxide
DICE: Dynamic Integrated Model of Climate and the Economy
DOT: Department of Transportation
EIS: Environmental Impact Statement
FUND:Climate Framework for Uncertainty, Negotiation, and Distribution
GHG: Greenhouse gas
IAM: Integrated Assessment Model
IPCC: Intergovernmental Panel on Climate Change
MY: Model Year
OECD: Organization for Economic Cooperation and Development
OMB: Office of Management and Budget
PAGE: Policy Analysis for the Greenhouse Effect
PRTP: Pure Rate of Time Preference
S&P: Standard and Poor's
SCC: Social Cost of Carbon
SRES: Special Report on Emissions Scenarios
THC: Thermohaline circulation
WGII: Working group II
ORNL: Oak Ridge National Laboratory
DOHC: Dual Overhead Camshafts
SOHC: Single Overhead Camshaft
OHV: Overhead Valve (pushrod)
CCP: Couple Cam Phasing
DCP: Dual (independent) Cam Phasing
DVVL: Discrete (two-step) Variable Valve Lift
CVVL: Continuous Variable Valve Lift
Deac: Cylinder Deactivation
CVA: Camless Valve Actuation (full)
Turbo: Turbocharging and engine downsizing
GDI: Gasoline Direct Injection
Diesel: Diesel with advanced aftertreatment
HCCI: Homogenous Charge Compression Ignition (gasoline)
LUB: Low-friction engine lubricants
EFR: Engine Friction Reduction
L4: Lock-up 4-speed automatic transmission
L5: Lock-up 5-speed automatic transmission
L6: Lock-up 6-speed automatic transmission
DCT6: 6-speed Dual Clutch Transmission
CVT: Continuously Variable Transmission
ASL: Aggressive Shift Logic
-------
This Page Intentionally Left Blank
-------
Executive Summary
The Environmental Protection Agency (EPA) and the National Highway Traffic
Safety Administration (NHTSA) are issuing a joint proposal to establish new standards
for light-duty highway vehicles that will reduce greenhouse gas emissions and improve
fuel economy. The joint proposed rulemaking is consistent with the National Fuel
Efficiency Policy announced by President Obama on May 19, 2009, responding to the
country's critical need to address global climate change and to reduce oil consumption.
EPA is proposing greenhouse gas emissions standards under the Clean Air Act, and
NHTSA is proposing Corporate Average Fuel Economy standards under the Energy
Policy and Conservation Act, as amended. These standards apply to passenger cars,
light-duty trucks, and medium-duty passenger vehicles, covering model years 2012
through 2016. They require these vehicles to meet an estimated combined average
emissions level of 250 grams of CO2 per mile in MY 2016 under EPA's GHG program,
and 34.1 mpg in MY 2016 under NHTSA's CAFE program and represent a harmonized
and consistent national program (National Program). These standards are designed such
that compliance can be achieved with a single national vehicle fleet whose emissions and
fuel economy performance improves each year from MY2012 to 2016. This document
describes the supporting technical analysis for areas of these jointly proposed rules which
are consistent between the two agencies.
NHTSA and EPA have coordinated closely to create a nationwide joint fuel
economy and GHG program based on consistent compliance structures and technical
assumptions. To the extent permitted under each Agency's statutes, NHTSA and EPA
have incorporated the same compliance flexibilities, such as averaging, banking, and
trading of credits, and the same testing protocol for determining the agencies' respective
fleet-wide average proposed standards. In addition, the agencies have worked together to
create a common baseline fleet and to harmonize most of the costs and benefit inputs
used in the agencies' respective modeling processes for this joint proposed rule.
Chapter 1 of this Draft TSD provides an explanation of the agencies' new
methodology used to develop the baseline and reference case vehicle fleets, including the
technology composition of these fleets, and how the agencies projected vehicle sales into
the future. One of the fundamental features of this technical analysis is the development
of these fleets, which are used by both agencies in their respective models. In order to
determine technology costs associated with this joint rulemaking, it is necessary to
consider the vehicle fleet absent a rulemaking as a "business as usual" comparison. In
past CAFE rulemakings, NHTSA has used confidential product plans submitted by
vehicle manufacturers to develop the reference case fleet. In responding to comments
from these previous rulemakings that the agencies make these fleets available for public
review, the agencies created a new methodology for creating baseline and reference fleets
using data, the vast majority of which is publicly available.
-------
Chapter 2 of this document discusses how NHTSA and EPA developed the
mathematical functions which provide the bases for manufacturers' car and truck
standards. NHTSA and EPA worked together closely to develop regulatory approaches
that are fundamentally the same, and have chosen to use an attribute-based program
structure based on the footprint attribute, like NHTSA's current Reformed CAFE
program. The agencies revisited other attributes as candidates for the standard functions,
but concluded that footprint remains the best option for balancing the numerous technical
and social factors. However, the agencies did adjust the shape of the footprint curve, in
contrast to the 2011 CAFE rule, the CO2 or fuel consumption curve is a piecewise linear
or constrained linear function, rather than a constrained logistic function. In determining
the shape of the footprint curve, the agencies considered factors such as the magnitudes
of CO2 reduction and fuel savings, how much that shape may entice manufacturers to
comply in a manner which circumvents the overall goals of the joint program, whether
the standards' stringencies are technically attainable, and the mathematical flexibilities
inherent to such a function
Chapter 3 contains a detailed analysis of NHTSA and EPA's technology
assumptions on which the proposed regulations were based. Because the majority of
technologies that reduce GHG emissions and improve fuel economy are identical, it was
crucial that NHTSA and EPA use common assumptions for values pertaining to
technology availability, cost, and effectiveness. The agencies collaborated closely in
determining which technologies would be considered in the rulemaking, how much these
technologies would cost the manufacturers (directly) in the time frame of the proposed
rules, how these costs will be adjusted for learning as well as for indirect cost multipliers,
and how effective the technologies are at accomplishing the goals of improving fuel
efficiency and GHG emissions.
Chapter 4 of this TSD provides a full description and analysis of the economic
factors considered in this joint proposal. EPA and NHTSA harmonized many of the
economic and social factors, such as the discount rates, fuel prices, social costs of carbon,
the magnitude of the rebound effect, the value of refueling time, and the social cost of
importing oil and fuel.
-------
The Baseline Vehicle Fleet
CHAPTER 1: The Baseline Vehicle Fleet
The passenger cars and light tracks sold currently in the United States, and those
which are anticipated to be sold in the MY 2012-2016 timeframe, are highly varied and
satisfy a wide range of consumer needs. From two-seater miniature cars to 11-seater
passenger vans to large extended cab pickup trucks, American consumers have a great
number of vehicle options to accommodate their utility needs and preferences. Recent
volatility in oil prices and the state of the economy have demonstrated that consumer
demand and choice can be sensitive to these factors. Although it is impossible for anyone
or any organization to precisely predict the future, a characterization and quantification
are required for the future fleet to assess impacts of rules which would affect that future
fleet. In order to do this, the various leading publically available sources are examined,
and a series of models are relied upon that help us to project the composition of a
reference fleet. This chapter describes the process for accomplishing this.
1.1 Why do the agencies establish a baseline and reference vehicle fleet?
In order to calculate the impacts of the EPA and NHTSA proposed regulations, it is
necessary to estimate the composition of the future vehicle fleet absent these proposed
regulations in order to conduct comparisons. EPA in consultation with NHTSA has
developed a comparison fleet in two parts. The first step was to develop a baseline fleet
based on model year 2008 data. EPA and NHTSA create a baseline fleet in order to track
the volumes and types of fuel economy-improving and COi-reducing technologies which
are already present in today's fleet. Creating a baseline fleet helps to keep, to some
extent, the agencies' models from adding technologies to vehicles that already have these
technologies, which would result in "double counting" of technologies' costs and
benefits. The second step was to project the baseline fleet sales into MYs 2011-2016.
This is called the reference fleet, and it represents the fleet that would exist in MYs 2011-
2016 absent any change from current regulations. The third step was to add technologies
to that fleet such that each manufacturer's average car and truck COi levels are in
compliance with their MY 2011 CAFE standards. This final "reference fleet" is the light
duty fleet estimated to exist in MYs 2012-2016 without these proposed rales. All of the
agencies' estimates of emission reductions/fuel economy improvements, costs, and
societal impacts are developed in relation to the respective reference fleets. The chapter
describes the first two steps of the development of the baseline and reference fleets. The
third step of technology addition is developed separately by each agency as the outputs of
the OMEGA and Volpe models. The process is described in section II of the preamble
and in each agency's respective RIAs.
1-1
-------
Draft Joint Technical Support Document
1.2 The 2008 baseline vehicle fleet
1.2.1 Why did the agencies choose 2008 as the baseline model year?
For this proposed rule, the baseline vehicle fleet developed by EPA in
consultation with NHTSA and is comprised of model year 2008 data. MY 2008 was used
as the basis for the baseline vehicle fleet, because it is the most recent model year for
which full data is publicly available. Vehicle manufacturers have 90 days after their last
vehicle is produced to submit their CAFE data to EPA.1 Most manufacturers interpret
this to mean 90 days after the end of the calendar year. For example, in calendar year
2007, model year 2008 vehicles were tested and certified by the EPA. These MY 2008
vehicles were then sold in the latter part (often fall) of 2007 until the following fall of
2008. In early 2009 (calendar year), the manufacturers then submit their total sales of
MY 2008 vehicles. After these sales figures were submitted, EPA and NHTSA
combined the sales with the previously measured and reported fuel economies to
calculate the sales-weighted average fleet fuel economy. The analysis for this present
rulemaking was conducted in early to mid 2009 calendar year, thus the full sales figures
for MY 2009 vehicles have not yet been submitted. Even though the fuel economies (and
some other specifications) of the MY 2009 vehicles were known, since they were tested
earlier, the sales were not yet known for each company exactly. Therefore, the agencies
chose to use MY 2008 as the baseline since it was the most complete data set available.
NHTSA and EPA may update the baseline to MY 2009 or MY 2010 between this notice
and the final rule if the data is complete and if time permits.
1.2.2 On what data is the baseline vehicle fleet based?
As part of the CAFE program, EPA measures vehicle COi emissions and converts
them to mpg and generates and maintains the federal fuel economy database. Most of the
information about the 2008 vehicle fleet was gathered from EPA's emission certification
and fuel economy database, most of which is available to the public. The data obtained
from this source included vehicle production volume, fuel economy, carbon dioxide
emissions, fuel type, number of engine cylinders, displacement, valves per cylinder,
engine cycle, transmission type, drive, hybrid type, and aspiration. However, EPA's
certification database does not include a detailed description of the types of fuel
economy-improving/COi-reducing technologies considered in this proposal, because this
level of information is not necessary for emission certification or fuel economy testing.
Thus, the agency augmented this description with publicly-available data which includes
more complete technology descriptions from Ward's Automotive Group.2>a In a few
instances when required vehicle information was not available from these two sources
(such as vehicle footprint), this information was obtained from publicly-accessible
1 Note that WardsAuto.com is a fee-based service, but all information is public to subscribers.
1-2
-------
The Baseline Vehicle Fleet
internet sites such as Motortrend.com, Edmunds.com and other sources to a lesser extent
(such as articles about specific vehicles revealed from internet search engine research.3'15
The baseline vehicle fleet for the analysis in this rule is comprised of publicly-
available data to the largest extent possible. However, a few relatively low-impact
technologies were added based on confidential information provided from some
manufacturers (within their product plan submissions to NHTSA and EPA). This was
done because the data was not available from any other source. These technologies
include low friction lubricants, electric power steering, improved accessories, and low
rolling resistance tires. This confidential information has been excised from the baseline
data submitted to the docket, though the summary results are still used, so that any
specific information cannot be traced back to any specific manufacturer. This
discrepancy between the public baseline and the one used by the agencies is relatively
minor and should only result in small differences in the outputs of the Volpe and
OMEGA models for certain manufacturers.
Creating the 2008 baseline fleet Excel file was an extremely labor intensive
process. EPA in consultation with NHTSA first considered using EPA's CAFE
certification data, which contains most of the required information. However, since the
deadline for manufacturers to report this data did not allow enough time for early
modeling review, it was necessary to start this process using an alternative data source.
The agencies next considered using EPA's vehicle emissions certification data,
which contains much of the required information, however it lacked the production
volumes that are necessary for the OMEGA and Volpe models. The data set also
contains some vehicle models manufacturers have certified, but not produced. A second
data source which would supply production volumes and eliminate extraneous vehicles
was needed. Data from a paid subscription to Ward's Automotive Group was used as the
second source for data, which contains production volumes and vehicle specifications.
The vehicle emissions certification dataset came in two parts, an engine file and
a vehicle file. Since there was a common index in the two files, the engine and vehicle
data were easily combined into one spreadsheet. The agencies had hoped to supplement
this dataset with production volume data from Ward's Automotive Group but the Ward's
data does not have production volumes for individual vehicles down to the resolution of
the specific engine and transmission level. Although production volumes from Ward's
Automotive Group could not be used, the subscription did provide specific details on
individual vehicles and engines. The Ward's data used came in two parts (engine file and
vehicle file), and also required mapping. In this case, mapping was more difficult since
there was no common index between the two files. A new index was implanted in the
engine file and a search equation in the vehicle file, which identified most of the vehicle
and engine combinations. Each vehicle and engine combination was reviewed and
corrections were made manually when the search routine failed to give the correct engine
and vehicle combination. The combined Ward's data was then mapped to the vehicle
b Motortrend.com and Edmunds.com are free, no-fee internet sites.
1-3
-------
Draft Joint Technical Support Document
emissions certification data by creating a new index in the combined Ward's data and
using the same process that was used to combine the Ward's engine and vehicle files. In
the next step, CAFE certification data had to be merged in order to fill out the needed
production volumes.
NHTSA and EPA reviewed the CAFE certification data for model year 2008 as it
became available. The CAFE certification dataset could have been used with the Ward's
data without the vehicle emission certification dataset, but was instead appended to the
combined Ward's and vehicle certification dataset. The two former datasets were then
mapped into the CAFE dataset using the same Excel mapping technique described above.
Finally, EPA and NHTSA obtained the remaining attribute and technology data, such as
footprint, curb weight, and others (for a complete list of data with sources see Table 1-1
below) from other sources (such as the internet and the confidential product plan data),
thus completing the baseline dataset.
This was the first time a baseline fleet was created using this method. Given the
long delay before the CAFE certification data became available, EPA explored creating
the alternative dataset. It is possible to create the same baseline with CAFE certification
data, the Ward's engine data, a limited amount of product plan data, and some internet
searches.
Table 1-1 below shows the baseline data that was compiled, its definition
(description) and source. The data that is marked "not available" is data NHTSA would
normally get from product plans. Some of the desired model inputs, such as the presence
of low rolling resistance tires, reduced engine friction, improved accessories, are not
available from public sources and the agencies had to rely on manufacturer's confidential
product plans. The Technology Effectivenesses Basis and the Cost Effectiveness Basis
values reflect the percent of a technology package's effectiveness and cost present in the
baseline fleet, and they are described in further detail in chapters 3.1 and 3.5,
respectively. Those technologies that are not accounted for in the baselinethat is, the
ones marked "not available"run the risk of getting double counted by the agencies'
models, but those effects are expected to be small.
1-4
-------
The Baseline Vehicle Fleet
Table 1-1 Data, Definitions, and Sources
Data Item
Manufacturer
CERT Manufacturer Name
Name Plate
Model
Reg Class
Class
CSM Class
Traditional Car/Truck
EPA Interpretation of New NHTSA
Car/Truck
Total Production Volume
Fuel Econ.
(mpg)
Curb
Weight (Ib)
Area (sf)
Fuel
Disp
(lit.)
Effective Cyl
Actual Cylinders
Valves Per Cylinder
Valve Type
Engine Cycle
Horsepower
Torque
Trans Type
Trans
Num of Gears
Structure
Drive
Definition
Common name of company that manufactured
vehicle. May include more name plates than Cert
Manufacturer Name.
Certification name of company that manufactured
vehicle
Name of Division
Name of Vehicle
EPA Fuel Economy Class Name
If a car's Footprint<43 then "SubCmpctAuto"
If a car's 43<=Footprint<46 then "CompactAuto"
If a car's 46<=Footprint<53 then "MidSizeAuto"
If a car's Footprint >=53 then "LargeAuto"
If a S.U.V.'s Footprint < 43 then "SmallSuv"
If a S.U.V.'s 43<=Footprint<46 then
"MidSizeSuv"
If a S.U.V's Footprint >=46 then "LargeSuv"
If a Truck's Footprint < 50 then "SmallPickup"
If a Truck's Footprint>=50 then "LargPickup"
If a Van's Structure is Ladder then "Cargo Van"
If a Van's Structure is Unibody then "Minivan"
CSM Worldwides class for the vehicle. Used to
weight vehicles based on CSM data.
C= Car, T=Truck. As defined in the certification
database. Not used in calculations.
C= Car, T= Truck. EPA Best guess, some incorrect.
Using NHTSA below. Left in for historical
reasons. Not used in calculations.
Total number of vehicles produced for that model.
EPA Unadjusted Fuel Economy
Vehicles Curb Weight
Average Track x Wheelbase
Gas or Diesel
Engine Cylinder Displacement Size in Liters
Number of Cylinder + 2 if the engine has a turbo or
super charger.
Actual Number of Engine Cylinders
Number of Valves Per Actual Cylinder
Type of valve actuation.
As Defined by EPA Cert. Definition
Max. Horsepower of the Engine
Max. Torque of the Engine
A=Auto AMT=Automated Manual M=Manual
CVT= Continuously Variable Transmission
Type Code with number of Gears
Number of Gears
Ladder or Unibody
Fwd, Rwd, Awd, 4wd
Where The Data is From
Certification data
Certification data
Certification data
Certification data
Certification data
Derived From Certification
data and Footprint
CSM Worldwide
Certification data
EPA Derived from NHTSA
Definition
Certification data
Certification data
Some from Edmonds.com or
Motortrend.com, Others from
product plans with a subset
verified with Edmonds.com or
Motortrend.com for accuracy.
Calculated from track width
and wheel base
Wards
Wards/Certification data
Derived From Certification
data.
Certification data
Certification data
Wards (Note: Type E is from
Cert Data)
Wards
Wards
Wards
Certification data
Certification data
Certification data
General Internet Searches using
Google.com
Certification data
1-5
-------
Draft Joint Technical Support Document
Wheelbase
Track Width
(front)
Track Width
(rear)
Footprint
Curb
Weight
GVWR
HYBRID
HYBRID TYPE
Turbo
Super Charged
Turbo and/or Super Charged
SOHC
DOHC
OHV
1
LUB
2
EFR
3
CCPS
4
DVVLS
5
DEACS
6
ICP
7
DCP
8
DVVLD
9
CVVL
10
DEACD
11
DEACO
12
CCPO
Length of Wheelbase
Length of Track Width in inches
Length of Track Width in inches
Average Track x Wheelbase
Curb Weight of the Vehicle
Gross Vehicle Weight Rating of the Vehicle
Is the Vehicle a Hybrid?
Type of Hybrid
Turbo Charged Engine
Super Charged
Either or both Turbo charged / Super Charged
Single Overhead Cam Engine
Dual Overhead Cam Engine
Overhead Valve Engine
Low friction lubricants incremental to base
engine
Engine friction reduction incremental to base
engine
VVT - Coupled Cam Phasing (CCP) on SOHC
Discrete Variable Valve Lift (DVVL) on SOHC
Cylinder Deactivation on SOHC
VVT - Intake Cam Phasing (ICP)
VVT - Dual Cam Phasing (DCP) (independent)
Discrete Variable Valve Lift (DVVL) on DOHC
Continuously Variable Valve Lift (CVVL)
Cylinder Deactivation on DOHC
Cylinder Deactivation on OHV
VVT - Coupled Cam Phasing (CCP) on OHV
Some from Edmonds.com or
Motortrend.com, Others from
product plans with a subset
verified with Edmonds.com or
Motortrend.com for accuracy.
Some from Edmonds.com or
Motortrend.com, Others from
product plans with a subset
verified with Edmonds.com or
Motortrend.com for accuracy.
Some from Edmonds.com or
Motortrend.com, Others from
product plans with a subset
verified with Edmonds.com or
Motortrend.com for accuracy.
Calculated from track width
and wheel base
Some from Edmonds.com or
Motortrend.com, Others from
product plans with a subset
verified with Edmonds.com or
Motortrend.com for accuracy.
Some from Edmonds.com or
Motortrend.com, Others from
product plans with a subset
verified with Edmonds.com or
Motortrend.com for accuracy.
Certification data
Certification data
Certification data
Certification data
Certification data
Wards
Wards
Wards
Obtained from product plans
from manufacturers that report
this information.2
Obtained from product plans
from manufacturers that report
this information.2
Wards1
Wards1
Wards1
Wards1
Wards1
Wards1
Wards1
Wards1
Wards1
Wards1
1-6
-------
The Baseline Vehicle Fleet
13
DVVLO
14
CDOHC
15
SGDI
16
CBRST
17
TRBDS
18
EGRB
19
DSLT
20
DSLC
21
6MAN
22
IATC
23
CVT
24
NAUTO
25
DCTAM
26
EPS
27
IACC
28
MHEV
29
HVIA
30
BISG
31
CISC
32
PSHEV
33
2MHEV
34
PHEV
35
MSI
36
MS2
37
MS5
38
ROLL
39
LDB
40
SAXU
Discrete Variable Valve Lift (DVVL) on OHV
Conversion to DOHC with DCP
Stoichiometric Gasoline Direct Injection (GDI)
Combustion Restart
Turbocharging and Downsizing
Exhaust Gas Recirculation (EGR) Boost
Conversion to Diesel (from TRBDS)
Conversion to Diesel (from CBRST)
6-Speed Manual/Improved Internals
Improved Auto. Trans. Controls/Externals
Continuously Variable Transmission
6/7/8-Speed Auto. Trans with Improved Internals
Dual Clutch or Automated Manual Transmission
Electric Power Steering
Improved Accessories
12V Micro-Hybrid
Higher Voltage/Improved Alternator
Belt Mounted Integrated Starter Generator
Crank Mounted Integrated Starter Generator
Power Split Hybrid
2-Mode Hybrid
Plug-in Hybrid
Material Substitution (1%)
Material Substitution (2%)
Material Substitution (5%)
Low Rolling Resistance Tires
Low Drag Brakes
Secondary Axle Disconnect-Unibody
Wards1
Not available
Wards1
Not available
Not available
Not available
Not available
Not available
Certification data
Not available
Certification data
Certification data
Certification data
Obtained from product plans
from manufacturers that report
this information.2
Obtained from product plans
from manufacturers that report
this information.2
Not available
Not available
Certification data
Certification data
Certification data
Certification data
Not available
Not available
Not available
Not available
Obtained from product plans
from manufacturers that report
this information.2
Not available
Not available
1-7
-------
Draft Joint Technical Support Document
41
SAXL
Import Car
Index
Volume 2011
Volume 2012
Volume 2013
Volume 2014
Volume 2015
Volume 2016
Vehicle Type Number
Footprint: PU Average
Footprint: PU Cutoff
NHTSA Defined New NHTSA
Car/Truck
C02
Thesholded Footprint
TEB Tech 1
TEB Tech 2
TEB Tech 3
TEB Tech 4
Secondary Axle Disconnect-Ladder
Cars Imported
Index Used to link EPA and NHTSA baselines
Projected Production Volume for 201 1
Projected Production Volume for 2012
Projected Production Volume for 2013
Projected Production Volume for 2014
Projected Production Volume for 2015
Projected Production Volume for 2016
Vehicle Type Number assigned to a vehicle based
on its number of cylinders, valves per cylinder, and
valve actuation technology.
Car Foot Print is normal. Truck footprint is the
production weighted average for each vehicle.
Same as above, but footprint valve for trucks
limited to 66 before weighting. Meaning valves
greater than 66 equal 66.
New NHTSA Car Truck value as determined by
NHTSA. Used in calculations.
COi calculated from MPG. CC>2 weighted 1.15
times higher for diesel vehicles.
Footprint valve that will be set to 41 for values less
than 41, Will be set to 56 for car values > 56, and
will be set to 66 for truck values >66
Used in Summary Calculations
Percentage of Tech Package 1
Percentage of Tech Package 2
Percentage of Tech Package 3
Percentage of Tech Package 4
Not available
Certification data
Created
Calculated based on 2008
volume and AEO and CSM
adjustment factors.
Calculated based on 2008
volume and AEO and CSM
adjustment factors.
Calculated based on 2008
volume and AEO and CSM
adjustment factors.
Calculated based on 2008
volume and AEO and CSM
adjustment factors.
Calculated based on 2008
volume and AEO and CSM
adjustment factors.
Calculated based on 2008
volume and AEO and CSM
adjustment factors.
Mapping done by Agencies
Derived from data from
Edmunds.com or
Motortrend.com. Production
volumes or specific footprints
from product plans.
Derived from data from
Edmunds.com or
Motortrend.com. Production
volumes or specific footprints
from product plans.
NHTSA
Certification data
Derived from data from
Edmunds.com or
Motortrend.com. Production
volumes or specific footprints
from product plans.
Calculated using a macro in
another spread sheet.
Calculated based on the
technology in the "Data" sheet.
Calculated using a macro in
another spread sheet.
Calculated based on the
technology in the "Data" sheet.
Calculated using a macro in
another spread sheet.
Calculated based on the
technology in the "Data" sheet.
Calculated using a macro in
another spread sheet.
Calculated based on the
technology in the "Data" sheet.
1-8
-------
The Baseline Vehicle Fleet
TEB Tech 5
TEB Tech 6
CEB Tech 1
CEB Tech 2
CEB Tech 3
CEB Tech 4
CEB Tech 5
CEB Tech 6
Percentage of Tech Package 5
Percentage of Tech Package 6
Percentage of Cost of Tech Package 1
Percentage of Cost of Tech Package 2
Percentage of Cost of Tech Package 3
Percentage of Cost of Tech Package 4
Percentage of Cost of Tech Package 5
Percentage of Cost of Tech Package 6
Calculated using a macro in
another spread sheet.
Calculated based on the
technology in the "Data" sheet.
Calculated using a macro in
another spread sheet.
Calculated based on the
technology in the "Data" sheet.
Calculated using a macro in
another spread sheet.
Calculated based on the
technology in the "Data" sheet.
Calculated using a macro in
another spread sheet.
Calculated based on the
technology in the "Data" sheet.
Calculated using a macro in
another spread sheet.
Calculated based on the
technology in the "Data" sheet.
Calculated using a macro in
another spread sheet.
Calculated based on the
technology in the "Data" sheet.
Calculated using a macro in
another spread sheet.
Calculated based on the
technology in the "Data" sheet.
Calculated using a macro in
another spread sheet.
Calculated based on the
technology in the "Data" sheet.
Notes:
1. For engines not available in the WardsAuto.com engine file, an internet search was done to find this information.
2. These data were obtained from manufacturer's product plans. They were obtain to block (where possible) the model from
adding technology that was already on a vehicle.
3. Ward's Automotive Group data obtained from "2008 Light Vehicle Engines" .
The sales volumes for the MY 2008 baseline fleet are included in the section
below on reference fleet under the MY 2008 columns. Table 1-2 displays the
technologies present in the baseline fleet.
1-9
-------
Draft Joint Technical Support Document
Table 1-2 2008 Technology Percentages
All Trucks
All Cars
BMW Trucks
BMW Cars
Chrysler Trucks
Chrysler Cars
Ford Trucks
Ford Cars
Subaru Trucks
Subaru Cars
GM Trucks
GM Cars
Honda Trucks
Honda Cars
Hyundai Trucks
Hyundai Cars
Tata Trucks
Tata Cars
Kia Trucks
Kia Cars
Mazda Trucks
Mazda Cars
Daimler Trucks
Daimler Cars
Mitsubishi Trucks
Mitsubishi Cars
Nissan Trucks
Nissan Cars
Porsche Trucks
Porsche Cars
Suzuki Trucks
Suzuki Cars
Toyota Trucks
Toyota Cars
VW Trucks
VW Cars
.c
L*
=
0.0%
2.5%
0.0%
0.0%
0.0%
0.9%
0.0%
3.1%
0.0%
0.0%
0.0%
2.6%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
12.4%
17.4%
0.0%
0.0%
0.0%
0.0%
0.6%
43.1%
T3
&
!S
U
4>
&.
=
t/3
0.2%
0.1%
0.0%
0.0%
0.0%
0.0%
0.1%
0.9%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
20.3%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
SOHC
23.7%
17.4%
0.0%
13.9%
38.9%
20.8%
62.9%
14.2%
70.5%
69.2%
0.0%
0.0%
64.3%
57.2%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.6%
0.0%
36.2%
54.6%
100.0%
100.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
85.0%
DOHC
47.6%
73.3%
100.0%
86.0%
4.1%
71.5%
34.3%
85.8%
29.5%
30.8%
30.8%
56.2%
35.7%
42.8%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
99.4%
100.0%
63.8%
45.4%
0.0%
0.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
15.0%
>
O
28.7%
9.2%
0.0%
0.0%
56.9%
7.7%
2.8%
0.0%
0.0%
0.0%
69.2%
43.8%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.1%
M
J
61.4%
48.5%
0.0%
0.0%
76.9%
28.5%
3.9%
6.4%
0.0%
0.0%
92.7%
92.0%
0.0%
0.0%
100.0%
100.0%
0.0%
0.0%
100.0%
100.0%
0.0%
0.0%
0.0%
0.0%
100.0%
100.0%
100.0%
100.0%
0.0%
0.0%
0.0%
0.0%
89.1%
69.9%
0.0%
0.0%
Vj
U
U
4.1%
1.4%
0.0%
11.9%
0.0%
0.0%
26.5%
4.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.9%
0.0%
0.0%
37.7%
42.9%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
DVVLS
5.8%
7.4%
0.0%
11.9%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
64.3%
57.2%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.9%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
t/3
u
<
w
Q
0.0%
1.3%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
10.7%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.9%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
u
5.3%
3.2%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
14.7%
11.5%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.9%
46.7%
30.3%
0.0%
0.0%
0.0%
0.0%
100.0%
28.7%
0.0%
0.0%
0.0%
0.2%
78.9%
1.9%
U
Q
36.6%
47.9%
94.7%
85.8%
3.9%
41.7%
9.3%
5.2%
29.5%
30.8%
30.8%
28.8%
31.5%
42.6%
29.6%
32.2%
100.0%
100.0%
17.2%
10.2%
99.3%
99.9%
0.0%
0.0%
0.0%
0.0%
100.0%
100.0%
100.0%
82.6%
0.0%
0.0%
78.3%
99.4%
99.4%
14.1%
DVVLD
1.5%
0.2%
0.0%
1.5%
0.0%
0.0%
0.0%
0.0%
27.4%
0.9%
0.0%
0.0%
4.2%
0.3%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.9%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
100.0%
0.0%
0.0%
0.0%
0.0%
0.0%
78.9%
0.8%
1-10
-------
The Baseline Vehicle Fleet
Table 1-2 2008 Technology Percentages
All Trucks
All Cars
BMW Trucks
BMW Cars
Chrysler Trucks
Chrysler Cars
Ford Trucks
Ford Cars
Subaru Trucks
Subaru Cars
GM Trucks
GM Cars
Honda Trucks
Honda Cars
Hyundai Trucks
Hyundai Cars
Tata Trucks
Tata Cars
Kia Trucks
Kia Cars
Mazda Trucks
Mazda Cars
Daimler Trucks
Daimler Cars
Mitsubishi Trucks
Mitsubishi Cars
Nissan Trucks
Nissan Cars
Porsche Trucks
Porsche Cars
Suzuki Trucks
Suzuki Cars
Toyota Trucks
Toyota Cars
VW Trucks
VW Cars
J
>
U
2.8%
5.2%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
31.5%
42.6%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.9%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
DEACD
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.9%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
DEACO
10.5%
1.2%
0.0%
0.0%
0.0%
5.2%
0.0%
0.0%
0.0%
0.0%
39.1%
4.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.9%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
o
u
u
1.4%
5.7%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
5.0%
28.8%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.9%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
DVVLO
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.9%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
3
o
t/3
2.9%
4.7%
6.3%
7.7%
0.2%
0.0%
1.0%
0.0%
0.0%
0.0%
0.0%
6.0%
4.2%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
23.5%
11.4%
15.9%
1.7%
0.0%
0.0%
0.0%
0.0%
100.0%
17.4%
0.0%
0.0%
5.7%
7.6%
100.0%
46.0%
Z
<
§
c
1.0%
2.3%
0.7%
14.6%
4.3%
0.9%
0.0%
1.2%
0.0%
4.1%
0.0%
1.1%
0.0%
3.1%
0.0%
0.2%
0.0%
0.0%
0.0%
0.0%
0.0%
3.8%
0.0%
0.3%
0.0%
1.4%
1.2%
4.5%
1.4%
51.9%
0.0%
0.0%
1.4%
0.1%
0.0%
6.4%
H
U
2.6%
11.5%
0.0%
0.5%
3.6%
18.5%
1.8%
1.9%
0.0%
0.0%
0.0%
0.0%
0.0%
3.2%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.1%
0.1%
0.0%
0.0%
14.4%
44.2%
16.5%
69.4%
0.0%
0.0%
0.0%
0.0%
4.7%
16.7%
0.0%
5.4%
NAUTO
19.9%
18.5%
99.3%
83.3%
21.4%
3.9%
18.8%
36.8%
0.0%
0.0%
18.6%
12.3%
0.0%
0.0%
25.0%
0.0%
100.0%
91.6%
0.0%
0.0%
57.2%
19.2%
99.3%
67.4%
23.3%
7.6%
0.0%
0.0%
98.6%
0.0%
0.0%
0.0%
19.2%
22.5%
100.0%
65.7%
DCTAM
0.0%
0.8%
0.0%
1.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
10.4%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
14.3%
t/J
w
0.9%
12.6%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
16.5%
32.3%
0.0%
0.0%
0.0%
0.0%
0.0%
64.0%
0.0%
0.0%
O
t/3
3
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
o
t/3
u
0.0%
0.4%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
3.2%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
1-11
-------
Draft Joint Technical Support Document
Table 1-2 2008 Technology Percentages
All Trucks
All Cars
BMW Trucks
BMW Cars
Chrysler Trucks
Chrysler Cars
Ford Trucks
Ford Cars
Subaru Trucks
Subaru Cars
GM Trucks
GM Cars
Honda Trucks
Honda Cars
Hyundai Trucks
Hyundai Cars
Tata Trucks
Tata Cars
Kia Trucks
Kia Cars
Mazda Trucks
Mazda Cars
Daimler Trucks
Daimler Cars
Mitsubishi Trucks
Mitsubishi Cars
Nissan Trucks
Nissan Cars
Porsche Trucks
Porsche Cars
Suzuki Trucks
Suzuki Cars
Toyota Trucks
Toyota Cars
VW Trucks
VW Cars
PSHEV
0.8%
2.7%
0.0%
0.2%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
1.1%
0.0%
0.0%
0.0%
0.0%
4.7%
16.9%
0.0%
0.0%
2MHEV
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.5%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
w
-
0.0%
0.0%
0.0%
0.2%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
T3
4>
0>
=.
t/3
t
o
Is
<
37.9%
39.1%
0.0%
0.0%
34.9%
67.0%
55.9%
36.3%
62.1%
69.6%
75.5%
77.9%
0.0%
0.0%
29.8%
76.4%
0.0%
0.0%
23.1%
51.5%
10.1%
33.3%
0.0%
0.0%
62.3%
30.8%
0.0%
7.0%
0.0%
0.0%
0.0%
76.2%
9.0%
35.7%
0.0%
0.0%
"§
0>
=.
t/3
Ifi
3
=
<
37.3%
22.9%
0.0%
0.0%
35.5%
8.2%
20.9%
18.2%
29.5%
2.4%
5.0%
5.4%
99.7%
87.3%
45.0%
16.7%
0.0%
8.4%
75.7%
41.0%
29.6%
29.7%
0.7%
22.0%
0.0%
4.5%
81.6%
19.1%
0.0%
33.7%
99.5%
13.1%
63.9%
19.5%
0.0%
0.0%
"§
0>
=.
t/3
^0
.3
=
<
18.6%
16.4%
99.3%
82.6%
21.4%
3.9%
18.8%
36.8%
0.0%
0.0%
19.0%
12.3%
0.0%
0.0%
25.0%
0.0%
100.0%
91.6%
0.0%
0.0%
57.2%
19.2%
0.0%
0.0%
23.3%
7.6%
0.0%
0.0%
98.6%
0.0%
0.0%
0.0%
19.2%
19.8%
100.0%
65.7%
"§
0>
=.
t/3
r-
|
<
1.4%
1.7%
0.0%
0.8%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
99.3%
67.4%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
T3
4>
0>
=.
t/3
90
o
Is
<
0.0%
0.4%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
2.6%
0.0%
0.0%
8
4>
=.
t/J
ft
"a
=
B
!S
§
1.1%
4.9%
0.0%
0.3%
0.3%
1.4%
2.7%
5.6%
8.4%
23.9%
0.4%
3.2%
0.3%
6.4%
0.2%
6.7%
0.0%
0.0%
1.2%
7.5%
3.0%
13.9%
0.0%
0.0%
0.0%
11.6%
0.7%
0.0%
0.0%
14.4%
0.5%
10.7%
1.9%
5.5%
0.0%
8.2%
a.
t/3
c
"a
=
B
!S
§
1.0%
2.3%
0.7%
14.6%
4.3%
0.9%
0.0%
1.2%
0.0%
4.1%
0.0%
1.1%
0.0%
3.1%
0.0%
0.2%
0.0%
0.0%
0.0%
0.0%
0.0%
3.8%
0.0%
0.3%
0.0%
1.4%
1.2%
4.5%
1.4%
51.9%
0.0%
0.0%
1.4%
0.1%
0.0%
6.4%
H
U
2.6%
11.5%
0.0%
0.5%
3.6%
18.5%
1.8%
1.9%
0.0%
0.0%
0.0%
0.0%
0.0%
3.2%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.1%
0.1%
0.0%
0.0%
14.4%
44.2%
16.5%
69.4%
0.0%
0.0%
0.0%
0.0%
4.7%
16.7%
0.0%
5.4%
Automated Manual
5 Speed
0.0%
0.3%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
10.4%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
Automated Manual
6 Speed
0.0%
0.5%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
14.3%
Automated Manual
7 Speed
0.0%
0.0%
0.0%
1.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
These technology tables indicate that some manufacturers are already
implementing some of these technologies in their 2008 fleet. VW stands out as having a
significant number of turbocharged direct injection engines, though it is uncertain
1-12
-------
The Baseline Vehicle Fleet
whether their engines are also downsized. Some of the valve and cam technologies are
quite common today: for example, nearly half the fleet has dual cam phasing, while
Honda and GM are have considerable levels of engines with cylinder deactivation
(DEAC). Honda also has already implemented continuously variable valve lift (CVVL)
on a majority of their engines. Moreover, Toyota has the highest penetration of hybrid
technologies. Finally, regarding the technologies obtained from the product plans: many
of the manufacturers who reported the lubricating oil information, indicated that low
friction (or low viscosity) oil will be used more frequently in the future. Therefore, it is
likely that this technology is underreported for the fleet as a whole since it was assumed
that the usage of low viscosity oils was zero for the companies who did not provide
product plans. The lack of reporting of these types of technologies (if they exist in the
fleet) will probably lead to a slightly higher cost of compliance for these companies than
what the agencies show in the Preamble for the MY 2011 costs. However, as described
in the section II.D of the Preamble, these technologies alone do not necessarily guarantee
a better fuel economy compared to other vehicles in their footprint class. There are many
other considerations at play, most notably, weight and power.
The section below provides further detail on the conversion of the MY 2008
baseline into the MY 2011-2016 reference fleet. It also described more of the data
contained in the baseline spreadsheet.
1.3 The MY 2011-2016 Reference Fleet
The reference fleet aims to reflect the current market conditions and expectations
about conditions of the vehicle fleet during the rulemaking time frame. Fundamentally,
constructing this fleet involved projecting the MY 2008 baseline fleet into the MY 2011-
2016 model years. It also included the assumption that none of the models had changes
during this period. Projecting what the fleet will look like in the future is a process that is
inherently full of uncertainty. NHTSA and EPA therefore relied on many sources of
reputable information to make these projections.
1.3.1 On what data is the baseline vehicle fleet based?
EPA and NHTSA have based the projection of total car and light truck sales on
recent projections made by the Energy Information Administration (EIA). EIA publishes
a long-term projection of national energy use annually called the Annual Energy Outlook
(AEO).4 This projection utilizes a number of technical and econometric models which
are designed to reflect both economic and regulatory conditions expected to exist in the
future. In support of its projection of fuel use by light-duty vehicles, EIA projects sales
of new cars and light trucks. Due to the state of flux of both energy prices and the
economy, EIA published three versions of its 2009 Annual Energy Outlook. The
Preliminary 2009 report was published early (in November 2008) in order to reflect the
1-13
-------
Draft Joint Technical Support Document
dramatic increase in fuel prices which occurred during 2008, after the development of the
2008 AEO. The official 2009 report was published in March of 2009. A third 2009
report was published a month later which reflected the economic stimulus package passed
by Congress earlier this year. For the analysis in this rule, NHTSA and EPA use the sales
projections of this latest report, referred to as the updated 2009 AEO.
In their updated 2009 report, EIA projects that total light-duty vehicle sales
gradually recover from their currently depressed levels by roughly 2013. In 2016, car
and light truck sales are projected to be 9.5 and 7.1 million units, respectively. While the
total level of sales of 16.6 million units is similar to pre-2008 levels, the fraction of car
sales is higher than that existing in the 2000-2007 timeframe. This reflects the impact of
higher fuel prices, as well as the effects of the EISA (which called for a minimum fleet
average of 35 mpg by 2020). Note that EIA's definition of cars and trucks follows that
used by NHTSA prior to the MY 2011 CAFE final rule published earlier this year. The
MY 2011 CAFE final rule reclassified a number of 2-wheel drive sport utility vehicles
from the truck fleet to the car fleet. EIA's sales projections of cars and trucks for the
2011-2016 model years under both the new and old NHTSA truck definition are shown in
Table 1-3 and Table 1-4 below, respectively. Actual vehicle sales for MY 2008 are also
shown under the new NHTSA truck definition. Slight differences exist between the total
car and truck sales projections. These differences are less than 0.1% and are due to the
iterative process described further below to shift sales between car and truck market
segments.
Table 1-3 AEO Volumes New NHTSA Car Truck Definition After Projections
All Trucks
All Cars
Total
2008
Baseline
Sales
5,620,847
8,220,517
13,841,364
2011
Projected
Volume
5,547,326
8,235,204
13,782,530
2012
Projected
Volume
5,530,583
9,255,624
14,786,207
2013
Projected
Volume
5,720,280
9,977,341
15,697,621
2014
Projected
Volume
5,810,895
10,479,350
16,290,245
2015
Projected
Volume
5,717,300
10,890,967
16,608,267
2016
Projected
Volume
5,586,953
11,067,608
16,654,561
Table 1-4 AEO Volumes Old NHTSA Car Truck Definition Before Projections
All Trucks
All Cars
Total
2011
Projected
Volume
7,007,298
6,954,580
13,961,878
2012
Projected
Volume
7,725,263
7,125,693
14,850,955
2013
Projected
Volume
8,261,773
7,391,939
15,653,713
2014
Projected
Volume
8,776,298
7,440,095
16,216,393
2015
Projected
Volume
9,241,209
7,334,371
16,575,580
2016
Projected
Volume
9,468,365
7,112,689
16,581,055
In addition to a shift towards more car sales, sales of segments within both the car
and truck markets have also been changing and are expected to continue to change in the
future. Manufacturers are introducing more crossover models which offer much of the
utility of SUVs but using more car-like designs. In order to reflect these changes in fleet
1-14
-------
The Baseline Vehicle Fleet
makeup, EPA and NHTSA considered several available forecasts. After reviewing
information from a number of forecasting organizations, the agencies decided to purchase
forecasts from two well-known industry analysts, CSM-Worldwide (CSM), and J.D.
Powers.5 NHTSA and EPA decided to use the forecast from CSM, for several reasons.
One, CSM agreed to allow us to publish their high level data, on which the forecast is
based, in the public domain. Two, it covered nearly all the timeframe of greatest
relevance to this proposed rule (2012-2015 model years). Three, it provided projections
of vehicle sales both by manufacturer and by market segment. Four, it utilized market
segments similar to those used in the EPA emission certification program and fuel
economy guide. As discussed further below, this allowed the CSM forecast to be
combined with other data obtained by NHTSA and EPA. The breakdown of car and
truck sales by manufacturer and by market segment for the 2016 model year and beyond
were assumed to be the same as CSM's forecast for the 2015 calendar year.
1.3.2 How do the agencies develop the reference vehicle fleet?
The process of producing the 2011-2016 reference fleet involved combining the
baseline fleet with the projection data described above. This was a complex multistep
procedure, which is described in this section.
1.3.2.1 How was the 2008 baseline data merged with the CSM data?
Merging the 2008 baseline data with the 2011-2016 CSM data required a
thorough mapping of certification vehicles to CSM vehicles by individual make and
model. One challenge the agencies faced when determining a reference case fleet was
that the sales data projected by CSM had different market segmentation than the data
contained in EPA's internal database. In order to create a common segmentation between
the two databases, side-by-side comparison of the specific vehicle models in both
datasets was performed, and additional "CSM segment" modifier in the spreadsheet was
created, thus mapping the two datasets. The reference fleet sales based on the "CSM
segmentation" was then projected
The baseline data and reference fleet volumes are available to the public. The
baseline Excel spreadsheet in the docket is the result of the merged files.6 It provides
specific details on the sources and definitions for the data. The Excel file contains
several tabs. They are: "Data", "Data Tech Definitions", "SUM", "SUM Tech
Definitions", "Truck Vehicle Type Map", and "Car Vehicle Type Map". "Data" is the
tab with the raw data. "Data Tech Definitions" is the tab where each column is defined
and its data source named. "SUM" is the tab where the raw data is processed to be used
in the OMEGA and Volpe models. The "SUM" tab minus columns A-F and minus the
Generic vehicles is the input file for the models. The "Generic" manufacturer (shown in
the "SUM" tab) is the sum of all manufacturers and is calculated as a reference, and for
data verification purposes. It is used to validate the manufacturers' totals. It also gives
1-15
-------
Draft Joint Technical Support Document
an overview of the fleet. Table 1-5 shows the sum of the models chosen. The number of
models is determined by the number of unique segment and vehicle type combinations.
These combinations of segment and vehicle type (the vehicle type number is the same as
the technology package number) are determined by the technology packages discussed
in.the EPA DRIA (chapter 1). "SUM Tech Definitions" is the tab where the columns of
the "SUM" tab are defined. The "Truck Vehicle Type Map" and "Car Vehicle Type
Map" map the number of cylinder and valve actuation technology to the "tech package"
vehicle type number.
Table 1-5 Models from the SUM Tab Model
Model
Car Like LargeSuv >=V8 Vehicle Type: 13
Car Like LargeSuv V6 Vehicle Type: 16
Car Like LargeSuv V6 Vehicle Type: 12
Car Like LargeSuv V6 Vehicle Type: 9
Car Like LargeSuv 14 and 15 Vehicle Type: 7
Car Like MidSizeSuv V6 Vehicle Type: 8
Car Like MidSizeSuv V6 Vehicle Type: 5
Car Like MidSizeSuv 14 Vehicle Type: 7
Car Like SmallSuv V6 Vehicle Type: 12
Car Like SmallSuv V6 Vehicle Type: 4
Car Like SmallSuv 14 Vehicle Type: 3
LargeAuto >=V8 Vehicle Type: 13
LargeAuto >=V8 Vehicle Type: 10
LargeAuto >=V8 Vehicle Type: 6
LargeAuto V6 Vehicle Type: 12
LargeAuto V6 Vehicle Type: 5
MidSizeAuto >=V8 Vehicle Type: 13
MidSizeAuto >=V8 Vehicle Type: 10
MidSizeAuto >=V8 (7 or >) Vehicle Type: 6
MidSizeAuto V6 Vehicle Type: 12
MidSizeAuto V6 Vehicle Type: 8
MidSizeAuto V6 Vehicle Type: 5
MidSizeAuto 14 Vehicle Type: 3
In the combined EPA certification and CSM database, all of the 2008 vehicle
models were assumed to continue out to 2016, though their volumes changed in
proportion to CSM projections. Also, any new models expected to be introduced within
the 2009-2016 timeframe are not included in the data. These volumes are reassigned to
the existing models. All MY 2011-2016 vehicles are mapped to the existing vehicles by
a process of mapping to manufacturer market share and overall segment distribution. The
mappings are discussed in the next section. Further discussion of this limitation is
discussed below in section 1.3.4. The statistics of this fleet will be presented below since
further modifications were required to the volumes as the next section describes.
1-16
-------
The Baseline Vehicle Fleet
1.3.2.2 How were the CSM forecasts normalized to the AEO forecasts?
The next step in the production of the reference fleet is one of the more
complicated steps to explain. Here, the projected CSM forecasts for relative sales of cars
and tracks by manufacturer and by market segment was normalized (set equal) to the
total sales estimates of the updated 2009 Annual Energy Outlook (AEO). NHTSA and
EPA used projected car and truck volumes for this period from AEO 2009. However, the
AEO projects sales only at the car and truck level, not at the manufacturer and model-
specific level, which are needed for the analysis. The CSM data provided year-by-year
percentages of cars and trucks sold by each manufacturer as well as the percentages of
each vehicle segment. Using these percentages normalized to the AEO-projected
volumes then provided the manufacturer-specific market share and model-specific sales
for model years 2011-2016 (it is worth clarifying that the agencies are not using the
model-specific sales volumes from CSM, only the volumes by manufacturer and
segment). This process is described in greater detail in the following paragraphs.
In order to determine future production volumes, multipliers were developed by
manufacturer and vehicle segment that could be applied to MY 2008 volumes. The
process for developing the multipliers is complicated, but is easiest to explain as a three-
step process, though the first step is combined with both the second and third step, so
only one multiplier per manufacturer and vehicle segment is developed.
The three steps are:
1. Adjust total car and truck sales to match AEO projections.
2. Adjust Car sales to match CSM market share projections for each manufacturer
and car segment.
3. Adjust Truck sales to match CSM market share projections for each manufacturer
and truck segment.
The first step is the adjustment of total car and truck sales in 2008 to match AEO
projections of total car and track sales in 2011-2016. The volumes for all of the trucks in
2008 were added up (TruckSum2008), and so were the volumes of all the cars
(CarSum2008). A multiplier was developed to scale the volumes in 2008 to the AEO
projections. The example equation below shows the general form of how to calculate a
car or truck multiplier. The AEO projections are shown above in Table 1-3.
Example Equation :
TruckMultiplier(Year X) = AEOProjectionforTracks(Year X) / TrackSum2008
CarMultiplier(Year X) = AEOProjectionforCars(Year X) / CarSum2008
Where: Year X is the model year of the multiplier.
The AEO projection is different for each model year. Therefore, the multipliers
are different for each model year. The multipliers can be applied to each 2008 vehicle as
a first adjustment, but multipliers based solely on AEO have limited value since it can
1-17
-------
Draft Joint Technical Support Document
only give an adjustment that will give the correct total numbers of cars and trucks without
the correct market share or vehicle mix. A correction factor based on the CSM data,
which does contain market share and vehicle segment mix, is therefore necessary, so
combining the AEO multiplier with CSM multipliers (one per manufacturer, segment,
and model year) will give the best multipliers.
There were several steps in developing an adjustment for Cars based on the CSM
data. CSM provided data on the market share and vehicle segment distribution. The first
step in determining the adjustment for Cars was to total the number of Cars in each
vehicle segment by manufacturer in MY 2008. A total for all manufacturers in each
segment was also calculated. The next step was to multiply each segment the volume of
each segment for each manufacturer by the CSM market share. The AEO multiplier was
also applied at this time. This gave projected volumes with AEO total volumes and
market share correction for Cars. This is shown in the "Adjusted for 2011 AEO and
Manufacturer Market Share" column of Table 1-6.
The next step is to adjust the sales volumes for CSM vehicle segment distribution.
The process for adjusting for vehicle segment is more complicated than a simple one step
multiplication. In order to keep manufacturers' volumes constant and still have the
correct vehicle segment distribution, vehicles need to move from segment to segment
while maintaining constant manufacturers' totals. Six rules and one assumption were
applied to accomplish the shift. The assumption (based on the shift in vehicle sales in the
last year) is that people are moving to smaller vehicles in the rulemaking time frame. A
higher level (less detailed) example of this procedure is provided in the preamble section
II.
1. Vehicles from CSM's "Luxury Car," "Specialty Car," and "Other Car" segments,
if reduced will be equally distributed to the remaining four categories ("Full-Size
Car," "Mid-Size Car," "Small Car," "Mini Car"). If these sales increased, they
were taken from the remaining four categories so that the relative sales in these
four categories remained constant.
2. Vehicles from CSM's "Luxury Car," "Specialty Car," and "Other Car" segments,
if increased will take equally from the remaining categories ("Full-Size Car,"
"Mid-Size Car," "Small Car," "Mini Car").
3. All manufacturers have the same multiplier for a given segment shift based on
moving all vehicles in that segment to achieve the CSM distribution. Table 1-6
shows how the 2011 vehicles moved and the multipliers that were created for
each adjustment. This does not mean that new vehicle segments will be added
(except for Generic Mini Car described in the next step) to manufacturers that do
not produce them. Vehicles within each manufacturer will be shifted as close to
the distribution as possible given the other rules. Table 1-7 has the percentages of
Cars per CSM segment. These percentages are multiplied by the total number of
vehicles in a given year to get the total sales in the segment. Table 1-6 shows the
totals for 2011 in the "2011 AEO-CSM Sales Goal" column.
4. When "Full-Size Car," "Mid-Size Car," "Small Car" are processed, if vehicles
need to move in or out of the segment, they will move into or out of the next
1-18
-------
The Baseline Vehicle Fleet
6.
smaller segment. So, if Mid-Size Cars are being processed they can only move to
or be taken from Small Cars. Note: In order to accomplish this, a "Generic Mini
Car" segment was added to manufacturers who did not have a Mini (type) Car in
production in 2008, but needed to shift down vehicles from the Small Car
segment.
The data must be processed in the following order: "Luxury Car," "Specialty
Car," "Other Car," "Full-Size Car," "Mid-Size Car," "Small Car." The "Mini
Car" does not need to be processed separately. By using this order, it works out
that vehicles will always move toward the correct distribution. There are two
exceptions, BMW and Porsche only have "Luxury Car," "Specialty Car," and
"Other Car" vehicles, so their volumes were not changed or shifted since these
rules did not apply to them.
When an individual manufacturer multiplier is applied for a segment, the vehicles
move to or from the appropriate segments as specified in the previous rules and as
shown in Table 1-6.
Table 1-6 Example: 2011 Model Year Volume Shift
CSM Segment
All Full-Size Car
All Luxury Car
All Mid-Size Car
All Mini Car
All Small Car
All Specialty Car
All Others
2008 MY
Sales
730,355
1,057,875
1,970,494
599,643
1,850,522
754,547
3,259
Adjusted for
2011 AEO
and
Manufacturer
Market Share
501,245
1,076,470
1,946,981
686,738
2,007,527
783,982
4,355
Luxury, Specialty,
Other Adjustment
520,885
970,680
2,033,087
733,339
2,099,343
637,785
12,178
Full Size
Adjustment
411,025
970,680
2,142,948
733,339
2,099,343
637,785
12,178
Midsize
Adjustment
411,025
970,680
1,838,095
733,339
2,404,196
637,785
12,178
Small Car
Adjustment
411,025
970,680
1,838,095
991,309
2,146,226
637,785
12,178
2011 AEO-
CSM Sales
Goal
411,025
970,666
1,838,095
991,349
2,146,226
637,759
12,178
Number Vehicles that shift and Where
All Full-Size Car
All Luxury Car
All Mid-Size Car
All Mini Car
All Small Car
All Specialty Car
All Others
19,640
(105,790)
86,107
46,601
91,816
(146,198)
7,823
(109,861)
-
109,861
-
-
-
-
-
-
(304,853)
-
304,853
-
-
-
-
-
257,970
(257,970)
-
-
Individual Manufacturer Multiplier
All Full-Size Car
All Luxury Car
All Mid-Size Car
All Mini Car
All Small Car
All Specialty Car
All Others
0.902
0.814
2.796
0.789
0.858
1.352
1-19
-------
Draft Joint Technical Support Document
Table 1-7 CSM - Percent of Cars per Segment
CSM Segment
Compact Car
Full-Size Car
Luxury Car
Mid-Size Car
Mini Car
Small Car
Specialty Car
Others
2011
0.0%
5.9%
13.9%
26.2%
14.1%
30.6%
9.1%
0.2%
2012
0.0%
4.3%
13.7%
29.0%
15.5%
27.5%
9.8%
0.2%
2013
0.0%
4.7%
14.6%
27.9%
15.5%
26.5%
10.7%
0.2%
2014
0.0%
4.9%
15.0%
27.5%
15.4%
26.1%
10.9%
0.2%
2015
0.0%
4.9%
15.1%
27.9%
15.4%
25.8%
10.6%
0.2%
2016
0.0%
4.9%
15.1%
27.9%
15.4%
25.8%
10.6%
0.2%
Mathematically, an individual manufacturer multiplier is calculated by making the
segment the goal and dividing by the previous total for the segment (shown in Table 1-6).
If the number is greater than 1, the vehicles are entering the segment, and if the number is
less than 1, the vehicles are leaving the segment. So, for example, if Luxury Cars have
an adjustment of 1.5, then for a specific manufacturer who has Luxury Cars, a multiplier
of 1.5 is applied to its luxury car volume, and the total number of vehicles that shifted
into the Luxury segment is subtracted from the remaining segments to maintain that
company's market share. On the other hand, if Large Cars have an adjustment of 0.7,
then for a specific manufacturer who has Large Cars, a multiplier of 0.7 is applied to its
Large Cars, and the total number of vehicles leaving that segment is transferred into that
manufacturer's Mid-Size Cars.
After the vehicle volumes are shifted using the above rules, a total for each
manufacturer and vehicle segment is maintained. The total for each manufacturer
segment for a specific model year (e.g., 2011 General Motors Luxury Cars) divided by
the MY 2008 total for that manufacturer segment (e.g., 2008 General Motors Luxury
Cars) is the new multiplier used to determine the future vehicle volume for each vehicle
model. This is done by taking the multiplier (which is for a specific manufacturer and
segment) times the MY 2008 volume for the specific vehicle model (e.g., 2008 General
Motors Luxury Car Cadillac CTS). This process is repeated for each model year (2011-
2016).
The method used to adjust CSM Trucks to the AEO market share was different
than the method used for Cars. The process for Cars is different than Trucks because it is
not possible to predict how vehicles would shift between segments based on current
market trends. This is because of the added utility of some trucks that makes their sales
more insensitive to factors like fuel price. Again, CSM provided data on the market
share and vehicle segment distribution. The process for having the fleet match CSM's
market share and vehicle segment distribution was iterative.
The following totals were determined:
The total number of trucks for each manufacturer in 2008 model year.
The total number of trucks in each truck segment in 2008 model year.
1-20
-------
The Baseline Vehicle Fleet
The total number of track in each segment for each manufacturer in 2008 model
year.
The total number of tracks for each manufacturer in a specific future model year
based on the AEO and CSM data. This is the goal for market share.
The total number of tracks in each truck segment in a specific future model year
based on the AEO and CSM data. This is the goal for vehicle segment
distribution. Table 1-8 has the percentages of Trucks per CSM segment
Table 1-8 CSM - Percent of Trucks per Segment
CSM Segment
Full-Size CUV
Full-Size Pickup
Full-Size SUV
Full-Size Van
Mid-Size CUV
Mid-Size MAV
Mid-Size Pickup
Mid-Size SUV
Mid-Size Van
Small CUV
Small MAV
Small SUV
2011
4.1%
20.0%
3.3%
3.5%
14.5%
3.8%
3.6%
3.2%
1 1 .9%
27.3%
1 .9%
2.8%
2012
3.3%
17.9%
2.8%
3.3%
15.8%
3.3%
5.8%
2.3%
1 1 .3%
28.4%
3.4%
2.4%
2013
4.0%
18.1%
1 .6%
4.0%
17.5%
2.6%
6.4%
1 .8%
10.7%
27.0%
4.1%
2.2%
2014
4.4%
20.1%
1 .2%
3.9%
18.1%
2.2%
6.4%
1 .6%
10.0%
26.1%
4.1%
2.0%
2015
4.5%
20.8%
1 .3%
4.0%
18.4%
2.2%
6.6%
1 .6%
8.7%
26.2%
4.1%
1 .8%
2016
4.5%
20.8%
1 .3%
4.0%
18.4%
2.2%
6.6%
1 .6%
8.7%
26.2%
4.1%
1 .8%
To start, two different types of tables were created. One table had each
manufacturer with its total sales for 2008 (similar to Table 1-9). This table will have the
goal for each manufacturer, and a column added for each iteration with the current total.
The second table has a truck segment total by manufacturer. The second table starts out
with a "Generic" manufacturer (Table 1-10) which is the table where the goal resides.
Each manufacturer (BMW is shown in Table 1-11) is then listed below the "Generic"
manufacturer. With each iteration, a new total is added for each segment that is
calculated and added to the table. This is not shown in the tables below. A process of
first adjusting the numbers in the tables to the goal for market share distribution was
done. This was followed by adjusting to the goal for vehicle segment distribution. Each
time an adjustment was done a new column was added. An adjustment was done by
creating a multiplier (either segment distribution based or manufacturer distribution
based) and applying it to each vehicle segment total in the current iteration. A
manufacturer-based multiplier is calculated by taking the goal total for a manufacturer
and dividing by the current total (starting with 2008 model year volumes) for a
manufacturer. A segment distribution-based multiplier is calculated by taking the goal
distribution volumes in the Generic manufacturer set and dividing them by the current
volume. Table 1-9, Table 1-10, and Table 1-11 below illustrate two iterations using
BMW.
1-21
-------
Draft Joint Technical Support Document
Table 1-9 Manufacturer Truck Totals
BMW
2008 Model Year Sales
61,324
Manufacturer Distribution 2011 Volume Goal
139948
Multiplier for Iteration 1
139948/61324=2.28
Table 1-10 Segment Specific Truck Totals for All Manufacturers
Manufacturer
Generic1
Generic
Generic
Generic
Generic
Generic
Generic
Generic
Generic
Generic
Generic
Generic
CSM Segment
Full-Size Pickup
Mid-Size Pickup
Full-Size Van
Mid-Size Van
Mid-Size MAV
Small MAV
Full-Size SUV
Mid-Size SUV
Small SUV
Full-Size CUV
Mid-Size CUV
Small CUV
2008 Model Year
Sales
1,195,073
598,197
33,384
719,529
191,448
235,524
530,748
347,026
377,262
406,554
798,335
1,441,589
Segment Distribution
2011 Volume Goal
1,390,343
251,433
244,393
826,733
263,765
133,309
232,375
221,607
194,496
287,313
1,007,583
1,901,229
Multipliers
1.30
0.38
8.30
1.37
1.26
0.57
0.46
0.70
0.75
0.69
0.99
1.16
Note: 1. Generic means all manufacturers.
Table 1-11 Segment Specific Truck Totals for BMW
Manufacturer
BMW
BMW
BMW
BMW
BMW
BMW
BMW
BMW
BMW
BMW
BMW
BMW
CSM Segment
Full-Size Pickup
Mid-Size Pickup
Full-Size Van
Mid-Size Van
Mid-Size MAV
Small MAV
Full-Size SUV
Mid-Size SUV
Small SUV
Full-Size CUV
Mid-Size CUV
Small CUV
Total BMW Vehicles
2008 Model Year
Sales
3,882
36,409
21,033
61,324
Iteration 1 Adjust for
Market Share
2.28*3,882=8,551
2.28*36,409=83,013
2.28*21,033=47,955
139,819
Iteration 2 Adjust for
Segment Distribution
1.26*8,851=11,137
0.99*83,013=82,603
1.16*47,955=55,592
149,332
Using this process, the numbers will get closer to the goal of matching CSM's market
share for each manufacturer and distribution for each vehicle segment after each of the
iterations. The iterative process is carried out until the totals nearly match the goals.
1-22
-------
The Baseline Vehicle Fleet
After 19 iterations, all numbers were within 0.01% of CSM's distributions. The
calculation iterations could have been stopped sooner, but they were continued to observe
how the numbers would converge.
After the market share and segment distribution were complete, the totals need to be
used to create multipliers that could be applied to the original individual 2008 model year
vehicle volumes (each unique manufacture models volume). The total for each
manufacturer segment divided by the 2008 model year total for each manufacturer
segment gives a multiplier that can be applied to each vehicle based on its manufacturer
and segment.
The above process is done for each model year needed (2011-2016). The multipliers
are then applied to each vehicle in 2008 model year, which gives a volume for each
vehicle in 2011 through 2016 model year. The "reference case" (which is the technology
being applied to 2012-2016 fleet to meet the 2011 cafe standard) is discussed below in
section 4.1, with the results being presented in section 4.2.
1.3.3 What are the sales volumes and characteristics of the reference fleet?
Table 1-12 and Table 1-13 below contain the sales volumes and percentages that
result from the process above for MY 2008 and 2011-2016. The forecast from CSM used
in this proposal is CSM's forecast from the 2nd quarter of 2009. CSM also provided to
EPA an updated forecast in the 3rd quarter of 2009, which we were unable to use for this
proposal due to time constraints. However, we have placed a copy of the 3rd Quarter
CSM forecast in the public docket for this rulemaking, and we will consider its use, and
any further updates from CSM or other data received during the comment period when
developing the analysis for the final rule.7
Table 1-12 Vehicle Segment Volumes
Reference Class
Segment3
Large Auto
Mid- Size Auto
Compact Auto
Sub-Compact Auto
All Cars
Large Pickup
Small Pickup
Large SUV
Mid-Size SUV
Small SUV
Mini Van
Cargo Van
All Trucks
Actual and Projected Sales Volume
2008
557,693
3,097,859
1,976,424
1,364,434
6,996,410
1,581,880
177,497
2,783,949
1,263,360
285,355
642,055
110,858
6,844,954
2011
390,306
2,649,352
1,966,447
1,887,350
6,893,455
1,535,300
57,694
2,670,474
1,469,957
84,525
749,650
321,476
6,889,076
2012
361,186
3,153,248
2,128,221
2,114,147
7,756,803
1,533,514
112,916
2,642,285
1,563,424
138,914
717,454
320,898
7,029,404
2013
382,212
3,470,051
2,318,084
2,240,751
8,411,098
1,620,983
140,210
2,713,398
1,544,252
180,559
699,786
387,335
7,286,523
2014
390,640
3,722,450
2,480,435
2,348,611
8,942,136
1,776,746
138,240
2,718,252
1,502,608
178,377
657,057
376,831
7,348,109
2015
410,847
3,929,107
2,582,171
2,452,342
9,374,467
1,813,145
136,667
2,705,080
1,490,577
160,839
561,199
366,294
7,233,800
2016
416,829
4,022,893
2,643,374
2,512,431
9,595,527
1,785,188
145,324
2,625,901
1,446,932
156,230
543,558
355,901
7,059,034
a This is no longer the CSM segment, but the segment used for the subsequent analysis in this rule.
1-23
-------
Draft Joint Technical Support Document
Table 1-13 Manufacturer Volumes
Total Vehicles
All Trucks
All Cars
BMW Trucks
BMW Cars
Chrysler Trucks
Chrysler Cars
Ford Trucks
Ford Cars
Subaru Trucks
Subaru Cars
GM Trucks
GMCars
Honda Trucks
Honda Cars
Hyundai Trucks
Hyundai Cars
Tata Trucks
Tata Cars
Kia Trucks
Kia Cars
Mazda Trucks
Mazda Cars
Daimler Trucks
Daimler Cars
Mitsubishi Trucks
Mitsubishi Cars
Nissan Trucks
Nissan Cars
Porsche Trucks
Porsche Cars
Suzuki Trucks
Suzuki Cars
Toyota Trucks
Toyota Cars
VW Trucks
VWCars
0)
§1 JS
* 1 rc
M
13,841,364
6,844,954
6,996,410
61,324
291,796
956,792
700,413
846,596
1,021,792
82,546
116,035
1,512,047
1,607,460
505,140
1,006,639
53,158
337,869
55,584
9,596
59,472
221,980
55,885
246,661
79,135
208,052
15,371
85,358
305,546
717,869
18,797
18,909
35,319
79,339
951,136
1,260,364
26,999
290,385
"8 2
gS 1
8P
13,782,531
6,889,076
6,893,455
139,959
272,572
402,659
193,704
943,583
1,229,762
117,896
153,644
1,313,927
1,156,366
571,178
996,396
127,199
570,431
31,344
18,572
97,706
302,347
59,782
317,948
97,926
177,493
9,268
67,779
420,666
793,873
21,401
30,909
25,194
91,355
1,059,161
1,473,999
108,477
388,054
"8 2
33 1
sp
14,786,207
7,029,404
7,756,803
142,861
283,471
282,472
178,635
1,025,575
1,348,260
88,877
149,192
1,209,642
1,484,580
669,948
1,155,008
141,028
580,538
42,168
22,878
78,789
304,524
71,930
313,489
109,351
172,172
10,098
64,843
477,897
958,696
20,767
31,605
30,391
86,537
1,019,375
1,697,762
109,415
423,433
1 2
23 |
8P
15,697,621
7,286,523
8,411,098
141,949
323,191
244,441
175,072
1,107,535
1,424,345
77,593
135,946
1,320,438
1,620,301
724,500
1,352,607
135,751
558,975
46,256
32,822
74,596
309,667
71,136
318,669
115,684
190,133
11,440
61,169
478,571
1,031,569
20,639
35,813
29,750
86,220
1,019,048
1,862,201
100,952
458,641
15 «
3-38
S « 3
* 08 0
£>
16,290,246
7,348,109
8,942,136
144,022
352,248
222,545
169,046
1,202,141
1,446,097
76,913
135,141
1,342,856
1,708,507
687,208
1,493,242
134,274
562,862
53,335
35,534
72,192
331,198
66,919
338,487
125,189
204,335
9,124
56,478
474,558
1,073,307
19,379
38,470
30,545
81,480
1,045,671
1,985,033
104,023
467,885
15 «
!£2 o S
S « s
* 08 0
£>
16,608,267
7,233,800
9,374,467
139,034
371,668
112,381
136,583
1,230,889
1,503,175
73,425
130,588
1,355,820
1,789,813
691,641
1,562,496
132,978
590,579
48,583
40,586
69,894
347,533
59,361
340,069
113,366
229,562
7,349
52,368
454,488
1,104,272
16,822
36,175
27,204
75,965
1,081,323
2,114,273
102,743
465,263
15 «
*> | s
5 « s
* 08 0
£>
16,654,561
7,059,034
9,595,527
135,569
380,804
109,674
138,602
1,202,442
1,511,354
74,135
131,022
1,322,512
1,820,234
675,173
1,593,092
129,763
596,891
47,372
41,584
68,310
351,081
57,998
345,489
110,541
235,205
7,171
53,459
444,471
1,123,486
17,273
37,064
26,526
77,427
1,057,837
2,154,115
100,186
476,699
Table 1-14 also shows how the change in fleet make-up may affect the footprint
distributions over time. The resulting data indicates that the average footprint level of the
1-24
-------
The Baseline Vehicle Fleet
fleet will shrink by 1.2% (1% for cars and 2% for trucks, and sales weighted). Thus, the
combination of AEO and CSM projections produces a future vehicle fleet which is only
slightly smaller in terms of footprint than that which was sold in 2008. At the same time,
2008 sales volumes were likely significantly lower than sales earlier in the decade due to
the spike in fuel prices. These projections indicate that the footprint size of the future
fleet will not rebound in the future.
Table 1-14 Production Foot Print Mean
Model Year
2008
2011
2012
2013
2014
2015
2016
Foot Print Mean
for
Car & Truck
Combined
49
48.9
48.5
48.5
48.6
48.5
48.4
Foot Print Mean
for
Car
45.4
45
44.9
45
45
45
45
Foot Print Mean
for
Truck
54.1
54.8
54.5
54.7
55.1
55.2
55.2
Table 1-15 and Table 1-16 below show the changes in engine cylinders over the
model years. The current assumptions show that engines will be downsized over the
timing of this rule. The biggest shift occurs between MY 2008 and 2013. This shift is
caused by the expected changes in class and segment mix as predicted by AEO and CSM,
and does not represent engine downsizing attributable to the proposed rules.
Table 1-15 Truck Percentages of 4, 6, 8 Cylinder Engines by Model Year
Model
All Trucks 8 Cylinders
All Trucks 6 Cylinders
All Trucks 4 Cylinders
Percent of Total Sales of 4,6,8 Cylinders Trucks
2008
33.7%
55.8%
10.5%
2011
32.3%
56.3%
1 1 .4%
2012
30.4%
56.8%
12.8%
2013
29.9%
57.6%
12.4%
2014
31.7%
56.5%
1 1 .9%
2015
32.6%
55.4%
12.0%
2016
32.6%
55.3%
12.1%
Table 1-16 Car Percentages of 4,6, 8 Cylinder Engines by Model Year
Model
All Cars 8 Cylinders
All Cars 6 Cylinders
All Cars 4 Cylinders
Percent of Total Sales of 4,6,8 Cylinders Cars
2008
5.2%
36.7%
58.1%
2011
4.1%
32.8%
63.1%
2012
3.9%
32.4%
63.7%
2013
4.2%
32.6%
63.1%
2014
4.3%
32.6%
63.1%
2015
4.4%
32.6%
62.9%
2016
4.5%
32.5%
63.1%
1-25
-------
Draft Joint Technical Support Document
1.3.4 How is the development of the baseline fleet for this proposal different
from NHTSA's historical approach and why is this approach
preferable?
As described in detail at the opening of this chapter, NHTSA has historically
based its analysis of potential new CAFE standards on detailed product plans the agency
has requested from manufacturers planning to produce light vehicles for sale in the
United States. Although the agency has never compelled manufacturers to submit such
information, most major manufacturers and some smaller manufacturers have voluntarily
provided it when requested.
As in this and other prior rulemakings, NHTSA has requested extensive and
detailed information regarding the models that manufacturers plan to offer, as well as
manufacturers' estimates of the volume of each model they expect to produce for sale in
the U.S. NHTSA's recent requests have sought information regarding a range of
engineering and planning characteristics for each vehicle model (e.g., fuel economy,
engine, transmission, physical dimensions, weights and capacities, redesign schedules),
each engine (e.g., fuel type, fuel delivery, aspiration, valvetrain configuration, valve
timing, valve lift, power and torque ratings), and each transmission (e.g., type, number of
gears, logic).
The information that manufacturers have provided in response to these requests
has varied in completeness and detail. Some manufacturers have submitted nearly all of
the information NHTSA has requested, for most or all of the model years covered by
NHTSA's requests, and have closely followed NHTSA's guidance regarding the
structure of the information. Other manufacturers have submitted partial information,
information for only a few model years, and/or information in a structure less amenable
to analysis. Still other manufacturers have not responded to NHTSA's requests or have
responded on occasion, usually with partial information.
In recent rulemakings, NHTSA has integrated this information and estimated
missing information based on a range of public and commercial sources (such as those
used to develop today's market forecast). For manufacturers which do not respond,
NHTSA has estimated fleet composition based on the latest-available CAFE compliance
data (the same data used as part of the foundation for today's market forecast). NHTSA
has then adjusted the size of the fleet based on AEO's forecast of the light vehicle market
and normalized manufacturers' market shares based on the latest-available CAFE
compliance data.
Compared to this approach, the market forecast the agencies have developed for
this analysis has both advantages and disadvantages.
Most importantly, today's market forecast is much more transparent. The
information sources used to develop today's market forecast are all either in the public
domain, available to the public upon request, or available commercially. As a result,
1-26
-------
The Baseline Vehicle Fleet
NHTSA and EPA are able to docket the market inputs actually used in the agencies'
respective modeling systems, such that any reviewer may independently repeat and
review the agencies' analyses. Previously, although NHTSA provided this type of
information to manufacturers upon request (e.g., GM requested and received outputs
specific to GM), NHTSA was otherwise unable to release market inputs and the most
detailed model outputs (i.e., the outputs containing information regarding specific vehicle
models) because doing so would violate requirements protecting manufacturers'
confidential business information from disclosure.8 Therefore, this approach provides
much greater opportunity for the public to replicate and review the agencies' analyses
and comment accordingly.
Another advantage of this market forecast is the agencies' ability to more fully
assess the incremental costs and benefits of the proposed standards. In the past two
years, NHTSA has requested and received three sets of future product plan submissions
from the automotive companies, most recently this past spring. These submissions are
intended to be the actual future product plans for the companies. In the most recent
submission it is clear that many of the firms have been and are clearly planning for future
CAFE standard increases for model years 2012 and later. This is not surprising, as
NHTSA's proposed standards for MY 2011-2015 and the 2007 passage of EISA have
prompted the companies to expect that there will be increases in the CAFE standards for
MY 2012 and later, as well as the likelihood of future GHG standards. The results for the
product plans for many firms show a significant increase in their projected application of
fuel-saving (and CCVreducing) technologies by MY 2016. However, for the purposes of
assessing the costs of the model year 2012-2016 standards, the use of the product plans
present a difficulty - specifically, how to assess the increased costs of the proposed future
standards if the companies have already anticipated the future standards, such that the
costs are now part of the agencies' baseline. The approach used for this proposal does
not raise this concern, as the underlying data comes from model year 2008 production.
On the other hand, manufacturers' plans for MY 2010 and MY 2011 are less subject to
this concern, because manufacturers' expectations regarding these model years were
informed by knowledge of the MY 2010 and MY 2011 standards. The agencies are
hopeful that manufacturers will agree to make public their plans for one or both of these
model years, potentially facilitating the development of a baseline fleet that better
balances multiple concerns.
In addition, by developing a baseline fleet from common sources, the agencies
have been able to avoid some errorsperhaps related to interpretation of requeststhat
have been observed in past responses to NHTSA's requests. For example, while
reviewing information submitted to support the March 2009 CAFE final rule, NHTSA
staff discovered that one manufacturer had misinterpreted instructions regarding the
specification of vehicle track width, leading to important errors in estimates of vehicle
footprints. Although the manufacturer resubmitted the information with corrections after
being alerted by the agency, this new approach helps the agencies reduce the potential for
such errors and inconsistencies by utilizing common data sources and procedures.
1-27
-------
Draft Joint Technical Support Document
An additional advantage of the approach used for this proposal is a consistent
projection of the change in fuel economy and COi emissions across the various vehicles
from the application of new technology. In the past, manufacturer product plans would
include the application of new fuel economy improvement technology for a new or
improved vehicle model with the resultant estimate from the manufacturer of the fuel
economy levels for the vehicle. However, manufacturers did not always provide to
NHTSA detailed analysis which showed how they forecasted what the fuel economy
performance of the new vehicle was - that is, whether it came from actual test data, from
vehicle simulation modeling, from best engineering judgment or some other
methodology. Thus, it was not always possible for NHTSA to review the methodology
used by the manufacturer, nor was it always possible to review what approach the
different manufacturers utilized from a consistency perspective. With the approach used
for this proposal, the baseline market data comes from actual vehicles which have actual
fuel economy test data, so there is no question what is the basis for the fuel economy or
CC>2 performance of the baseline market data, as it is actual measured data.
Another advantage of today's approach is that future market shares are based on a
forecast of what will occur in the future, rather than on a static value. In the past,
NHTSA has utilized a constant market share for each model year, based on the most
recent year available for example from the CAFE compliance data, that is, a forecast of
the 2011-2015 time frame where company market shares do not change. In the approach
used today, the forecasts from CSM have been utilized to determine how future market
shares among the companies may change over time.0
The approach that NHTSA and EPA have taken in developing the market forecast
does, however, have some disadvantages. Most importantly, it produces a market
forecast that does not represent some important changes likely to occur in the future.
Some of the changes not captured by this approach are specific. For example, the
current market forecast includes some vehicles for which manufacturers have announced
plans for elimination or drastic production cuts such as the Chevrolet Trailblazer, the
Chrysler PT Cruiser, the Chrysler Pacifica, the Dodge Magnum, the Ford Crown
Victoria, the Hummer H2, the Mercury Sable, the Pontiac Grand Prix, and the Pontiac
G5. These vehicle models appear explicitly in market inputs to NHTSA's analysis, and
are among those vehicle models included in the aggregated vehicle types (representative
vehicles) appearing in market inputs to EPA's analysis.
Conversely, the agencies' market forecast does not include some forthcoming
vehicle models, such as the Chevrolet Volt, the Chevrolet Camaro, the Ford Fiesta and
several publicly announced electric vehicles, including the announcements from Nissan.
Nor does it include several MY 2009 and 2010 vehicles, such as the Honda Insight, the
Hyundai Genesis and the Toyota Venza, as the starting point for vehicle definitions was
MY 2008. Additionally, the market forecast does not account for publicly-announced
technology introductions, such as Ford's EcoBoost system, whose product plans specify
c The agencies note, however, that CSM or other data could also be applied to product plan data in order to
adjust market share over time.
1-28
-------
The Baseline Vehicle Fleet
which vehicles and how many are planned to have this technology. Were the agencies to
rely on manufacturers' product plans, today's market forecast would account for not only
these specific examples, but also for similar examples that have not yet been announced
publicly.
The agencies believe that including specific new vehicle models such as the Volt
or Insight after MY 2008 would most likely have a small impact on the estimates of the
final technology required to comply with the proposed standards. However, including
them would make the standards appear to cost less relative to the reference case. First,
the projections of sales by vehicle segment and manufacturer include these expected new
vehicle models. Thus, to the extent that these new vehicles are expected to change
consumer demand, they should be reflected in the reference case. While projecting the
characteristics of the new vehicles with MY 2008 vehicles, the primary difference
between the new vehicles and 2008 vehicles in the same vehicle segment is the use of
additional CCVreducing/fuel-saving technology. Both the NHTSA and EPA models add
such technology to facilitate compliance with the proposed standards. Thus, the future
projections of the vehicle fleet generally shift vehicle designs towards those of these
newer vehicles. The advantage of the agencies approach is that it helps clarify the costs
of this proposal, as the cost of all fuel economy improvements beyond those required by
the MY 2011 CAFE standards are being assigned to the proposal. In some cases, the new
vehicles being introduced by manufacturers are actually in response to their anticipation
of this rulemaking. The agencies approach prevents some of these technological
improvements and their associated cost from being assumed in the baseline. Thus, the
added technology will not be considered to be free for the purposes of this rule.
We note that, as a result of these issues, the market file may show sales volumes
for certain vehicles during MYs 2012-2016 even though they will be discontinued before
that time frame. Although the agencies recognize that these specific vehicles will be
discontinued, they are included in the market forecast because they are useful for
representing successor vehicles that may appear in the rulemaking time frame to replace
the discontinued vehicles in that market segment.
Other market changes not captured by today's approach are broader. For
example, Chrysler Group LLC has announced plans to offer small- and medium-sized
cars using Fiat powertrains. The product plan submitted by Chrysler includes vehicles
that appear to reflect these plans. However, none of these specific vehicle models are
included in the market forecast the agencies have developed starting with MY 2008
CAFE compliance data. The product plan submitted by Chrysler is also more optimistic
with regard to Chrysler's market share during MYs 2012-2016 than the market forecast
projected by CSM and used by the agencies today.
Additionally, some technical information manufacturers that have provided in
product plans regarding specific vehicle models is, at least as far as NHTSA and EPA
have been able to determine, not available from public or commercial sources. While
such gaps do not bear significantly on the agencies' analysis, the diversity of pickup
1-29
-------
Draft Joint Technical Support Document
configurations necessitated utilizing a sales-weighted average footprint valued for many
manufacturers' pickups. Since the modeling only utilizes footprint in order to estimate
each manufacturer's CC>2 or fuel economy standard and all the other vehicle
characteristics are available for each pickup configuration, this approximation has no
practical impact on the projected technology or cost associated with compliance with the
various standards evaluated. The only impact which could arise would be if the relative
sales of the various pickup configurations changed, or if the agencies were to explore
standards with a different shape or attribute. This would necessitate recalculating the
average footprint value in order to maintain accuracy.
The agencies have carefully considered these advantages and disadvantages of
using a market forecast derived from public and commercial sources rather than from
manufacturers' product plans, and the agencies believe that the advantages outweigh the
disadvantages for the purpose of proposing standards for model years 2012-2016.
NHTSA's inability to release confidential market inputs and corresponding detailed
outputs from the CAFE model has raised serious concerns among many observers
regarding the transparency of NHTSA's analysis, as well as related concerns that the lack
of transparency might enable manufacturers to provide unrealistic information to try to
influence NHTSA's determination of the maximum feasible CAFE levels. While
NHTSA does not agree with some observers' assertions that some manufacturers have
deliberately provided inaccurate or otherwise misleading information, today's market
forecast is fully open and transparent, and is therefore not subject to such concerns.
With respect to the disadvantages, the agencies are hopeful that manufacturers
will, in the future, agree to make public their plans regarding model years that are very
near, such as MY 2010 or perhaps MY 2011, so that this information can be incorporated
into an analysis that is available for public review and comment. In any event, because
NHTSA and EPA are releasing market inputs used in the agencies' respective analyses,
manufacturers, suppliers, and other automobile industry observers and participant can
submit comments on how these inputs should be improved, as can all other reviewers.
d A full-size pickup might be offered with various combinations of cab style (e.g., regular, extended, crew)
and box length (e.g., 5l/2, 6l/2, 8') and, therefore, multiple footprint sizes. CAFE compliance data for
MY2008 data does not contain footprint information, and does not contain information that can be used to
reliably identify which pickup entries correspond to footprint values estimable from public or commercial
sources. Therefore, the agencies have used the known production levels of average values to represent all
variants of a given pickup line (e.g., all variants of the F-150 and the Sierra/Silverado) in order to calculate
the sales-weighted average footprint value for each pickup family. Again, this has little impact on the
results of the modeling effort. In the extreme, one single footprint value could be used for every vehicle
sold by a single manufacturer as long as the fuel economy standard associated with this footprint value
represented the sales-weighted, harmonic average of the fuel economy standards associated with each
vehicle's footprint values.
1-30
-------
The Baseline Vehicle Fleet
1.3.5 How does manufacturer product plan data factor into the baseline used
in this proposal?
In the spring of 2009, many manufacturers submitted product plans in response to
NHTSA's recent request that they do so. NHTSA and EPA both have access to these
plans, and both agencies have reviewed them in detail. A small amount of product plan
data was used in the development of the baseline. The specific pieces of data are:
Wheelbase
Track Width Front
Track Width Rear
EPS (Electric Power Steering)
ROLL (Reduced Rolling Resistance)
LUB (Advance Lubrication i.e. low weight oil)
IACC (Improved Electrical Accessories)
Curb Weight
GVWR (Gross Vehicle Weight Rating)
The track widths, wheelbase, curb weight, and GVWR could have been looked up
on the internet (159 were), but were taken from the product plans when available for
convenience. To ensure accuracy, a sample from each product plan was used as a check
against the numbers available from Motortrend.com. These numbers will be published in
the baseline file since they can be easily looked up on the internet. On the other hand,
EPS, ROLL, LUB, and IACC are difficult to determine without using manufacturer's
product plans. These items will not be published in the baseline file, but the data has
been aggregated into the agencies' baseline in the technology effectiveness and cost
effectiveness for each vehicle in a way that allows the baseline for the model to be
published without revealing the manufacturers data.
1-31
-------
Draft Joint Technical Support Document
References
1 CFEIS (Certification Fuel Emissions Information System) Database: CFEIS is an EPA database system.
2 WardsAuto.com: Used as a source for engine specifications.
3 Motortrend.com and Edmunds.com: Used as a source for foot print and vehicle weight data.
4
Energy Information Administration's 2009 Annual Energy Outlook.
CSM World Wide, CSM World Wide is a paid service provider; I.D. Power Automotive Forecasting, I.D. Power
Automotive Forecasting is a paid service provider.
6 The baseline Excel file ("2008-2016 Production Summary Data .Definitions Docket 08_27_2009") is
available in the docket (Docket EPA-HQ-OAR-2009-0472).
7 "CSM North America Sales Forecast Comparison 2Q09 3Q09 For Docket." 2nd and 3rd quarter
forecasting results from CSM World Wide (Docket EPA-HQ-OAR-2009-0472).
8 See 49 CFR Part 512.
1-32
-------
What are the Attribute-Based Curves the Agencies are Using, and How Were They Developed?
CHAPTER 2: What are the Attribute-Based Curves the Agencies
are Using, and How Were They Developed?
2.1 Standards are attribute-based and defined by a mathematical function
As discussed in Section II.B of the NPRM, NHTSA and EPA are setting attribute-
based CAFE and COi standards for MYs 2012-2016 passenger cars and light trucks which
standards are expressed by a mathematical function. EPCA, as amended by EISA, expressly
requires that CAFE standards for passenger cars and light trucks be based on one or more
vehicle attributes related to fuel economy, and be expressed in the form of a mathematical
function.1 The CAA has no such requirement, although such an approach is permissible
under section 202 (a) and EPA has used the attribute-based approach in issuing standards
under analogous provisions of the CAA (e.g., standards for non-road diesel engines using
engine size as the attribute). Given the advantages of using attribute-based standards for light
duty vehicles and given the goal of coordinating and harmonizing CC>2 standards promulgated
under the CAA and CAFE standards promulgated under EPCA, as expressed by President
Obama in his announcement of the new National Program and by the agencies in the joint
NOT, EPA is also proposing standards that are attribute-based and expressed as mathematical
functions.
Under an attribute-based standard approach, each vehicle model has a fuel economy or
CC>2 target, the stringency of which depends of the value of the attribute for the vehicle
model. Thus, fuel economy and COi targets are set for individual vehicles, and vary directly
with the attribute, i.e., becoming more stringent as the attribute decreases and vice versa. For
example, size-based (i.e., size-indexed) standards assign higher fuel economy targets (lower
COi targets) to smaller vehicles and lower fuel economy targets (higher CC>2 targets) to larger
vehicles. The CAFE or fleet-wide average COi level required of a particular manufacturer is
determined by calculating the sales-weighted average of the model-specific fuel economy or
CO2 targets.
The form of the fuel economy or COi standard can have a dramatic effect on the level
of improvement required from various types of vehicles and the relative stringency of the
standard for various manufacturers. These disparate impacts derive from two facts
concerning the current automotive market: 1) consumers currently demand a wide variety of
vehicle types and sizes, and 2) individual manufacturers focus their product offerings on
different portions of this demand. Therefore, a comparison of the advantages and
disadvantages of attribute-based and flat (universal) standards needs to consider both of these
factors.
Methods to reduce the CO2 emissions from and increase the fuel economy of
individual vehicles can be generally grouped into four main categories: 1) reducing vehicle
size, 2) reducing vehicle performance and/or utility, and 3) increasing the efficiency of the
powertrain and accessories maintaining constant performance and utility, and 4) reducing
2-1
-------
Draft Joint Technical Support Document
weight and aerodynamic drag while maintaining vehicle size and utility.a Stringent flat
standards encourage all four of these methods. However, the level of encouragement differs
between manufacturers. Manufacturers of generally larger vehicles face a more difficult task
in meeting a flat standard than those of generally smaller vehicles. Thus, the stringency of
flat standards will tend to be limited by the capabilities of the manufacturers of larger
vehicles. Manufacturers of smaller vehicles will tend to face lower compliance costs and may
not even need to change their vehicles at all.b Thus, the overall level of control may be
limited. The flat standard will encourage smaller, less powerful vehicles, as these vehicles
face the least compliance costs. Sales will tend to be shifted to manufacturers of generally
smaller vehicles as their overall compliance costs will be lower than those of manufacturers of
average or larger vehicles. Finally, because smaller vehicles face lower costs than larger
vehicles, consumers will be encouraged to purchase smaller vehicles.
Attribute-based standards are preferable to universal industry-wide average standards
for several reasons. First, if the shape is chosen properly, every manufacturer is more likely
to be required to continue adding more fuel efficient technology each year, because the level
of the compliance obligation of each manufacturer is based on its own particular product mix.
Second, attribute-based standards minimize the incentive for manufacturers to respond
to CAFE and CC>2 standards in ways that may be harmful to safety.0'2 Because each vehicle
model has its own target (based on the attribute chosen), attribute-based standards can,
depending on the corresponding mathematical function (i.e., the shape of the attribute based
standards), reduce or eliminate the incentive to build smaller vehicles simply to meet a fleet-
wide average. Since smaller vehicles are subject to more stringent fuel economy targets, a
manufacturer's increasing its proportion of smaller vehicles would not necessarily ease its
compliance obligation. For the purposes of this rule, the agencies believe that setting
attribute-based standards with careful attention to the slope of the target curves will not
encourage wholesale changes in fleet mix to either smaller or larger vehicles.
Third, attribute-based standards provide a more equitable regulatory framework for
different vehicle manufacturers.11'3 A universal industry-wide average standard imposes
disproportionate compliance obligations on manufacturers that need to change their product
plans and no obligation on those manufacturers that have no need to change their plans.
a Another method of increasing fleet fuel economy is to increase the number of high-fuel-economy vehicles and
reduce sales of low-fuel-economy vehicles, for instance, by changing their relative prices. EPA and NHTSA
believe that existing technologies allow automakers to meet the requirements of the proposed rules without
changing the mix of vehicles sold (although the proposed rules would not prohibit such changes in mix).
b Widespread trading of credits could encourage fuel economy improvements or CO2 emission reductions from
manufacturers of smaller vehicles. This would allow the stringency of the standard to be increased, but would
probably create a shift of wealth between the two sets of manufacturers. However, most manufacturers have
indicated that they would be unlikely to trade credits with their direct competitors.
c The 2002 NAS Report described at length and quantified the potential safety problem with average fuel
economy standards that specify a single numerical requirement for the entire industry. See NAS Report at 5,
finding 12.
d Id. at 4-5, finding 10.
2-2
-------
What are the Attribute-Based Curves the Agencies are Using, and How Were They Developed?
Attribute-based standards (that have the proper shape) spread the regulatory cost burden for
fuel economy more broadly across all of the vehicle manufacturers within the industry.
And fourth, attribute-based standards respect economic conditions and consumer
choice. Manufacturers are required to invest in technologies that improve the fuel economy
of their fleets, regardless of vehicle mix. Thus the consumer is likely to enjoy the same wide
range of vehicles to choose from in 2016 as they do now. The difference will be, of course,
that most of the vehicles will be more fuel efficient.
The agencies recognize that, because manufacturers' compliance obligations under
attribute-based standards are based in part on the mix of vehicles that they produce, the fuel
savings and emissions reductions produced under attribute-based standards can vary
depending on market conditions. For example, fuel prices lower than those anticipated at the
time of rulemaking will tend to shift consumer demand toward larger vehicles. If
manufacturers sell a greater number of larger vehicles than the agencies anticipate, fuel
savings and CCh reductions would be lower than anticipated. In contrast, if fuel prices rise
significantly, more fuel savings and COi reductions than anticipated should be likely.
One potential way to mitigate the variability of results under attribute-based standards
due to market conditions is through the use of explicit backstops, standards below which
manufacturers may not drop. For purposes of the CAFE program, EISA requires a backstop
for domestically-manufactured passenger carsa universal minimum, non-attribute-based
standard of either "27.5 mpg or 92 percent of the average fuel economy projected by the
Secretary of Transportation for the combined domestic and non-domestic passenger
automobile fleets manufactured for sale in the United States by all manufacturers in the model
year.. ."4, whichever is greater. In the MY 2011 final rule, the first rule setting standards since
EISA added the backstop provision to EPCA, NHTSA considered whether the statute
permitted the agency to set backstop standards for the other regulated fleets of imported
passenger cars and light trucks. Although commenters expressed support both for and against
a more permissive reading of EISA, NHTSA concluded in that rulemaking that its authority
was likely limited to setting only the backstop standard that Congress expressly provided, i.e.,
the one for domestic passenger cars. A backstop, however, is permissible under the CAA (as
are universal standards generally).
For purposes of the CAFE and COi standards proposed in this NPRM, NHTSA and
EPA recognize that the risk, even if small, does exist that low fuel prices in MYs 2012-2016
might lead indirectly to less than currently anticipated fuel savings and emissions reductions.
The NPRM seeks comment on whether backstop standards, or any other method within the
agencies' statutory authority, should and can be implemented for the import and light truck
fleets in order to achieve the fuel savings that attribute-based standards might not absolutely
guarantee.
2-3
-------
Draft Joint Technical Support Document
2.2 What attribute do the agencies use, and why?
Consistent with the MY 2011 CAFE standards, EPA and NHTSA are proposing to use
footprint as the attribute for the MY 2012-2016 CAFE standards and CO2 emissions
standards. There are several policy reasons why the agencies believe that footprint is the most
appropriate attribute on which to base the standards, as discussed below.
It is important that the CAFE and COi standards be set in a way that does not
encourage manufacturers to respond by selling vehicles that are in any way less safe.
Although NHTSA's research also indicates that reductions in vehicle mass tend to
compromise vehicle safety, footprint-based standards provide an incentive to use new and
advanced lightweight materials and structures that could be otherwise discouraged by weight-
based standards, because manufacturers can use them to improve a vehicle's fuel economy
without their use necessarily resulting in a change in the vehicle's target level of fuel
economy or COi emissions and without a substantial impact on the safety (in terms of
crashworthiness) of that vehicle.
Further, although the agencies recognize that weight is better correlated with fuel
economy and COiemissions than is footprint, the agencies continue to believe that there is
less risk of "gaming" (artificial manipulation of the attribute(s) to achieve a more favorable
target) by increasing footprint under footprint-based standards than by increasing vehicle
mass under weight-based standardsit is relatively easy for a manufacturer to add enough
weight to a vehicle to decrease its applicable fuel economy target a significant amount, as
compared to increasing vehicle footprint. The agencies also agree with concerns raised in
2008 by some commenters to NHTSA's MY 2011 CAFE rulemaking that there would be
greater potential for gaming under multi-attribute standards, such as standards under which
targets would also depend on attributes such as weight, torque, power, towing capability,
and/or off-road capability. Standards that incorporate such attributes in conjunction with
footprint would not only be significantly more complex, but by providing degrees of freedom
with respect to more easily-adjusted attributes, they would make it less certain that the future
fleet would actually achieve the average fuel economy improvement and CO2 reduction levels
projected by the agencies.
However, although NHTSA and EPA believe initially that footprint is the most
appropriate attribute upon which to base the proposed standards (and the MY 2011 CAFE
standards use footprint as the exclusive attribute), recognizing strong public interest in this
issue, the NPRM seeks comment on whether the agencies should consider setting standards
for the final rule based on another attribute or a combination of attributes. If commenters
suggest that the agencies should consider another attribute or a combination of attributes, the
agencies specifically request that the commenters address the concerns raised in the
paragraphs above regarding the use of other attributes, and explain how standards should be
developed using the other attribute(s) in a way that contributes more to fuel savings and COi
reductions than the footprint-based standards, without compromising safety and while
maintaining the variety of vehicle choices that consumers enjoy today.
2-4
-------
What are the Attribute-Based Curves the Agencies are Using, and How Were They Developed?
2.3 What mathematical function do the agencies use, and why?
The MY 2011 CAFE standards are defined by a continuous, constrained logistic
function, which takes the form of an S-curve, and is defined according to the following
formula:
I
TARGET =
Here, TARGET is the fuel economy target (in mpg) applicable to vehicles of a given
footprint (FOOTPRINT, in square feet), b and a are the function's lower and upper
asymptotes (also in mpg), e is approximately equal to 2.718,6 c is the footprint (in square feet)
at which the inverse of the fuel economy target falls halfway between the inverses of the
lower and upper asymptotes, and d is a parameter (in square feet) that determines how
gradually the fuel economy target transitions from the upper toward the lower asymptote as
the footprint increases. Figure 2-1 shows an example of a logistic target function, where b =
20 mpg, a= 30 mpg, c = 40 square feet, and d = 5 square feet.
35
30
25 -
9 20-
15 -
10 -
5 -
LIMIT "
LOWER
10
30 40 50 60 70
Footprint (Square Feet)
80
100
Figure 2-1 Sample Logistic Curve
e e is the irrational number for which the slope of the function y = number" is equal to 1 when x is equal to zero.
The first 8 digits ofe are 2.7182818.
2-5
-------
Draft Joint Technical Support Document
After fitting this mathematical form (separately) to the passenger car and light truck
fleets and determining the maximum feasible stringency of the standards (i.e., the vertical
positions of the curves), NHTSA arrived at the curves shown in Figure 2-2 to define the MY
2011 standards.
45
40 -
35 -
30 -
25 -
20
^^ Passenger Cars
Light Trucks
35 40 45 50 55 60 65 70
Footprint (sf)
Figure 2-2 MY 2011 CAFE Standards for Passenger Cars and Light Trucks
In finalizing the MY 2011 standards, NHTSA noted that the agency is not required to
use a constrained logistic function and indicated that the agency may consider defining future
CAFE standards in terms of a different mathematical function. NHTSA and EPA have done
so jointly in preparation for the proposed CAFE standards and COi emissions standards.
In revisiting this question jointly, EPA and NHTSA found that the final MY 2011
CAFE standard for passenger cars, though less steep than the MY 2011 standard NHTSA
proposed in 2008, continues to concentrate the sloped portion of the curve (from a compliance
perspective, the area in which upsizing results in a slightly lower applicable target) within a
relatively narrow footprint range (approximately 47-55 square feet). In practical terms, this
results in a portion of the fleet that has a large change in emissions and fuel consumption with
just a minor change in footprint. This potentially increases the incentive for gaming by
2-6
-------
What are the Attribute-Based Curves the Agencies are Using, and How Were They Developed?
upsizing slightly to meet less stringent standards. Further, most passenger car models have
footprints smaller than the curve's 51.4 square foot inflection point, and many passenger car
models have footprints at which the curve is relatively flat (see figure above). The resulting
fit to the data was forced by choice of asymptotes defining the constrained logistic function.
The agencies discuss below an alternative function that seeks to prevent this from occurring.
For both passenger cars and light trucks, a mathematical function that has some slope
at most footprints where vehicles are produced is advantageous in terms of fairly balancing
regulatory burdens among manufacturers, and in terms of providing a disincentive to respond
to new standards by downsizing vehicles in ways that compromise vehicle safety. For
example, a flat standard has the drawbacks of a universal standard: it may be very difficult for
a full-line manufacturer to meet, while requiring very little of a manufacturer concentrating on
small vehicles, and a flat standard may provide an incentive to manufacturers to downsize
certain vehicles, in order to "balance out" other vehicles subject to the same standard.
As a potential alternative to the constrained logistic function, NHTSA had, in
proposing MY 2011 standards, presented information regarding a constrained linear function
(or piecewise linear function). As shown in the 2008 NPRM, a constrained linear function
has the potential to avoid creating a localized region (in terms of vehicle footprint) over which
the slope of the function is relatively steep. Although NHTSA did not receive public
comments on this option, the agency indicated that it still believed a linear function
constrained by upper (on a gpm basis) and possibly lower limits could merit reconsideration
in future CAFE rulemakings.
Having re-examined a piecewise linear function or constrained linear function for
purposes of the proposed standards, NHTSA and EPA tentatively conclude that for both
passenger cars and light trucks, it remains meaningfully sloped over a wide footprint range,
thereby providing a well-distributed disincentive to downsize vehicles in ways that could
compromise highway safety. Also, the function proposed is not so steeply sloped that it
would provide a strong incentive to increase vehicle size in order to obtain a lower CAFE
requirement and higher CO2 limit, thereby compromising energy and environmental benefits.
Therefore, the CAFE and CC>2 emissions standards proposed in the NPRM are defined by
constrained linear or piecewise linear functions.
The NHTSA constrained linear function is defined according to the following formula:
1
TARGET = -
MIN
MAX\ cxFOOTPRINT + d,-\-
( a b
Here, TARGET is the fuel economy target (in mpg) applicable to vehicles of a given
footprint (FOOTPRINT, in square feet), b and a are the function's lower and upper
asymptotes (also in mpg), respectively, c is the slope (in gpm per square foot) of the sloped
portion of the function, and d is the intercept (in gpm) of the sloped portion of the function
(that is, the value the sloped portion would take if extended to a footprint of 0 square feet.
The MIN and MAX functions take the minimum and maximum, respectively of the included
2-7
-------
Draft Joint Technical Support Document
values; for example, MIN(1,2) = 1, MAX(1,2) = 2, and MIN[MAX(1,2),3)]=2. Figure 2-3
shows an example of a linear target function, where a = 0.0241 gpm (41.6 mpg), b = 0.032
gpm (31.2 mpg), c = 0.000531 gpm per square foot, and d = 0.002292 gpm (436 mpg).
Because the function is linear on a gpm basis, not an mpg basis, it is plotted on this basis.
I
0.050 -,
0.045 -
0.040 -
0.035 -
0.030 -
0.025 -
0.020 "^
0.015 -
0.010 -
0.005 -
0.000
I/LIMIT,.
SLOPE ,.--^-
* - ll^--"'"""""*^
1/LIMITUP
--INTERCEPT
35
40
45
50
Footprint (sf)
55
60
65
70
Figure 2-3 Sample Linear Function
The EPA piecewise linear function is defined according to the following
formula:
Target = a, if x < 1
Target = ex + d, if 1 < x < h
Target = b, if x > h
This equation takes the simplified form (consistent with NHTSA's function):
Target = MIN [ MAX ( c * x + d , a), b]
Where,
2-8
-------
What are the Attribute-Based Curves the Agencies are Using, and How Were They Developed?
Target = the CO2 target value for a given footprint (in g/mi)
a = the minimum target value (in g/mi COi)
b = the maximum target value (in g/mi CO2)
c = the slope of the linear function (in g/mi per sq ft CCh)
d = is the intercept or zero-offset for the line (in g/mi CO2)
x = footprint of the vehicle model (in square feet, rounded to the nearest tenth)
1 & h are the lower and higher footprint limits or constraints or the boundary ("kinks")
between the flat regions and the intermediate sloped line (in sq ft)
Figure 2-4 shows the basic curve.
o> ra
-------
Draft Joint Technical Support Document
technologically heterogeneous, NHTSA used the CAFE model to develop a fleet to which
nearly all the technologies listed in Ch. 3 of the TSDf were applied, by taking the following
steps: (1) treating all manufacturers as unwilling to pay civil penalties rather than applying
technology, (2) applying any technology at any time, irrespective of scheduled vehicle
redesigns or freshening, and (3) ignoring "phase-in caps" that constrain the overall amount of
technology that can be applied by the model to a given manufacturer's fleet. These steps
helped to increase technological parity among vehicle models, thereby providing a better basis
(than the baseline fleet) for estimating the statistical relationship between vehicle size and fuel
economy. In other words, this process normalized the fleet such that differences in
technology had a minimal contribution to the variation (scatter) in the data. This process also
served to slightly flatten the slope of the line (a goal sought after as described above).
In fitting the curves, EPA and NHTSA also continued to apply constraints to limit the
function's value for both the smallest and largest vehicles. Without a limit at the smallest
footprints, the functionwhether logistic or linearcan reach values that would be unfairly
burdensome for a manufacturer that elects to focus on the market for small vehicles;
depending on the underlying data, an unconstrained form could apply to the smallest vehicles
targets that are simply unachievable. Limiting the function's value for the smallest vehicles
ensures that the function remains technologically achievable at small footprints, and that it
does not unduly burden manufacturers focusing on small vehicles. Obviously, passenger
compartments can only become so small; this means, among other things, that a vehicle's
frontal areaa key determinant of aerodynamic dragcan only become so small. Also, as
vehicle size decreases, complying with federal motor vehicle safety standards can necessitate
the addition of structural elements that add mass and thereby reduce fuel economy and
increase emissions. On the other side of the function, without a limit at the largest footprints,
the function may provide no floor on required fuel economy. Also, the safety considerations
that support the provision of a disincentive for downsizing as a compliance strategy apply
weaklyif at allto the very largest vehicles. Limiting the function's value for the largest
vehicles leads to a function with an inherent absolute minimum level of performance, while
remaining consistent with safety considerations.
Before fitting the sloped portion of the constrained linear form, NHTSA and EPA
selected footprints above and below which to apply constraints (i.e., minimum and maximum
values) on the function. For passenger cars, the agencies noted that several manufacturers
offer small and, in some cases, sporty coupes below 41 square feet, examples including the
BMW Z4 and Mini, Saturn Sky, Honda Fit and S2000, Hyundai Tiburon, Mazda MX-5
Miata, Suzuki SX4, Toyota Yaris, and Volkswagen New Beetle. Because such vehicles
represent a small portion (less than 10 percent) of the passenger car market, yet often have
characteristics that could make it infeasible to achieve the very challenging targets that could
apply in the absence of a constraint, the agencies are proposing to "cut off the linear portion
of the passenger car function at 41 square feet. For consistency, the agencies are proposing to
f The agencies excluded diesel engines and strong hybrid vehicle technologies from this exercise (and only this
exercise) because the agencies expect that manufacturers would not need to rely heavily on these technologies in
order to comply with the proposed standards. NHTSA and EPA did include diesel engines and strong hybrid
vehicle technologies in all other portions of their analyses.
2-10
-------
What are the Attribute-Based Curves the Agencies are Using, and How Were They Developed?
do the same for the light truck function, although no light trucks are currently offered below
41 square feet. The agencies further noted that above 56 square feet, the only passenger car
model present in the MY 2008 fleet were four luxury vehicles with extremely low sales
volumesthe Bentley Arnage and three versions of the Rolls Royce Phantom. The agencies
are therefore proposing to "cut off the linear portion of the passenger car function at 56
square feet. Finally, the agencies noted that although public information is limited regarding
the sales volumes of the many different configurations (cab designs and bed sizes) of pickup
trucks, the largest pickups (e.g., the Ford F-150, GM Sierra/Silverado, Nissan Titan, and
Toyota Tundra), appear to fall above 66 square feet in footprint. The agencies are therefore
proposing to "cut off the linear portion of the light truck function at 66 square feet. These
"cut-off values of footprint serve as the constraints to the functions shown above. The lower
and upper levels are determined from where the sloped line meets these cut-off points.
In the NPRM, EPA and NHTSA invite comment on this approach to fitting the curves.
The agencies note that final decisions on this issue will play an important role in determining
the form and stringency of the final standards, the incentives those standards will provide
(e.g., with respect to downsizing small vehicles), and the relative compliance burden faced by
each manufacturer.
Having developed a set of baseline data to which to fit the mathematical fuel
consumption function, the initial values for parameters c and d were determined for cars and
trucks separately. Values for c and d were initially set such that the average (equivalently,
sum) of the absolute values of the differences was minimized between the "maximum
technology" fleet fuel consumption (within the footprints between the upper and lower limits)
and the straight line function defined above at the same corresponding vehicle footprints.
That is, c and d were determined by minimizing the average absolute residual, commonly
known as the MAD (Mean Absolute Deviation) approach, of the corresponding straight line.
The curve was fit in fuel consumption (i.e., CCh) space rather than fuel economy space
because the manufacturer targets are in terms of the harmonic average fuel economy (or
average CCh), and so it is more important that the curve fit the fuel consumption (i.e., CCh)
data well than that it fit the fuel economy data well. NHTSA also explained in the MY 2011
final rule that it chose to use MAD in this Step instead of minimizing the sum of the square
errors ("least squares," another common approach in curve fitting) in order to lessen the
influence of outliers. NHTSA believes that it is more appropriate to use unweighted data in
fitting the curve rather than weighting the data by sales because of large variations in model
sales.
Finally, the agencies calculated the values of the upper and lower limits based on the
corresponding footprint constraints discussed above (41 and 56 square feet for passenger cars,
and 41 and 66 square feet for light trucks).
The result of this methodology is shown below in Figure 2-5 and Figure 2-6 for
passenger cars and light trucks, respectively. The fitted curves are shown with the underlying
"maximum technology" passenger car and light truck fleets. For passenger cars, the mean
absolute deviation of the sloped portion of the function was 14 percent. For trucks, the
corresponding MAD was 10 percent.
2-11
-------
Draft Joint Technical Support Document
0.060 -,
0.055 -
0.050 -
0.045 -
0.040 -
0.035 -
0.030 -
0.025 -
0.020 -
0.015 -
0.010 -
0.005 -
0.000 -
35
. : . . .
.|. » j ...i* .. .: . :
, . .* »« !
:»:.#::.
40
45
50
55 60
Footprint (sf)
65
70
75
80
Figure 2-5 "Maximum Technology" Passenger Fleet with Fitted Constrained Linear Function
2-12
-------
What are the Attribute-Based Curves the Agencies are Using, and How Were They Developed?
0.060
Figure 2-6 "Maximum Technology" Light Truck with Fitted Constrained Linear Function
The agencies used these functional forms as a starting point to develop mathematical
functions defining actual proposed standards. As discussed in Sections II.B of the NPRM
preamble, the agencies transposed these functions vertically (i.e., on a gpm basis, uniformly
downward) to produce the maximum feasible passenger car and light truck CAFE standards,
and corresponding COi emissions standards.
2-13
-------
Draft Joint Technical Support Document
References
1 49 U.S.C. 32902(a)(3)(A).
2 The 2002 NAS Report described at length and quantified the potential safety problem with average fuel
economy standards that specify a single numerical requirement for the entire industry. See NAS Report at 5,
finding 12.
3 Id. at 4-5, finding 10.
4 49 U.S.C. 32902(b)(4).
2-14
-------
Technologies Considered in the Agencies' Analysis
CHAPTER 3: Technologies Considered in the Agencies'
Analysis
This Chapter of the joint TSD describes the technologies NHTSA and EPA
evaluated as potential inputs in their respective models and provides estimates of the
technologies' costs, effectiveness and availability. This Chapter also describes, in general
terms, how the agencies use these inputs in their respective models. For greater detail on
this subject, please see Section V of NHTSA's PRIA and Chapter 4 of EPA's draft RIA.
3.1 How do the agencies decide which technologies to include in the
analysis?
Technology assumptions, i.e., assumptions about their availability, cost,
effectiveness, and the rate at which they can be incorporated into new vehicles, are often
very controversial as they have a significant impact on the levels of the standards.
Agencies must, therefore, take great care in developing and justifying these
assumptions. In developing technology inputs for MY 2012-2016 standards the agencies
reviewed the technology assumptions that NHTSA used in setting the 2011 standards and
the comments that NHTSA received in response to its May 2008 Notice of Proposed
Rulemaking. In addition, the agencies reviewed the technology input assumptions
identified in EPA's 2008 Staff Technical Report1 and supplemented their review with
information from the sources described in the following section. This analysis between
the agencies continued until the publication of this notice and will continue until the final
rule is published.
3.1.1 Reports and papers in the literature
NHTSA and EPA have done extensive research in identifying the most credible
sources of information. These sources included: the 2002 NAS report on the effectiveness
and impact of CAFE standards;2 the 2004 study done by NESCCAF;3 the California Air
Resources Board (CARB) Initial Statement of Reasons in support of their carbon
rulemaking;4 a 2006 study done by Energy and Environmental Analysis (EEA) for the
Department of Energy;5 a study done by the Martec Group for the Alliance of
Automobile Manufacturers, and an update by the Martec Group to that study;6 and
vehicle fuel economy certification data. Both agencies also reviewed the published
technical literature which addressed the issue of COi emission control and fuel economy,
such as papers published by the Society of Automotive Engineers and the American
Society of Mechanical Engineers. In addition, confidential data submitted by vehicle
manufacturers in response to NHTSA's request for product plans,7 and confidential
information shared by automotive industry component suppliers in meetings with EPA
and NHTSA staff held during the second half of the 2007 calendar year were used as a
cross check of the public data mentioned above and not as a significant basis for this
rulemaking. EPA also has a contracted study ongoing with FEV that consists of
complete system tear-downs to evaluate technologies down to the nuts and bolts to arrive
3-1
-------
Draft Joint Technical Support Document
at very detailed estimates of the costs associated with manufacturing them (and, as noted,
the agencies used this analysis to estimate costs of turbocharging with downsizing).8
EPA and NHTSA reviewed all this information in order to develop the best estimates of
availability, cost and effectiveness of these fuel-saving/CCVreducing technologies. Note
that tear-down cost estimates from FEV for additional engine and transmission
technologies became available shortly before the release of this proposal, but were not
incorporated into the agencies' cost analysis. These preliminary results are available for
review and the agencies will consider this information for use in the final rule analysis.9
A detailed report will be submitted to the docket on these additional technologies during
the public comment period for this proposal. In addition, additional technologies are
expected to be completed from this study with FEV after the close of the comment period
which the agencies will also docket.
The agencies would also like to note that per the Energy Independence and
Security Act (EISA), the National Academies of Sciences is conducting an updated study
to update chapter 3 of their 2002 NAS Report, which presents technology effectiveness
estimates. The update will take a fresh look at that list of technologies and their
associated cost and effectiveness values.
The updated NAS report is expected to be available on September 30, 2009. As
the Report is received by the agencies, it will be placed in the respective dockets for this
rulemaking for the public's review and comment. Because this is expected to occur
during the comment period, the public is encouraged to check the docket regularly and
provide comments on the updated NAS Report by the closing of the public comment
period. As requested by the President in the January 26, 2009 Executive Order, NHTSA
and EPA will consider the updated NAS Report and any comments received on it, as
appropriate, in developing the technology cost and effectiveness estimates for the final
rule.
3.1.2 Fuel economy certification data
Where available, data from recent model years of EPA's Fuel Economy
Certification Data was used. CO2 and fuel consumption reduction estimates were
estimated from EPA's fuel economy database on the two-cycle (FTP city & highway) fuel
economy test results. During the standard fuel economy test cycles, direct measurements
of COi emissions are made. This data, along with other measurements, are then used to
calculate the estimated fuel economy performance in gallons of fuel consumed per mile.
Vehicle certification data are an obviously reliable source for determination of the CC>2
and fuel consumption reduction potential when a directly comparable vehicle was offered
both with and without the specific CC>2 and fuel consumption reducing technology,
because a comparison between the emissions data between the two vehicles directly
reflect the application of the technologies on the vehicle test cycles. Where possible,
technology-specific effectiveness numbers were extracted for vehicles where only the
specific technology would be changed from a reference vehicle, in order to eliminate any
confounding of values across several technologies. In some hybrid vehicle cases, the
3-2
-------
Technologies Considered in the Agencies' Analysis
exact same vehicle may not be offered, and a similar vehicle was selected for
comparison.
3.2 Which technologies will be applicable in the rulemaking time
frame?
Fuel consumption and COi emissions reductions are possible from a variety of
technologies whether they be engine-related (e.g., turbocharging), transmission-related
(e.g., six forward gears in place of four), accessory-related (e.g., electric power steering),
or vehicle-related (e.g., low rolling resistance tires). One of the key factors that NHTSA
and EPA considered in developing the proposed standards for each model year is the
availability and feasibility of fuel-saving/CCVreducing technologies. In order to analyze
the availability and feasibility of technologies in an objective way, the agencies reviewed
NHTSA's MY 2011 CAFE final rule and EPA's 2008 Staff Report and updated the
assumptions based on more current information. As a check on the new analysis,
NHTSA and EPA also reviewed the product plans submitted by manufacturers to the
agencies in response to NHTSA's March 2009 request for product plans.
Many technologies considered by the agencies are well known, readily available
during the rulemaking time period, and could be incorporated into vehicles once
production decisions are made. The agencies did not consider technologies in the
research stage because their effectiveness and/or costs are presently only known with
greater levels of uncertainty.
The technologies considered by the NHTSA and EPA models are briefly
described below. They fall under the five broad categories of engine, transmission,
vehicle, electrification/accessory, and hybrid technologies. A more detailed description
of each technology, and the technology's costs and effectiveness, is described in greater
detail in section 3.5 of this TSD.
Types of engine technologies that improve fuel economy and reduce CC>2
emissions include the following:
Low-friction lubricants - low viscosity and advanced low friction lubricants
oils are now available with improved performance and better lubrication.
Reduction of engine friction losses - can be achieved through low-tension
piston rings, roller cam followers, improved material coatings, more optimal
thermal management, piston surface treatments, and other improvements in
the design of engine components and subsystems that improve engine
operation.
Conversion to dual overhead cam with dual cam phasing - as applied to
overhead valves designed to increase the air flow with more than two valves
per cylinder and reduce pumping losses.
Cylinder deactivation - deactivates the intake and exhaust valves and prevents
fuel injection into some cylinders during light-load operation. The engine
3-3
-------
Draft Joint Technical Support Document
runs temporarily as though it were a smaller engine which substantially
reduces pumping losses
Variable valve timing - alters the timing or phase of the intake valve, exhaust
valve, or both, primarily to reduce pumping losses, increase specific power,
and control residual gases.
Discrete variable valve lift - increases efficiency by optimizing air flow over
a broader range of engine operation which reduces pumping losses.
Accomplished by controlled switching between two or more cam profile lobe
heights.
Continuous variable valve lift - is an electromechanically controlled system in
which cam period and phasing is changed as lift height is controlled. This
yields a wide range of performance optimization and volumetric efficiency,
including enabling the engine to be valve throttled.
Stoichiometric gasoline direct-injection technology - injects fuel at high
pressure directly into the combustion chamber to improve cooling of the
air/fuel charge within the cylinder, which allows for higher compression ratios
and increased thermodynamic efficiency.
Combustion restart - can be used in conjunction with gasoline direct-injection
systems to enable idle-off or start-stop functionality. Similar to other start-
stop technologies, additional enablers, such as electric power steering,
accessory drive components, and auxiliary oil pump, might be required.
Turbocharging and downsizing - increases the available airflow and specific
power level, allowing a reduced engine size while maintaining performance.
This reduces pumping losses at lighter loads in comparison to a larger engine.
Exhaust-gas recirculation boost - increases the exhaust-gas recirculation used
in the combustion process to increase thermal efficiency and reduce pumping
losses.
Diesel engines - have several characteristics that give superior fuel efficiency,
including reduced pumping losses due to lack of (or greatly reduced)
throttling, and a combustion cycle that operates at a higher compression ratio,
with a very lean air/fuel mixture, than an equivalent-performance gasoline
engine. This technology requires additional enablers, such as NOX trap
catalyst after-treatment or selective catalytic reduction NOX after-treatment.
Types of transmission technologies considered include:
Improved automatic transmission controls - optimizes shift schedule to
maximize fuel efficiency under wide ranging conditions, and minimizes losses
associated with torque converter slip through lock-up or modulation.
3-4
-------
Technologies Considered in the Agencies' Analysis
Six-, seven-, and eight-speed automatic transmissions - the gear ratio spacing
and transmission ratio are optimized for a broader range of engine operating
conditions.
Dual clutch or automated shift manual transmissions - are similar to manual
transmissions, but the vehicle controls shifting and launch functions. A dual-
clutch automated shift manual transmission uses separate clutches for even-
numbered and odd-numbered gears, so the next expected gear is pre-selected,
which allows for faster and smoother shifting.
Continuously variable transmission - commonly uses V-shaped pulleys
connected by a metal belt rather than gears to provide ratios for operation.
Unlike manual and automatic transmissions with fixed transmission ratios,
continuously variable transmissions can provide fully variable transmission
ratios with an infinite number of gears, enabling finer optimization of
transmission torque multiplication under different operating conditions so that
the powertrain can operate at its optimum efficiency.
Manual 6-speed transmission - offers an additional gear ratio, often with a
higher overdrive gear ratio, than a 5-speed manual transmission.
Types of vehicle technologies considered include:
Low-rolling-resistance tires - have characteristics that reduce frictional losses
associated with the energy dissipated in the deformation of the tires under
load, therefore reducing the energy needed to move the vehicle.
Low-drag brakes - reduce the sliding friction of disc brake pads on rotors
when the brakes are not engaged because the brake pads are pulled away from
the rotors.
Front or secondary axle disconnect for four-wheel drive systems - provides a
torque distribution disconnect between front and rear axles when torque is not
required for the non-driving axle. This results in the reduction of associated
parasitic energy losses.
Aerodynamic drag reduction - is achieved by changing vehicle shape or
reducing frontal area, including skirts, air dams, underbody covers, and more
aerodynamic side view mirrors.
Mass reduction and material substitution - Mass reduction encompasses a
variety of techniques ranging from improved design and better component
integration to application of lighter and higher-strength materials. Mass
reduction is further compounded by reductions in engine power and ancillary
systems (transmission, steering, brakes, suspension, etc.).
3-5
-------
Draft Joint Technical Support Document
Types of electrification/accessory and hybrid technologies considered include:
Electric power steering (EPS) - is an electrically-assisted steering system that
has advantages over traditional hydraulic power steering because it replaces a
continuously operated hydraulic pump, thereby reducing parasitic losses from
the accessory drive.
Improved accessories (IACC) - may include high efficiency alternators,
electrically driven (i.e., on-demand) water pumps and cooling fans. This
excludes other electrical accessories such as electric oil pumps and electrically
driven air conditioner compressors.
Air Conditioner Systems - These technologies include improved hoses,
connectors and seals for leakage control. They also include improved
compressors, expansion valves, heat exchangers and the control of these
components for the purposes of improving tailpipe CC>2 emissions as a result
of A/C use. These technologies are covered separately in the EPA RIA.
12-volt micro-hybrid (MHEV) - also known as idle-stop or start stop and
commonly implemented as a 12-volt belt-driven integrated starter-generator,
this is the most basic hybrid system that facilitates idle-stop capability. Along
with other enablers, this system replaces a common alternator with an
enhanced power starter-alternator, both belt driven, and a revised accessory
drive system.
Higher Voltage Stop-Start/Belt Integrated Starter Generator (BISG) -
provides idle-stop capability and uses a high voltage battery with increased
energy capacity over typical automotive batteries. The higher system voltage
allows the use of a smaller, more powerful electric motor and reduces the
weight of the motor, inverter, and battery wiring harnesses. This system
replaces a standard alternator with an enhanced power, higher voltage, higher
efficiency starter-alternator, that is belt driven and that can recover braking
energy while the vehicle slows down (regenerative braking).
Integrated Motor Assist (IMA)/Crank integrated starter generator (CISC) -
provides idle-stop capability and uses a high voltage battery with increased
energy capacity over typical automotive batteries. The higher system voltage
allows the use of a smaller, more powerful electric motor and reduces the
weight of the motor, inverter, and battery wiring harnesses. This system
replaces a standard alternator with an enhanced power, higher voltage, higher
efficiency starter-alternator that is crankshaft mounted and can recover
braking energy while the vehicle slows down (regenerative braking).
2-mode hybrid (2MHEV) - is a hybrid electric drive system that uses an
adaptation of a conventional stepped-ratio automatic transmission by
replacing some of the transmission clutches with two electric motors that
control the ratio of engine speed to vehicle speed, while clutches allow the
3-6
-------
Technologies Considered in the Agencies' Analysis
motors to be bypassed. This improves both the transmission torque capacity
for heavy-duty applications and reduces fuel consumption and CC>2 emissions
at highway speeds relative to other types of hybrid electric drive systems.
Power-split hybrid (PSHEV) - a hybrid electric drive system that replaces the
traditional transmission with a single planetary gearset and a motor/generator.
This motor/generator uses the engine to either charge the battery or supply
additional power to the drive motor. A second, more powerful
motor/generator is permanently connected to the vehicle's final drive and
always turns with the wheels. The planetary gear splits engine power between
the first motor/generator and the drive motor to either charge the battery or
supply power to the wheels.
Plug-in hybrid electric vehicles (PHEV) - are hybrid electric vehicles with the
means to charge their battery packs from an outside source of electricity
(usually the electric grid). These vehicles have larger battery packs with more
energy storage and a greater capability to be discharged. They also use a
control system that allows the battery pack to be substantially depleted under
electric-only or blended mechanical/electric operation.
Electric vehicles (EV) - are vehicles with all-electric drive and with vehicle
systems powered by energy-optimized batteries charged primarily from grid
electricity.
3.3 What technology assumptions have the agencies used for this
proposal?
3.3.1 How are the technologies applied in the agencies' respective models?
While both NHTSA and EPA are basing their fuel economy and emission
modeling on the same baseline vehicle fleet and cost and effectiveness estimates for
control technologies, differences in the Volpe and OMEGA models require that this
common information be processed prior to use in different ways. With respect to the
vehicle fleet, the Volpe Model evaluates the addition of technology to individual vehicle
configurations or models, while the OMEGA model does so for vehicle platforms broken
down further by engine size. Thus, NHTSA used the vehicle sales estimates described
above directly, while EPA combined the sales of certain vehicle models. The Volpe
Model evaluates technologies individually. This, coupled with the modeling of
individual vehicle models, means that only the presence or absence of any particular
technology needs to be indicated, as described above. OMEGA applies technology in
combinations or packages. This, plus the grouping of individual vehicle models, requires
that the total effectiveness of the technology already applied in the baseline fleet must be
3-7
-------
Draft Joint Technical Support Document
calculated and must be reflected as a percentage of the various technology packages
available to be added to those vehicles.
With respect to the cost and effectiveness of technology, as mentioned above, the
Volpe Model applies technologies individually. It does this following certain specified
pathways for several categories of technologies (e.g., engine, transmission, accessories,
etc.). The Volpe Model applies technology incrementally, so the effectiveness of each
subsequent technology needs to be determined relative to the previous one. The same is
true for cost. In addition, because of overlap in the effectiveness of certain technologies,
herein referred to as the dis-synergy, any such overlap between the next technology on a
specified pathway with those which have already been potentially applied in other
pathways must be determined. For example, the incremental effectiveness of switching
from a six-speed automatic transmission to a dual clutch transmission will depend on the
level of engine technology already applied (e.g., intake cam phasing on a port-fuel
injected engine or a down-sized, turbocharged, direct injection engine).
EPA's OMEGA model applies technologies in packages and according to a fixed
sequence for any particular group of vehicles. This requires that the overall cost and
effectiveness of each package be determined first, considering any and all dis-synergies
which may exist. Then, the incremental cost and effectiveness of each subsequent
package is determined relative to the prior one.
Thus, while the same baseline vehicle fleet and cost and effectiveness estimates
for technologies are being used in both the Volpe and OMEGA models, the form of the
actual inputs to the model will appear to be different. For more information on EPA's and
NHTSA's unique approaches to modeling, please refer to each agency's respective
draft/proposed RIA.
In order to estimate both technology costs and fuel consumption/ CO2 reduction
estimates, it is necessary for each agency to describe the baseline vehicle characteristics
from which the estimates can be compared. This "baseline" is different from the usage in
Chapter 1 of this joint TSD. In Chapter 1, the baseline fleet is the projected fleet in 2016
before accounting for technologies needed to meet the 2011 MY CAFE standards and
before accounting for changes in fleet composition attributable to that rule (those later
steps accounted for independently by each agency in developing their separate reference
fleets). In the present context, it indicates the vehicle types and technologies that will be
used for comparison from a strictly cost and effectiveness point of view. These baselines
may be slightly different for the two agencies. For EPA, unless noted elsewhere, the
baseline vehicle is defined as a vehicle with a port-fuel injected, naturally aspirated
gasoline engine with two intake and two exhaust valves and fixed valve timing and lift.
The baseline transmission is a 4-speed automatic, and the vehicle has no hybrid systems.
For NHTSA, unless noted elsewhere, the baseline vehicle is the actual vehicle as it exists
in the baseline fleet (NHTSA models each unique vehicle separately). For details on how
the technology cost and effectiveness estimates presented below are used in each
agency's model, refer to each agency's respective draft/proposed RIA.
3-8
-------
Technologies Considered in the Agencies' Analysis
3.3.2 How did the agencies develop technology cost and effectiveness estimates for
the proposal?
3.3.2.1 Considerations that affect technology effectiveness
Regarding estimates for technology effectiveness, NHTSA and EPA also
reexamined the estimates from NHTSA's MY 2011 CAFE final rule and EPA's ANPRM
and Staff Technical Report, which largely mirrored NHTSA's NPRM estimates in the
2008 proposed rule. The agencies also reconsidered other sources such as the 2002 NAS
Report, the 2004 NESCCAF report, recent CAFE compliance data (comparing similar
vehicles with different technologies against each other in fuel economy testing, such as a
Honda Civic Hybrid versus a directly comparable Honda Civic conventional drive), and
confidential manufacturer estimates of technology effectiveness. Using the BOM
framework utilized in MY 2011 CAFE final rule, NHTSA and EPA engineers reviewed
effectiveness information from the multiple sources for each technology. Together, they
compared the multiple estimates and assessed their validity, taking care to ensure that
common BOM definitions and other vehicle attributes such as performance, refinement,
and drivability were taken into account. However, because the agencies' respective
models employ different numbers of vehicle subclasses and use different technology
decision trees to arrive at the standards, direct comparison of BOMs was somewhat more
complicated. To address this and to assure an apples-to-apple comparison, NHTSA and
EPA developed mapping techniques, devising technology packages and corresponding
incremental technology estimates. This approach helped compare incremental and
packaged estimates and derive results that are consistent and could be translated into the
respective models of the agencies. In general, most effectiveness estimates used in both
the MY 2011 CAFE final rule and the 2008 EPA staff report were determined to be
accurate and were carried forward without significant change into this rulemaking. When
NHTSA and EPA's estimates for effectiveness diverged slightly due differences in how
agencies apply technologies to vehicles in their respective models, the agencies will
report the ranges for the effectiveness values used in each model, as well as the reasons
the range is reasonable.
3.3.2.1.1 Technology synergies
The benefits of these technologies are generally but not always additive. That is,
adding cylinder deactivation to dual cam phasing does not necessarily result in a 10
percent improvement. This is true because some of the technologies address the same
shortcomings of the internal combustion engine, although in different ways. Where that
is the case, negative synergies are said to exist between two (or more) technologies. It is
also true that combining some technologies provides more benefit than a simple additive
effect. In those cases, positive synergies are said to exist. The analysis accounts for
these positive and negative synergies wherever appropriate so as to ensure no double or
-------
Draft Joint Technical Support Document
insufficient counting of effects. Synergies are discussed in more detail in each agency's
respective draft/proposed RIA.
3.3.2.2 Considerations that affect technology costs
As a general matter, the best way to derive technology cost estimates is to conduct
real- world tear down studies. These studies break down each technology into its
respective components, evaluate the costs of each component, and build up the costs of
the entire technology based on the contribution of each component. As such, tear down
studies require a significant amount of time and are very costly. EPA has begun
conducting tear down studies to assess the costs of 4-5 technologies under a contract with
FEV Inc., an independent engine and powertrain systems research, design and
development company. To date, only one technology (downsizing and turbo charging)
has been evaluated. Note that tear-down cost estimates from FEV for additional engine
and transmission technologies became available shortly before the release of this
proposal, but were not incorporated into the agencies' cost analysis. These preliminary
results are available for review and the agencies will consider this information for use in
the final rule analysis.10 A detailed report will be submitted to the docket on these
additional technologies during the public comment period for this proposal. In addition,
additional technologies are expected to be completed from this study with FEV after the
close of the comment period which the agencies will also docket. The agencies will
consider these studies and any comments received on them, as practicable and
appropriate, in developing technology cost assumptions for the final rule. As noted
above, the agencies relied on some of the FEV findings for estimating costs in this
rulemaking.
Regarding the other technologies, because tear down studies were not yet
available, the agencies decided to pursue the Bill of Materials (BOM) approach as
outlined in NHTSA's 2011 final rule and EPA's 2008 staff report. This approach was
recommended to NHTSA by Ricardo, an international engineering consulting firm
retained by NHTSA to aid in the analysis of public comments on its proposed standards
for MYs 2011-2015 because of its expertise in the area of fuel economy technologies. A
BOM approach is similar in concept to the approach used in tear down studies. The
difference is that under a BOM approach, the build up of cost estimates is conducted
based on a review of cost and effectiveness estimates for each component from available
literature. To the extent that the agencies departed from the 2011 CAFE final rule and
the 2008 staff report estimates, the agencies explained the reasons and provided
supporting analyses. For example, while NHTSA and Ricardo engineers had relied
considerably in the MY 2011 final rule on the 2008 Martec Report for costing contents of
some technologies, upon further joint review and for purposes of the MY 2012-2016
standards, the agencies decided that some of the costing information in that report was no
longer accurate due to downward trends in commodity prices since the publication of that
report. The agencies reviewed, revalidated or updated cost estimates for individual
components based on new information.
3-10
-------
Technologies Considered in the Agencies' Analysis
The subsequent sections describe the costs associated with the new vehicle
technologies. The costs represent the piece costs for an individual piece of hardware or
system, e.g., an intake cam phaser to provide variable valve timing. Costs are presented
in terms of their hardware incremental compliance cost. This means that they include all
potential costs associated with their application on vehicles, not just the cost of their
physical parts. The basis for the piece costs presented here is NHTSA's 2011 CAFE
FRM and EPA's 2008 Staff Technical Report.11'12 In those analyses, piece costs were
estimated based on a number of sources for cost related information. The objective was
to use those sources of information considered to be most credible for projecting the costs
of individual vehicle technologies. As a result of meetings between EPA and NHTSA
staff in the first half of 2009, where both piece costs and fuel consumption efficiencies
were discussed in detail, cost estimates have been adjusted slightly. Where estimates
differ between sources, engineering judgment was used to arrive at what is believed to be
the best cost estimate available today, and explained the basis for that exercise of
judgment. Thus while NHTSA and EPA found that much of the cost information used in
NHTSA's MY 2011 final rule and EPA's 2008 staff report was consistent to a great
extent, the agencies, in reconsidering information from many sources, revised several
component costs of several major technologies information (turbocharging downsizing,
mild and strong hybrids, diesels, SGDI, Valve Train Lift Technologies).
Note that EPA's 2008 staff report presented costs in terms of 2006 dollars. For
this analysis, costs are expressed in terms of 2007 dollars. Any costs presented here that
are based on costs expressed in 2006 dollars have been adjusted to reflect 2007 dollars
using the Gross Domestic Product price deflator (see Appendix 3.A).13'a The following
discussion summarizes the piece cost estimates and how these data sources were used to
arrive at the best estimate of piece costs. The following sections describe individual
technologies and their respective cost and effectiveness estimates. For purposes of this
joint TSD, these estimates were developed jointly by NHTSA and EPA. Due to
differences in the modeling and technology application methodologies employed by the
two agencies, the estimates used by each agency may not match exactly but will be
consistent given the different modeling tools used. Additional details are also provided in
each agency's respective proposed/draft RIAs.
3.3.2.2.1 Direct manufacturing costs
Building on NHTSA's estimates developed for the MY 2011 CAFE final rule
and EPA's Advanced Notice of Proposed Rulemaking, which relied on the 2008 Staff
Technical Report,14 the agencies took a fresh look at technology cost and effectiveness
values for purposes of the joint proposal under the National Program. For costs, the
agencies reconsidered both the direct or "piece" costs and indirect costs of individual
components of technologies. For the direct costs, the agencies followed a bill of
materials (BOM) approach as noted above. A bill of materials, in a general sense, is a list
of components that make up a systemin this case, an item of fuel economy-improving
1 The adjustment used to convert from 2006 dollars to 2007 dollars was 1.03.
3-11
-------
Draft Joint Technical Support Document
technology. In order to determine what a system costs, one of the first steps is to
determine its components and what they cost.
NHTSA and EPA estimated these components and their costs based on a number
of sources for cost-related information. The objective was to use those sources of
information considered to be most credible for projecting the costs of individual vehicle
technologies. For example, while NHTSA and Ricardo engineers had relied considerably
in the MY 2011 CAFE final rule on the 2008 Martec Report for costing contents of some
technologies, upon further joint review and for purposes of the MY 2012-2016 standards,
the agencies decided that some of the costing information in that report was no longer
accurate due to downward trends in commodity prices since the publication of that report.
The agencies reviewed, revalidated or updated cost estimates for individual components
based on new information. Thus though NHTSA and EPA found that much of the cost
information used in NHTSA's MY 2011 CAFE final rule and EPA's staff report was
consistent, the agencies, in reconsidering information from many sources, revised several
component costs of several major technologies (turbocharging downsizing, mild and
strong hybrids, diesels, SGDI, Valve Train Lift Technologies).15 These are discussed at
length below. For one technology (turbocharging/downsizing), the agencies relied, to the
extent possible, on the tear down data available and scaling methodologies used in EPA's
ongoing study with FEV. This study consists of complete system tear-down to evaluate
technologies down to the nuts and bolts to arrive at very detailed estimates of the costs
associated with manufacturing them.15 The confidential information provided by
manufacturers under their product plan submissions to the agencies or discussed in
meetings between the agencies and the manufacturers and suppliers served largely as a
check on publicly-available data.
Once costs were determined, they were adjusted to ensure that they were all
expressed in 2007 dollars using the GDP price deflator as described in Appendix 3.A.
Indirect costs were accounted for using a new approach developed by EPA for this
rulemaking and explained below rather than using the Retail Price Equivalent (RPE)
multiplier of 1.5 as was done in NHTSA's 2011 FRM and EPA's 2008 staff report.
NHTSA and EPA also considered how costs should be adjusted to reflect manufacturer
learning as discussed below. Lastly, costs were adjusted by modifying or scaling content
assumptions to account for differences across the range of vehicle sizes and functional
requirements, and adjusted the associated material cost impacts to account for the revised
content, although these adjustments were different for each agency due to the different
vehicle subclasses used in their respective models.
b The 2002 NAS Report, the 2004 study done by NESCCAF, the California Air Resources Board (CARD)
Initial Statement of Reasons in support of their carbon rulemaking, a 2006 study done by Energy and
Environmental Analysis (EEA) for the Department of Energy, a study done by Martec for the Alliance of
Automobile Manufacturers and the 2008 Martec Report which updated that study, and vehicle fuel
economy certification data, and confidential data submitted by manufacturers in response to the March
2009 request for product plans.
3-12
-------
Technologies Considered in the Agencies' Analysis
3.3.2.2.2 Cost markups to account for indirect costs
Indirect costs include production-related costs (research, development, and other
engineering), business-related costs (corporate salaries, pensions), and retail-sales-related
costs (dealer support, marketing). For this analysis, direct cost estimates were first
developed for each technology or system at the auto manufacturer level, i.e., the price
paid by the manufacturer to a Tier 1 component supplier. To these costs, an indirect cost
markup factor was then applied that varied by the best estimate of the particular
technology's complexity. This section describes the approach to determining the indirect
cost multipliers (ICM) used in this analysis and the specific multipliers used for each
piece of technology.
3.3.2.2.3 Concept behind and development of indirect cost multipliers
If all desirable data were available, when a new technology is implemented, the
costs of that technology would include the direct and indirect costs particular to that
technology. For instance, some changes may involve new tooling, while others may not;
some may affect the way the car is marketed, while others are of limited interest to
consumers. In a world of full information, the indirect costs of a new technology would
be calculated specifically for that technology. In practice, though, it is often difficult, if
not impossible, to identify the indirect costs specific to a new technology.
The automotive industry, EPA, and NHTSA have commonly used retail price
equivalent (RPE) multipliers to approximate the indirect costs associated with a new
technology. The RPE is a ratio of total revenues to direct manufacturing costs. Because,
by definition, total revenues = direct costs + indirect costs + profit, the RPE is the factor
that, when multiplied by direct manufacturing costs, recovers total revenue. This
multiplication is accurate only in the aggregate; it does not in reality apply to any specific
technology. The RPE is a way to estimate indirect costs on the assumption that indirect
costs are constant across all technologies and processes in a company. In the MY 2011
CAFE final rule NHTSA utilized a 1.5 RPE multiplier.
In fact, however, the indirect costs of new technologies vary, both with the
complexity of the technology and with the time frame. For instance, a hybrid-electric
engine is likely to involve greater research and development and marketing costs per
dollar of direct costs than low-rolling-resistance tires; the research and development costs
of any technology are likely to decrease over time. In recognition of this concern, EPA
contracted with RTI International to provide a current estimate of the RPE multiplier and
to examine whether the indirect costs of new technologies are likely to vary across
technologies. The report "Automobile Industry Retail Price Equivalent and Indirect Cost
Multipliers," by Alex Rogozhin, Michael Gallaher, and Walter McManus,16 calculates the
RPE multiplier as 1.46 in 2007. The report then develops indirect cost (1C) multipliers
that vary with the complexity of technology and the time frame. While any multiplier is
only an approximation of the true indirect costs of a new technology, the 1C multipliers in
this report move away from the assumption that the proportion of indirect costs is
3-13
-------
Draft Joint Technical Support Document
constant across all technologies and take into account some of the variation in these costs.
The multipliers developed in this report are presented in Table 3-1 .
The indirect cost multipliers used adjustment factors, developed by a team of EPA
engineers with expertise in the auto industry, which accounted for the differences in
complexity of the specific technologies under study. To examine the sensitivity of the
results to different technologies of the same complexity, and to provide more detailed
documentation of the development of the adjustment factors, EPA convened a second
panel,17 with NHTSA's input, to develop adjustment factors for three different
technologies. This latter process allowed for estimates of the variation in adjustment
factors, and thus in the variation of indirect cost multipliers. These results are also
presented in Table 3-1.
Table 3-1 Indirect Cost Multipliers
STUDY
RTI Report
EPA Memo: Average
Standard Deviation
Median
Max
Min
Multipliers Used in
this Analysis
TECHNOLOGY COMPLEXITY
Short Run
Low
1.05
1.16
0.14
1.12
1.43
1.00
1.11
Medium
1.20
1.29
0.15
1.26
1.53
1.02
1.25
High
1.45
1.64
0.21
1.66
2.15
1.37
1.45 1.64
Long Run
Low
1.02
1.12
0.14
1.06
1.42
1.00
1.07
Medium
1.05
1.20
0.13
1.20
1.45
1.01
1.13
High
1.26
1.39
0.15
1.40
1.69
1.12
1.26 1.39
The table shows minor differences in the multipliers for low- and medium-
complexity technologies (roughly 0.1), but larger differences in the high-complexity
technologies. The EPA and NHTSA engineers who reviewed the results believed that the
differences reflected actual differences in the technologies under study. In particular, for
low complexity, low-rolling-resistance tires (the application in the RTI Report) would
involve lower indirect costs than aerodynamic improvements (the application in the EPA
memo); and, for medium complexity, dual-clutch transmissions (the application in the
RTI Report) should have a smaller multiplier than engine downsizing done in conjunction
with turbocharging (the application in the EPA Memo). For these two cases, EPA and
NHTSA considered these technologies to span the range of technologies assigned to
those classes; the costs in this study, then, use the averages of the values of the two
reports, as shown in the last line of Table 3-1. For high complexity technologies, the
agencies felt the technologies assigned to these categorieshybrid-electric vehicles in
the RTI Report; plug-in hybrid electric vehicles in the EPA Memowere sufficiently
different that each deserved a different category. This is discussed in more detail in the
next section which highlights the multipliers used for each specific technology.
3-14
-------
Technologies Considered in the Agencies' Analysis
3.3.2.2.3.1 Application of specific indirect cost multipliers to each technology
As noted in the previous section, a different ICM was applied to each
technology's direct cost to arrive at its compliance cost. These different ICMs were
chosen based on the complexity of the technology in the opinions of staff engineers at
EPA and NHTSA, most of whom have several years of experience in the auto industry.
As shown in Table 3-1, ICMs were developed via two separate processes: that presented
in the RTI report; and that presented in the EPA Memo. While all of the ICMs generated
via these two processes were in general agreement, some differences did exist. In
determining how to deal with these differences, EPA and NHTSA agreed that, for the low
and medium complexity technologies, a simple average of the two values would be used.
However, for the high complexity technologies, it was decided that two separate high-
multipliers should be used. The lower multiplier, deemed high, would be applied to those
technologies of high complexity but with some level of use in the marketplace today.
Such technologies would be power-split and 2-mode hybrid electric vehicles. The higher
multiplier, deemed high+, would be applied to those technologies of high complexity but
with no, or essentially no, use in the current fleet. Such technologies would be plug-in
hybrids and full electric vehicles. Table 3-2 shows the complexity level for each
technology considered in this analysis.
Table 3-2 Complexity Levels of Technologies
LOW COMPLEXITY
Low friction lubes
Engine friction reduction
VVT-intake (OHC)
VVT-coupled (OHC & OHV)
VVT-dual (OHC)
Cylinder deactivation (OHC &
OHV)
VVLT-discrete (OHC)
GDI-stoich
Turbocharge (with no downsize)
Downsize (with no turbocharge)
Cam configuration changes (with
no downsize)
Aggressive shift logic
Early torque converter lockup
5-speed auto transmission
6-speed auto transmission
6-speed manual transmission
Improved accessories
Electric power steering
Low rolling resistance tires
Low drag brakes
Secondary axle disconnect
Improved aerodynamics
Weight reduction -3%
Weight reduction -5%
MEDIUM COMPLEXITY
VVLT-continuous (OHV)
VVLT-discrete (OHV)
GDI-lean burn
Turbocharge with downsize
Conversion to diesel
Dual clutch transmission
Continually variable trans
42 Volt conversion
Weight reduction - 20%
Integrated starter generator
(stop-start)
Combustion restart
Cooled EGR/EGR boost
HIGH COMPLEXTIY
VVLT-continuous
(OHC)
Camless valve actuation
Homogeneous charge
CI
Weight reduction - 30%
Integrated motor assist
hybrid
2-mode hybrid
Power-split hybrid
HIGH+
COMPLEXITY
Plug-in hybrid
Full electric vehicle
3-15
-------
Draft Joint Technical Support Document
Weight reduction - 10%
A/C Leakage Reduction
A/C Efficiency Improvement
The estimates of vehicle compliance costs cover the years of implementation of
the program - 2012 through 2016. In EPA's analysis, compliance costs have also been
estimated for the years following implementation to shed light on the long term - 2022
and later - cost impacts of the proposal. The year 2022 is used by EPA because the
short-term and long-term markup factors described above are applied in five year
increments with the 2012 through 2016 implementation span and the 2017 through 2021
span both representing the short-term.
3.3.2.2.4 Cost reduction through manufacturer learning
For some of the technologies considered in this analysis, manufacturer learning
effects would be expected to play a roll in the actual end costs. The "learning curve" or
"experience curve" describes the reduction in unit production costs as a function of
accumulated production volume. In theory, the cost behavior it describes applies to
cumulative production volume measured at the level of an individual manufacturer,
although it is often assumedas both agencies have done in past regulatory analysesto
apply at the industry-wide level particularly in industries that utilize many common
technologies and component supply sources.18 Both agencies believe there are indeed
many factors that cause costs to decrease over time. Research in the costs of
manufacturing has consistently shown that, as manufacturers gain experience in
production, they are able to apply innovations to simplify machining and assembly
operations, use lower cost materials, and reduce the number or complexity of component
parts. All of these factors allow manufacturers to lower the per-unit cost of production
(i.e., the manufacturing learning curve).19
NHTSA has a detailed description of the learning effect in the 2011 CAFE final
rule, and EPA has a detailed description of the learning effect in the 2008 staff report and
in several past rules.20 The description here is focused on how the learning effect has
been used in the analyses support each agency's rule.
Most studies of the effect of experience or learning on production costs appear to
assume that cost reductions begin only after some initial volume threshold has been
reached, but not all of these studies specify this threshold volume. The rate at which
costs decline beyond the initial threshold is usually expressed as the percent reduction in
average unit cost that results from each successive doubling of cumulative production
volume, sometimes referred to as the learning rate. Many estimates of experience curves
do not specify a cumulative production volume beyond which cost reductions would no
longer occur, instead depending on the asymptotic behavior of the effect for learning
rates below 100 percent to establish a floor on costs. Table 3-3 summarizes estimates of
learning rates derived from studies of production costs for various products.
3-16
-------
Technologies Considered in the Agencies' Analysis
Table 3-3 Estimated Learning Rates and Associated Volumes for Various Products
PRODUCT
Photovoltaic cells
Wind turbines
Gas turbines
Semiconductors
Automobile assembly
Truck manufacturing
Battery-electric LDV
Fuel cell hybrid LDV
Fuel cell LDV powertrain
COSTS
AFFECTED
Total costs
Total costs
Total costs
Total costs
Assembly labor
Total costs
Total costs
Total costs
Total costs
THRESHOLD
VOLUME
Not reported
100MW
100MW
Not reported
Not reported
Not reported
10,000 units
10,000 units
10,000 units
LEARNING
RATE
20%
20%
10%
13-24%
16%
10%
10%
16%
19%
In past rulemaking analyses, as noted above, both agencies have used a learning
curve factor of 20 percent for each doubling of production volume. NHTSA has used
this approach in analyses supporting recent CAFE rules. In its analysis, EPA has
simplified the approach by using an "every two years" based learning progression rather
than a pure production volume progression (i.e., after two years of production it was
assumed that production volumes would have doubled and, therefore, costs would be
reduced by 20 percent). This approach has served EPA well, especially considering that
those rulemaking analyses have reflected programs in which every new engine or vehicle
beginning in year one of implementation would be equipped with the newly required
piece of technology.
In the 2011 CAFE final rule, NHTSA considered not only this volume-based
learning as described above, but also "time-based" learning. The time-based learning,
estimated by NHTSA at three percent per year, occurs in years following the volume-
based learning steps and represents the smaller scale learning that occurs as
manufacturers continue to innovate. The time-based learning is, in effect, represented by
the flattened out, asymptotic portion of the learning curve.
For this analysis, each agency has employed both volume-based and time-based
learning effects. In NHTSA's analysis, volume-based learning is estimated to result in 20
percent lower costs as production volumes of the given technology reach 300,000 units
and again when they reach 600,000 units within the entire industry. The count of
production units begins in the 2012 model year and/or with the first units produced,
whichever is earlier. In EPA's analysis, as noted above, volume-based learning is
estimated to result in 20 percent lower costs after two full years of implementation (i.e.,
the 2014 MY costs are 20 percent lower than the 2012 and 2013 model year costs). In
each agency's analysis, time-based learning is estimated to result in 3 percent lower costs
in each year following first introduction of a given technology. Once two volume-based
learning steps have occurred, or once the five year time-based learning period is
completed, learning is assumed to be complete, at least for purposes of the analyses done
to support each agency's rule.
3-17
-------
Draft Joint Technical Support Document
Both agencies considered learning impacts on most but not all of the technologies
expected to be used because some of the expected technologies are already used rather
widely in the industry and, presumably, learning impacts have already occurred.
Volume-based learning was considered for only a handful of technologies that are
considered to be new or emerging technologies such as the hybrids and electric vehicles.
Most technologies have been considered to be more established given their current use in
the fleet and, hence, the lower time-based learning has been applied. The learning effects
applied to each technology are summarized in Table 3-4. Volume-based learning has
been applied to all of the hybrid and electric vehicle technologies with the exception of
the power-split hybrid. For each analysis, the power-split hybrid (e.g., the Toyota Prius)
was considered to have been in the market for enough years that the volume-based
learning effects have already passed and all that remains are the smaller, time-based
learning effects (this is consistent with NHTSA's 2011 CAFE FRM).
3-18
-------
Technologies Considered in the Agencies' Analysis
Table 3-4 Learning Effects Applied to Each Technology
TECHNOLOGY
Low friction lubricants
Engine friction reduction
VVT - intake cam phasing
VVT - coupled cam phasing
VVT - dual cam phasing
Cylinder deactivation
Discrete VVLT
Continuous VVLT
GDI - stoichiometric
GDI - lean burn
Turbocharge (with no downsize)
Downsize (with no turbocharge)
Turbocharge with downsize
Cam configuration changes (with no downsize)
Homogeneous charge CI
Conversion to diesel
Aggressive shift logic
Early torque converter lockup
5-speed automatic
6-speed automatic
6-speed DCT - dry clutch
6-speed DCT - wet clutch
6-speed manual
CVT
Stop-Start
IMA/ISA/BSG (includes engine downsize)
2-Mode hybrid electric vehicle
Power-split hybrid electric vehicle
Plug-in hybrid electric vehicle
Plug-in hybrid electric vehicle
Full electric vehicle
Improved high efficiency alternator & electrification of accessories
Upgrade to 42 volt electrical system
Electric power steering (12 or 42 volt)
Aero drag reduction (20% on cars, 10% on trucks)
Low rolling resistance tires
Low drag brakes (ladder frame only)
Secondary axle disconnect (unibody only)
Front axle disconnect (ladder frame only)
A/C Leakage Reduction
A/C Efficiency Improvements
LEARNING EFFECT
None
None
Time
Time
Time
Time
Time
Time
Time
Time
Time
Time
Time
Time
Time
Time
Time
Time
Time
Time
Time
Time
Time
Time
Volume
Volume
Volume
Time
Volume
Volume
Volume
Time
Time
Time
Time
None
None
None
Time
Time
Time
Table 3-20 through Table 3-25 show Volpe model costs and effectiveness values
for each particular technology described throughout this chapter. The costs shown are
applicable for the 2012 model year. The Volpe Model handles learning effects within the
model itself so that individual technology costs in the 2016 model year would be lower
than those shown in the tables.
Table 3-26 through Table 3-35 show OMEGA model costs and effectiveness
values for each particular technology described throughout this chapter. The costs shown
are applicable for the 2016 model year and, therefore, represent fully learned costs in the
context of EPA's analysis. For technologies added in years prior to 2016, EPA has
3-19
-------
Draft Joint Technical Support Document
backed out the learning effects relative to the costs shown in the tables. For example,
Table 3-28 shows the small car stop-start vehicle cost as $351 in 2016. In the 2012
model year, this cost would be higher since the volume-based learning reflected in the
2016 cost would not have occurred yet. Backing out two volume-based learning steps
(i.e., dividing $351 by 80% twice) would result in a 2012 cost estimate of $548.
The different handling of learning effects across the two analyses (that done by
EPA and that by NHTSA) would have little impact on the overall results of the analyses.
NHTSA handles learning effects in the CAFE model. EPA handles learning effects
outside of the OMEGA model since OMEGA is run for only one model year - 2016.
Because the CAFE model adds technology on a year-by-year basis, it applies volume-
based learning effects once the applicable production volume thresholds have been
exceeded. In contrast, EPA would not apply volume-based learning until year 3
(assumed to be the point at which initial production volumes have doubled). Most
technologies in both analyses are subject to time-based learning which is applied on an
annual basis in the same way by both agencies (with the exception that NHTSA applies
time-based learning in the CAFE model while EPA applies it outside the OMEGA
model). Importantly, by the 2016 model year, both agencies will have applied two
volume-based learning steps to the appropriate technologies (hybrid and electric vehicles
only) and four time-based learning steps to the appropriate technologies (most other
technologies) to arrive at the same technology costs.
3.4 Specific technologies considered and estimates of costs and
effectiveness
It is important to note that the cost and effectiveness values presented by NHTSA
and EPA in this Draft TSD are used when creating input files for the agencies' respective
models but are not literally the inputs themselves. The Docket to this proposed
rulemaking contains the specific input files EPA and NHTSA used when modeling the
technology cost.
3.4.1 What data sources did the agencies evaluate?
Refer to section 3.1 for a discussion of the specific data sources and reference
material used for the analysis of this rulemaking.
3.4.2 Individual technology descriptions and cost/effectiveness estimates
The sections that follow describe the technologies applied by both EPA and
NHTSA. It is important to note that there are distinct differences between the agencies in
how cost and effectiveness estimations are reported and applied.
3-20
-------
Technologies Considered in the Agencies' Analysis
EPA reports the total cost and effectiveness associated with a respective
technology. The technologies are then applied to a vehicle type as a package. To
calculate a cost for a vehicle package, the individual technologies are summed. To
estimate the overall effectiveness, the lumped parameter method is applied.
NHTSA applies cost and effectiveness through a series of decision trees. The
progressive nature of this methodology results in conditions for which a portion of a
technology has already been added. In this case, NHTSA would add an incremental
value which is the difference between the total cost of a technology and the portion
already existing. An example of this is cylinder deactivation. At the point in the decision
tree that it is applied, valvetrain technologies which could enable this feature have
already been added. NHTSA will add the delta in cost, which in this example accounts
for the NVH attributes.
3.4.2.1 Engine Technologies
NHTSA and EPA have reviewed the engine technology estimates used in
NHTSA's MY 2011 CAFE final rule and EPA's 2008 staff report. In doing so NHTSA
and EPA reconsidered all available sources and updated the estimates as appropriate. The
section below describes each of the engine technologies considered for this rulemaking.
3.4.2.1.1 Low Friction Lubricants (LUB)
One of the most basic methods of reducing fuel consumption in gasoline engines
is the use of lower viscosity engine lubricants. More advanced multi-viscosity engine
oils are available today with improved performance in a wider temperature band and with
better lubricating properties. This can be accomplished by changes to the oil base stock
(e.g., switching engine lubricants from a Group I base oils to lower-friction, lower
viscosity Group III synthetic) and through changes to lubricant additive packages (e.g.,
friction modifiers and viscosity improvers). The use of 5W-30 motor oil is now
widespread and auto manufacturers are introducing the use of even lower viscosity oils,
such as 5W-20 and OW-20, to improve cold-flow properties and reduce cold start friction.
However, in some cases, changes to the crankshaft, rod and main bearings and changes to
the mechanical tolerances of engine components may be required. In all cases, durability
testing would be required to ensure that durability is not compromised. The shift to lower
viscosity and lower friction lubricants will also improve the effectiveness of valvetrain
technologies such as cylinder deactivation, which rely on a minimum oil temperature
(viscosity) for operation.
Several manufacturers have previously commented confidentially, that low
friction lubricants could have an effectiveness value between 0 to 1 percent. For
purposes of this NPRM, NHTSA and EPA are using effectiveness estimates within this
range. Therefore 0.5 percent was used in both models.
3-21
-------
Draft Joint Technical Support Document
The 2002 NAS study estimated the low friction lubricant RPE at $8 to $11 using a
1.4 markup factor. The NESCCAF study showed an RPE of $5 to $15 with a 1.4
markup. The EEA report to DOE showed manufacturer costs of $10 to $20 with no
markup. Confidential Business Information (CBI) data estimates an average incremental
cost of $3 for the use of low friction lubricants. EPA's 2008 Staff Report also confirms
this $3 cost (2006$). Both NHTSA and EPA believe that manufacturer's estimates are
the most accurate, and thus continue to believe that the $3 cost estimate is appropriate
and independent of vehicle class since the engineering work required should apply to any
engine size. Applying an indirect cost multiplier (ICM) of 1.11, for a low complexity
technology, results in a compliance cost of $3 per vehicle (2007$) for a MY 2012 through
MY 2016 vehicle (learning effects are not applied to low friction lubes).0
3.4.2.1.2 Engine Friction Reduction (EFR)
In addition to low friction lubricants, manufacturers can also reduce friction and
improve fuel consumption by improving the design of engine components and
subsystems. Approximately 10 percent of the energy consumed by a vehicle is lost to
friction, and just over half is due to frictional losses within the engine.21 Examples
include improvements in low-tension piston rings, piston skirt design, roller cam
followers, improved crankshaft design and bearings, material coatings, material
substitution, more optimal thermal management, and piston and cylinder surface
treatments. Additionally, as computer-aided modeling software continues to improve,
more opportunities for evolutionary friction reductions may become available.
All reciprocating and rotating components in the engine are potential candidates
for friction reduction, and minute improvements in several components can add up to a
measurable fuel economy improvement. The 2002 NAS, NESCCAF and EEA reports as
well as confidential manufacturer data suggested a range of effectiveness for engine
friction reduction to be between 1 to 3 percent. NHTSA and EPA continue to believe
that this range is accurate. Because of the incremental nature of the CAFE model,
NHTSA needed to continue to use the narrower range of 1-2, which was also used in the
MY 2011 CAFE final rule.
In the MY 2011 CAFE final rule, NHTSA estimated a range from $13 to $49
using a 1.5 RPE on a per cylinder basis, or $9 to $33 without RPE (2007$). In the 2008
NPRM engine friction reduction was estimated to cost up to $14 without RPE on a per
cylinder basis (2006$). EPA's 2008 staff report estimated this at $11 using a 1.5 RPE on
a per cylinder basis, or $7 without RPE (2006$). After review, NHTSA and EPA believe
that the cost estimate is closer to the lower end of the MY 2011 CAFE final rule range
and thus for this rulemaking is $13 per cylinder compliance cost (2007$), including the
low complexity ICM markup value of 1.11, for a MY 2012 through MY 2016 vehicle
c Note that the costs developed for low friction lubes for this analysis reflect the costs associated with any
engine changes that would be required as well as any durability testing that may be required.
3-22
-------
Technologies Considered in the Agencies' Analysis
(learning effects are not applied to engine friction reduction). This cost is multiplied by
the number of engine cylinders.
3.4.2.1.3 Variable Valve Timing (WT)
Variable valve timing (VVT) classifies a family of valve-train designs that alter
the timing of the intake valve, exhaust valve, or both, primarily to reduce pumping losses,
increase specific power, and control the level of residual gases in the cylinder. VVT
reduces pumping losses when the engine is lightly loaded by controlling valve timing
closer to an optimum needed to sustain horsepower and torque. VVT can also improve
volumetric efficiency at higher engine speeds and loads. Additionally, VVT can be used
to alter (and optimize) the effective compression ratio where it is advantageous for
certain engine operating modes (e.g., in the Atkinson Cycle).
VVT has now become a widely adopted technology: in MY 2007, over half of all
new cars and light trucks had engines with some method of variable valve timing.22
Manufacturers are currently using many different types of variable valve timing, which
have a variety of different names and methods. The three major types of VVT are listed
below.
Each of the three implementations of VVT uses a cam phaser to adjust the
camshaft angular position relative to the crankshaft position, referred to as "camshaft
phasing." The phase adjustment results in changes to the pumping work required by the
engine to accomplish the gas exchange process. The majority of current cam phaser
applications use hydraulically-actuated units, powered by engine oil pressure and
managed by a solenoid that controls the oil pressure supplied to the phaser.
3.4.2.1.3.1 Intake Cam Phasing (ICP)
Valvetrains with ICP, which is the simplest of the cam phasing technologies, can
modify the timing of the inlet valves by phasing the intake camshaft while the exhaust
valve timing remains fixed. This requires the addition of a cam phaser on each bank of
intake valves on the engine. An in-line 4-cylinder engine has one bank of intake valves,
while V-configured engines have two banks of intake valves.
NHTSA's MY 2011 CAFE final rule and EPA 2008 Staff Report estimated an
effectiveness of 1 to 2 percent for ICP, which was supported by the NESCCAF report and
a majority of confidential manufacturer comments. The agencies have found no
additional sources to suggest strongly that this estimate is inaccurate, and so have
employed it for this NPRM.
As for costs, NHTSA's MY 2011 CAFE final rule estimated a $61 RPE ($41 non-
RPE) cost per cam phaser, based on the 2008 Martec Report and confidential
manufacturer data (2007$). In the 2008 staff report, EPA estimated this at $59 using a
1.5 RPE or $40 without RPE (2006$). NHTSA and EPA believe that this estimate
remains accurate. Using the new indirect cost multiplier of 1.11, for a low complexity
3-23
-------
Draft Joint Technical Support Document
technology, the compliance cost per cam phaser would be $45 per bank , yielding a $45
cost for and in-line engine configurations and $90 for V-engine configurations for a MY
2012 vehicle (2007$). Time-based learning is applied to TCP so the MY 2016 cost would
be $40 for in in-line engine and $80 for the V-engine (2007$).
3.4.2.1.3.2 Coupled Cam Phasing (CCPS and CCPO)
Valvetrains with coupled (or coordinated) cam phasing can modify the timing of
both the inlet valves and the exhaust valves an equal amount by phasing the camshaft of a
single overhead cam (SOHC) engine or an overhead valve (OHV) engine.d For overhead
cam engines, this requires the addition of a cam phaser on each bank of the engine. Thus,
an in-line 4-cylinder engine has one cam phaser, while SOHC V-engines have two cam
phasers. For overhead valve (OHV) engines, which have only one camshaft to actuate
both inlet and exhaust valves, CCP is the only VVT implementation option available and
requires only one cam phaser.6
Based on NHTSA's MY 2011 CAFE final rule, previously-received confidential
manufacturer data, and the NESCCAF report, NHTSA and EPA estimated the
effectiveness of CCP to be between 1 to 4 percent. NHTSA and EPA reviewed this
estimate for purposes of the NPRM, and continue to find it accurate.
The same cam phaser has been assumed for ICP and CCP applications, thus CCPs
cost per cam phaser is identical to ICPs. This results in a cost of $45 for in-line SOHC
and OHV engines and $90 for SOHC V-engine configurations for a MY 2012 vehicle
(2007$). With time-based learning applied, these costs for a 2016 MY vehicle would be
$40 and $80, respectively (2007$).
3.4.2.1.3.3 Dual Cam Phasing (DCP)
The most flexible VVT design is dual (independent) cam phasing, where the
intake and exhaust valve opening and closing events are controlled independently. This
option allows the option of controlling valve overlap, which can be used as an internal
EGR strategy. At low engine loads, DCP creates a reduction in pumping losses, resulting
in improved fuel consumption/reduced COi emissions. Increased internal EGR also
results in lower engine-out NOx emissions. The amount by which fuel consumption is
improved and CO2 emissions are reduced depends on the residual tolerance of the
combustion system. Additional improvements are observed at idle, where low valve
overlap could result in improved combustion stability, potentially reducing idle fuel
consumption.
d Although CCP appears only in the SOHC and OHV branches of the decision tree, it is noted that a single
phaser with a secondary chain drive would allow CCP to be applied to DOHC engines. Since this would
potentially be adopted on a limited number of DOHC engines NHTSA did not include it in that branch of
the decision tree.
e It is also noted that coaxial camshaft developments would allow other VVT options to be applied to OHV
engines. However, since they would potentially be adopted on a limited number of OHV engines NHTSA
did not include them in the decision tree.
3-24
-------
Technologies Considered in the Agencies' Analysis
NHTSA MY 2011 CAFE final rale estimated the effectiveness of DCP to be
between 3 to 5 percent relative to a base engine or 2 to 3 relative to an engine with ICP.
NHTSA and EPA believe that this estimate remains applicable for the NPRM.
As above, the same cam phaser has been assumed for ICP and DCP applications.
Thus, DCP's cost per cam phaser is identical to TCP's. DCP requires two cam phasers
per cylinder bank, one to control the intake valves and one to control the exhaust valves.
This results in a cost of $90, relative to an engine without ICP, or $45 relative to an
engine with ICP, minus $6 for the removal of the EGR valve, ultimately yielding costs of
$84 and $39 respectively for in-line DOHC configurations (all in 2007$). For V-
configuration engines, the cost is $180 relative to an engine without ICP, or $90 relative
to an engine with ICP, minus $6 for the removal of the EGR valve, ultimately yielding
costs of $174 and $84, respectively (all in 2007$). These costs are appropriate for a MY
2012 vehicle. With time-based learning applied, these costs for a 2016 MY vehicle
would be $157 and $73, respectively (2007$).
3.4.2.1.4 Variable Valve Lift (WL)
Controlling the lift of the valves provides a potential for further efficiency
improvements. By optimizing the valve-lift profile for specific engine operating regions,
the pumping losses can be reduced by reducing the amount of throttling required to
produce the desired engine power output. By moving the throttling losses further
downstream of the throttle valve, the heat transfer losses that occur from the throttling
process are directed into the fresh charge-air mixture just prior to compression, delaying
the onset of knock-limited combustion processes. Variable valve lift control can also be
used to induce in-cylinder mixture motion, which improves fuel-air mixing and can result
in improved thermodynamic efficiency. Variable valve lift control can also potentially
reduce overall valvetrain friction. At the same time, such systems may also incur
increased parasitic losses associated with their actuation mechanisms. A number of
manufacturers have already implemented VVLT into their fleets (Toyota, Honda, and
BMW), but overall this technology is still available for most of the fleet. There are two
major classifications of variable valve lift, described below:
1. Discrete Variable Valve Lift (DWLS, DWLD, DWLO)
DVVL systems allow the selection between two or three discrete cam profiles by
means of a hydraulically-actuated mechanical system. By optimizing the cam profile for
specific engine operating regions, the pumping losses can be reduced by reducing the
amount of throttling required to produce the desired engine power output. This increases
the efficiency of the engine. These cam profiles consist of a low and a high-lift lobe, and
may include an inert or blank lobe to incorporate cylinder deactivation (in the case of a 3-
step DVVL system). DVVL is normally applied together with VVT control. DVVL is
also known as Cam Profile Switching (CPS). DVVL is a mature technology with low
technical risk.
NHTSA's MY 2011 CAFE final rale, previously-received confidential
manufacturer data, and the NESCCAF report, all estimate the effectiveness of DVVL to
3-25
-------
Draft Joint Technical Support Document
be between 1 to 4 percent above that realized by VVT systems. NHTSA and EPA
believe this estimate continues to be applicable for the NPRM.
In the 2011 CAFE final rule, NHTSA estimated an RPE (1.5 markup factor) of
$201 for an inline 4-cylinder engine, $306 for a V6 engine and $396 for a V8 engine or
without the RPE markup $134, $204, $264, respectively (all in 2007$). In the 2008 staff
report, EPA estimated this at $169, $246, and $322, respectively, using a 1.5 RPE factor
(2006$) or $113, $164 and $215, respectively without markup (all in 2006$). For this
analysis, the agencies have estimated the compliance costs, including $25 for controls
and using a 1.11ICM for a low complexity technology, at $141 for an inline 4-cylinder
engine, $205 for a V6 engine and $293 for a V8 for a 2012 MY vehicle (2007$). With
time-based learning applied, these costs would be $125, $181, and $259, respectively for
a 2016 MY vehicle (2007$).
2. Continuously Variable Valve Lift (CVVL)
In CVVL systems, valve lift is varied by means of a mechanical linkage, driven
by an actuator controlled by the engine control unit. The valve opening and phasing vary
as the lift is changed and the relation depends on the geometry of the mechanical system.
BMW has considerable production experience with CVVL systems and has sold port-
injected "Valvetronic" engines since 2001. CVVL allows the airflow into the engine to
be regulated by means of intake valve opening reduction, which improves engine
efficiency by reducing pumping losses from throttling the intake system further upstream
as with a conventionally throttled engine.
Variable valve lift gives a further reduction in pumping losses compared to that
which can be obtained with cam phase control only, with CVVL providing greater
effectiveness than DVVL, since it can be fully optimized for all engine speeds and loads,
and is not limited to a two or three step compromise. There may also be a small
reduction in valvetrain friction when operating at low valve lift, resulting in improved
low load fuel consumption for cam phase control with variable valve lift as compared to
cam phase control only. Most of the fuel economy effectiveness is achieved with
variable valve lift on the intake valves only. CVVL is only applicable to double overhead
cam (DOHC) engines.
NHTSA's MY 2011 CAFE final rule estimated the effectiveness for CVVL at 1.5
to 3.5 percent over an engine with DCP, but also recognize that it could go up as high as
5% above and beyond DCP to account for the implementation of more complex CVVL
systems such as BMW's "valvetronic" engines. This coincides with EPA Staff report
estimates of the contribution of CVVL, which were based on the NESCCAF report, in
which CVVL could improve effectiveness by 4 percent (minivans) and up to 6 percent
(large cars) over dual cam phasing. Thus, the effectiveness range for CVVL in this joint
TSD ranges from 1.5 to 6 percent depending on the complexity level of the application.
However, due to the complexity and cost of this technology, the GHG model did not
consider it and the CAFE models projected very limited application of less complex
version of this technology respectively (i.e., 2 out of 1100 vehicles in the CAFE model).
3-26
-------
Technologies Considered in the Agencies' Analysis
The most recent submission of manufacturers' product plans confirmed that this
technology will not be applied by most manufacturers.
In the MY 2011 CAFE final rule, NHTSA estimated and RPE (1.5 factor) of
continuously variable valve lift to be $306 for an inline 4-cylinder engine, $432 for a V6
engine and $582 for a V8 engine or without RPE $204, $287, $388, respectively. In the
2008 staff report, EPA estimated this at $254, $466, and $508, respectively, using a 1.5
RPE factor (2006$) or $169, $311 and $339, respectively without markup (all in 2006$)
For this analysis, the agencies have estimated the compliance costs, using a 1.45 ICM for
a high complexity technology, at $277 for an inline 4-cylinder engine, $509 for a V6
engine and $554 for a V8 engine for a MY 2012 vehicle (2007$). With time-based
learning applied, these costs would become $245, $449, and $489, respectively for a MY
2016 vehicle (2007$).
3.4.2.1.5 Cylinder Deactivation (DEACS, DEACD, DEACO)
In conventional spark-ignited engines throttling the airflow controls engine torque
output. At partial loads, efficiency can be improved by using cylinder deactivation
instead of throttling. Cylinder deactivation (DEAC) can improve engine efficiency by
disabling or deactivating (usually) half of the cylinders when the load is less than half of
the engine's total torque capability - the valves are kept closed, and no fuel is injected -
as a result, the trapped air within the deactivated cylinders is simply compressed and
expanded as an air spring, with reduced friction and heat losses. The active cylinders
combust at almost double the load required if all of the cylinders were operating.
Pumping losses are significantly reduced as long as the engine is operated in this "part-
cylinder" mode.
Cylinder deactivation control strategy relies on setting maximum manifold
absolute pressures or predicted torque within which it can deactivate the cylinders. Noise
and vibration issues reduce the operating range to which cylinder deactivation is allowed,
although manufacturers are exploring vehicle changes that enable increasing the amount
of time that cylinder deactivation might be suitable. Some manufacturers may choose to
adopt active engine mounts and/or active noise cancellations systems to address NVH
concerns and to allow a greater operating range of activation. Manufacturers have stated
that use of DEAC on 4 cylinder engines would cause unacceptable NVH; therefore
cylinder deactivation has not been applied to 4-cylinder engines.
Cylinder deactivation has seen a recent resurgence thanks to better valvetrain
designs and engine controls. General Motors and Chrysler Group have incorporated
cylinder deactivation across a substantial portion of their V8-powered lineups. Honda
(Odyssey, Pilot) offers V6 models with cylinder deactivation.
Effectiveness improvements scale roughly with engine displacement-to-vehicle
weight ratio: the higher displacement-to-weight vehicles, operating at lower relative loads
for normal driving, have the potential to operate in part-cylinder mode more frequently.
3-27
-------
Draft Joint Technical Support Document
NHTSA and EPA reviewed the MY 2011 CAFE estimate and confirmed a range
of 0.5 to 6 percent depending on the valvetrain configuration and other existing
technologies. The OMEGA model, which is based on packages, applies 6% reduction in
CO2 emissions in all applications. The CAFE model, due to its incremental nature, uses
a range depending on the engine valvetrain configuration. For example, for DOHC
engines which are already equipped with DCP and DVVLD, there is little benefit that can
be achieved from adding cylinder deactivation since the pumping work has already been
minimized and internal Exhaust Gas Recirculation (EGR) rates are maximized, so the
effectiveness is only up to 0.5 percent for DEACD. For Single Overhead Camshaft
engines which have CCP and DVVLS applied, effectiveness ranged from 2.5 to 3 percent
for DEACS. For OHV engines, without VVT or VVL technologies, the effectiveness for
DEACO ranged from 3.9 to 5.5 percent.
NHTSA and EPA considered a range of $170 to $190 (or $28 to $190 depending
on whether an engine already has lost motion devices, oil control valves and camshaft
position sensors). This is a departure from NHTSA's 2011 final rule, which used a range
of $306 to $400. That range was primarily based on 2008 Martec Report and applied a
higher RPE value. In reviewing these assumptions, NHTSA and EPA amended the MY
2011 CAFE estimates and adjusted the estimates to include the new ICM low complexity
markup of 1.11. The EPA staff report estimated this ranging from $203 to $229 (RPE
using a 1.5 factor) or $135 to $153 without RPE (2006$). NHTSA's NPRM showed
estimates of a $170 for a 6-cylinder engine and $190 for an 8-cylinder enginewhen
adjusted for 2007 dollars and using the new ICM multipliers for engines that do not
have lost motion devices. These numbers were within the ranges described by the 2002
NAS and NESCCAF reports. The $170 and $190 costs are appropriate of a MY 2012
vehicle (2007$). With time-based learning applied, these would become $150 and $169,
respectively, for a MY 2016 vehicle (2007$).
If lost motion devices are on the engine, the cost of DEAC as applied to SOHC
and DOHC engines could be as low as $28.f This $288 accounts for the potential
additional application of active engine mounts on SOHC and DOHC engines and can
only be applied on 50 percent of the vehicles. Further, this SOHC and DOHC engine
estimate is relevant to the CAFE model only because the OMEGA model does not apply
technologies in the same incremental fashion as the CAFE model.
3.4.2.1.6 Conversion to Double Overhead Camshaft Engine with Dual Cam Phasing
(CDOHC)
f The $28 is an adjustment from the $75 estimate used in the MY 2011 final rule to account for the new
ICM markup factor and the fact that it could only be applied on up to 50 percent of the vehicles.
3-28
-------
Technologies Considered in the Agencies' Analysis
Double overhead camshaft engines achieve increased airflow at high engine
speeds, improve volumetric efficiency and reductions of the valvetrain's moving mass.
Such engines typically develop higher power at high engine speeds. Manufacturers may
choose to replace OHV engines with DOHC engine designs with dual cam phasing
(DCP). EPA and NHTSA estimate the effectiveness to be between 1 to 2.5 percent.
As for costs, NHTSA's MY 2011 CAFE final rule assumed that CDOHC would
have an RPE of $746 ($497 non-RPE) for a V8 engine, $590 ($393 non-RPE) for a V6
engine and $373 ($249 non-RPE) for inline 4-cylinder engine (2007$). In the 2008 staff
report, EPA did not estimate a cost for this specific technology. For purposes of this
NPRM, NHTSA revised the costs only by identifying this technology as a low
complexity technology and applying an indirect cost multiplier of 1.11 resulting in a
compliance cost of $552 for V8 engine, $436 for a V6 and $276 for an inline 4-cylinder
engine. The above costs are for a MY 2012 vehicle and with the application of time-
based learning will be slightly lower for MY 2016 vehicle
3.4.2.1.7 Stoichiometric Gasoline Direct Injection (SGDI)
Gasoline direct injection (GDI), or Spark Ignition Direct injection (SIDI), engines
inject fuel at high pressure directly into the combustion chamber (rather than the intake
port in port fuel injection). GDI requires changes to the injector design, an additional
high pressure fuel pump, new fuel rails to handle the higher fuel pressures and changes to
the cylinder head and piston crown design. Direct injection of the fuel into the cylinder
improves cooling of the air/fuel charge within the cylinder, which allows for higher
compression ratios and increased thermodynamic efficiency without the onset of
combustion knock. Recent injector design advances, improved electronic engine
management systems and the introduction of multiple injection events per cylinder firing
cycle promote better mixing of the air and fuel, enhance combustion rates, increase
residual exhaust gas tolerance and improve cold start emissions. GDI engines achieve
higher power density and match well with other technologies, such as boosting and
variable valvetrain designs.
Several manufacturers have recently introduced vehicles with GDI engines,
including VW/Audi, BMW, Toyota (Lexus IS 350) and General Motors (Chevrolet
Impala and Cadillac CTS 3.6L). BMW, GM, Ford and VW/Audi have announced their
plans to increase dramatically the number of GDI engines in their portfolios.
NHTSA's MY 2011 CAFE final rule estimated the effectiveness of SGDI to be
between 2 and 3 percent. NHTSA and EPA reviewed estimates from the Auto Alliance
of American Manufacturers, which projects 3 percent gains in fuel efficiency and a 7
percent improvement in torque. The torque increase provides the opportunity to
downsize the engine allowing an increase in efficiency of up to a 5.8 percent. NHTSA
and EPA also reviewed other published literature, reporting 3 percent effectiveness for
SGDI.23 Another source reports a 5 percent improvement on the NEDC drive cycle.24
3-29
-------
Draft Joint Technical Support Document
Confidential manufacturer data reported an efficiency effectiveness range of 1 to 2
percent. NHTSA and EPA determined that the range of 2 to 3 percent continues to be
appropriate. Combined with other technologies (i.e., boosting, downsizing, and in some
cases, cooled EGR), SGDI can achieve greater reductions in fuel consumption and COi
emissions compared to engines of similar power output.
The NHTSA and EPA cost estimates for SGDI take into account the changes
required to the engine hardware, engine electronic controls, ancillary and Noise Vibration
and Harshness (NVH) mitigation systems. Through contacts with industry NVH
suppliers, and manufacturer press releases, the agencies believe that the NVH treatments
will be limited to the mitigation of fuel system noise, specifically from the injectors and
the fuel lines. To that end, the agencies estimate SGDI costs at $251 for an inline 4-
cylinder and $326 for V6 and $353 for V8 including the low complexity ICM markup
value of 1.11 (2007$). The preceding costs are for a MY 2012 vehicle. With time-based
learning applied, these costs would be $222, $287, and $301 for a MY 2016 vehicle
(2007$).
3.4.2.1.8 Combustion Restart (CBRST)
Combustion restart allows "start-stop" functionality of DI engines through the
implementation of an upgraded starter with bi-directional rotation to allow precise
crankshaft positioning prior to subsequent fuel injection and spark ignition, allowing
engine restart. This method of implementing engine stop/start functionality allows not
only the fuel savings from not idling the engine, but also reduces fuel consumption as the
engine speeds up to its operational speed. A Direct Injection (DI) fuel system is required
for implementation of this technology.
NHTSA reviewed the MY 2011 CAFE final rule assumptions and determined that
due to technical risks with its implementation, this technology will be made available in
the CAFE model in MY 2014 at the earliest. Some of the risks are associated with
unresolved issues regarding the impact of very high or very low ambient air temperatures
on the ability to start the engine in the described manner. Although the starter motor can
provide fail-safe starting capability in these temperature limited areas, strategies must be
developed to manage the transitions. Others relate to production readiness.
Additional hardware is required to implement combustion restart, beyond SGDI.
This includes a battery sensor, incremental wiring and high current switching, an
incremental crank position sensor, and, in the case of an automatic transmission
applications, a transmission oil pump to allow for torque converter continuity.
BMW has published a 3.5 percent fuel consumption effectiveness over the NEDC
drive cycle for combustion restart, 5 and AVL a 4.8 percent effectiveness.26 However,
these reported effectiveness levels could potentially be reduced significantly on the EPA
combined drive cycle, as combustion restart does not save fuel on the highway drive
3-30
-------
Technologies Considered in the Agencies' Analysis
cycle. Therefore, NHTSA estimates the fuel consumption effectiveness for CBRST to
range from 2 to 2.5 percent.
Regarding the cost estimate, NHTSA determined that the estimate of $118 from
the 2008 Martec Report cost estimates for individual pieces was the best available. The
total RPE cost (excluding transmission pump) is $141 at high volumes, which includes
$70 for upgrading the starter, $10 for a battery sensor and wiring, $10 for high current
switch and $4 for crank sensor a totaling $94 (non-RPE) cost. Applying an indirect cost
multiplier of 1.25, for a medium complexity technology, results in a compliance cost of
$118 for a MY 2012 vehicle and will be reduced in future years with the application of
time-based learning.
3.4.2.1.9 Turbocharging and Downsizing (TRBDS)
The specific power of a naturally aspirated engine is primarily limited by the rate
at which the engine is able to draw air into the combustion chambers. Turbocharging and
supercharging (grouped together here as boosting) are two methods to increase the intake
manifold pressure and cylinder charge-air mass above naturally aspirated levels.
Boosting increases the airflow into the engine, thus increasing the specific power level,
and with it the ability to reduce engine displacement while maintaining performance.
This effectively reduces the pumping losses at lighter loads in comparison to a larger,
naturally aspirated engine.
Almost every major manufacturer currently markets a vehicle with some form of
boosting. While boosting has been a common practice for increasing performance for
several decades, turbocharging has considerable potential to improve fuel economy and
reduce CC>2 emissions when the engine displacement is also reduced. Specific power
levels for a boosted engine often exceed 100 hp/L, compared to average naturally
aspirated engine power densities of roughly 70 hp/L. As a result, engines can
conservatively be downsized roughly 30 percent to achieve similar peak output levels. In
the last decade, improvements to turbocharger turbine and compressor design have
improved their reliability and performance across the entire engine operating range. New
variable geometry turbines and ball-bearing center cartridges allow faster turbocharger
spool-up (virtually eliminating the once-common "turbo lag") while maintaining high
flow rates for increased boost at high engine speeds. However, even with turbocharger
improvements, maximum engine torque at very low engine speed conditions, for example
launch from standstill, is increased less than at mid and high engine speed conditions.
The potential to downsize engines may be less on vehicles with low displacement to
vehicle mass ratios in order to provide adequate acceleration from standstill, particularly
up grades or at high altitudes.
Use of GDI systems with turbocharged engines and air-to-air charge air cooling
also reduces the fuel octane requirements for knock limited combustion and allows the
3-31
-------
Draft Joint Technical Support Document
use of higher compression ratios. Ford's "Ecoboost" downsized, turbocharged GDI
engines introduced on MY 2010 vehicles allow the replacement of V8 engines with V6
engines with improved in 0-60 mph acceleration and with fuel economy improvements of
up to 12 percent.27
NHTSA and EPA reviewed NHTSA's MY 2011 CAFE final rule estimates that a
turbocharged and downsized engine at equivalent performance levels would offer an
effectiveness improvement of 2 to 5 percent over a naturally-aspirated SGDI engine of
comparable performance. This would equate to a 12 to 14 effectiveness improvement
over baseline fixed-valve engine, similar to the estimate for Ford's Ecoboost. NHTSA
and EPA kept these estimates for this NPRM.
Recently published data with advanced spray-guided injection systems and more
aggressive engine downsizing targeted towards reduced fuel consumption and CO2
emissions reductions indicate that the potential for reducing CC>2 emissions for
turbocharged, downsized GDI engines may be as much as 15 to 30% relative to port-fuel-
injected engines.14'15'16'17'18 EPA and NHTSA seek comment on how best to determine
these values. Confidential manufacturer data suggests an incremental range of fuel
consumption and COi emission reduction of 4.8 to 7.5 percent for turbocharging and
downsizing. Other publicly-available sources suggest a fuel consumption and COi
emission reduction of 8 to 13 percent compared to current-production naturally-aspirated
engines without friction reduction or other fuel economy technologies: a joint technical
paper by Bosch and Ricardo suggesting fuel economy gain of 8 to 10 percent for
downsizing from a 5.7 liter port injection V8 to a 3.6 liter V6 with direct injection using a
wall-guided direct injection system;28 a Renault report suggesting a 11.9 percent NEDC
fuel consumption gain for downsizing from a 1.4 liter port injection in-line 4-cylinder
engine to a 1.0 liter in-line 4-cylinder engine, also with wall-guided direct injection;29 and
a Robert Bosch paper suggesting a 13 percent NEDC gain for downsizing to a
turbocharged DI engine, again with wall-guided injection.30 These reported fuel
economy benefits show a wide range depending on the GDI technology employed.
As noted above, the agencies relied on engine teardown analyses conducted by
EPA, FEV and Munro to develop costs for turbocharged GDI engines.31 Teardown
studies are one of the most effective ways to estimate technology costs. The study
showed the cost of a turbocharger (variable geometry turbo, air-to-air charge air cooler,
auxiliary cooling pump, lubes, upgrades to the exhaust manifold, and necessary controls)
for the 14 DOHC engine studied to be $372 (2007$). The study also showed an engine
downsizing savings for downsizing a 2.4L DOHC engine to a 1.6L DOHC engine of $60
(2007$). These savings were the result of reduced bore spacing, a shorter camshaft and
crankshaft, removal of two balance shafts and associated hardware, and a smaller intake
manifold. These values ($372 and -$60) have been used for 14 engines adding boost and
for 14 engines undergoing downsizing to a smaller 14, respectively. These costs are
applicable for a MY 2012 vehicle. With time-based learning applied, these costs would
be $329 and -$53 for a MY 2016 vehicle.
3-32
-------
Technologies Considered in the Agencies' Analysis
Consistent with the 2008 staff report, the agencies estimate direct manufacturing
costs associated with downsizing to be $50 per cylinder, $10 per valve, and $100 per cam
shaft for the 2015 model year (2006$). Applying a 1 year of time-based learning to these
costs, and adjusting them to 2007 dollars, results in 2016 direct manufacturing costs of
$50 per cylinder, $10 per valve, and $100 per cam shaft. A summary of the final costs
and how they were calculated is shown in Table 3-5 for a MY 2016 vehicle in 2007$.
For a MY 2012 vehicle, 4 years of time-based learning at 3% per year would be added
back to the values shown in Table 3-5.
3-33
-------
Draft Joint Technical Support Document
Table 3-5 Turbocharging and Downsizing and Other Camshaft Configuration Costs in 2016 (2007 dollars)
Turbo w/o downsize
Downsize w/o turbo
Turbo with downsize
c«
, OJ
H 00
CS C
U 2
o
technology
Turbocharge (single)
Turbocharge (twin)
Downsize to 14 DOHC
Downsize to 14 DOHC
Downsize to 14 DOHC
Downsize to 14 DOHC
Downsize to 13 DOHC
Downsize to V6 DOHC
Downsize to V6 DOHC
Downsize to V6 DOHC
Downsize to V6 DOHC
Downsize to 14 DOHC & add
turbo
Downsize to 14 DOHC & add
turbo
Downsize to 14 DOHC & add
turbo
Downsize to 14 DOHC & add
turbo
Downsize to 13 DOHC & add
turbo
Downsize to V6 DOHC & add
twin turbo
Downsize to V6 DOHC & add
twin turbo
Downsize to V6 DOHC & add
twin turbo
Downsize to V6 DOHC & add
twin turbo
Convert to V6 DOHC
Convert to V6 DOHC
Convert to V8 DOHC
Convert to V8 DOHC
Convert to V8 DOHC
incremental to
Base engine
Base engine
V6 DOHC
V6 SOHC
V60HV
14 DOHC (larger)
14 DOHC
V8 DOHC
V8 SOHC 2V
V8 SOHC 3V
V80HV
V6 DOHC w/o turbo
V6 SOHC w/o turbo
V6 OHV w/o turbo
14 DOHC (larger) w/o
turbo
14 DOHC w/o turbo
V8 DOHC w/o turbo
V8 SOHC 2V w/o turbo
V8 SOHC 3V w/o turbo
V8 OHV w/o turbo
V6 SOHC
V6OHV
V8 SOHC 2V
V8 SOHC 3V
V8OHV
Direct manufacturing costs in 2016
Turbo
13/14
$329
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
$398
$398
$398
$329
$329
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
V6/V8
n/a
$598
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
$598
$598
$598
$598
n/a
n/a
n/a
n/a
n/a
Engine downsize costs
cylinder #
change at
$50/
n/a
n/a
-2
_2
-2
0
-1
-2
_2
+2
-2
-2
_2
_2
0
-1
-2
-2
_2
_2
0
0
0
0
0
valve #
change at
$10/
n/a
n/a
-8
+4
+4
0
-4
-8
+8
0
+8
-8
+4
+4
0
-4
-8
+8
0
+8
+12
+12
+16
+8
+16
Cam*
change at
$100/
n/a
n/a
_2
0
+3
0
0
0
+2
+2
+3
_2
0
+3
0
0
0
+2
+2
+3
+2
+3
+2
+2
+3
Resultant
Downsize
cost
n/a
n/a
-$379
-$60
$239
-$53
-$90
-$179
$179
$299
$279
-$379
-$60
$239
-$53
-$90
-$179
$179
$100
$279
$319
$418
$359
$279
$419
1C Mark
up
1.11
1.11
1.11
1.11
1.11
1.11
1.11
1.11
1.11
1.11
1.11
1.25
1.25
1.25
1.25
1.25
1.25
1.25
1.25
1.25
1.11
1.11
1.11
1.11
1.11
Compliance
Cost*
$366
$663
-$337
-$53
$265
-$47
-$80
-$160
$199
$332
$310
$214
$453
$797
$372
$344
$613
$971
$872
$1,096
$354
$464
$398
$310
$509
* Note that, where downsizing results in cost savings, the compliance cost is calculated as the 1C markup less 1
cost. The absolute value of the direct manufacturing cost is then subtracted from that to arrive at the end result.
manufacturing cost of-$379, the compliance cost would be (1.11-1) x |-$379 | - |-$379 | =-$337.
which is then multiplied by the absolute value of the direct manufacturing
For example, for the V6 DOHC downsized to the 14 DOHC at a direct
3-34
-------
3.4.2.1.10 Cooled Exhaust Gas Recirculation/EGR Boost (EGRB)
Cooled exhaust gas recirculation (cooled EGR) or EGR Boost is a combustion
concept that involves utilizing EGR as a charge dilutant for controlling combustion
temperatures and cooling the EGR prior to its introduction to the combustion system.
Higher exhaust gas residual levels at part load conditions reduce pumping losses for
increased fuel economy. Cooled EGR reduces knock sensitivity which enables the use of
more optimal spark advance or enables compression ratio to be increased for improved
thermal efficiency, and increased fuel economy. Currently available turbo, charge air
cooler, and EGR cooler technologies are sufficient to demonstrate the feasibility of this
concept.
However, this remains a technology with a number of issues that still need to be
addressed and for which there is no production experience. EGR system fouling
characteristics could be potentially worse than diesel EGR system fouling, due to the
higher HC levels found in gasoline exhaust. Turbocharger compressor contamination
may also be an issue for low pressure EGR systems. Additionally, transient controls of
boost pressure, EGR rate, cam phasers and intake charge temperature to exploit the
cooled EGR combustion concept will require development beyond what has already been
accomplished by the automotive industry. These are all "implementation readiness"
issues that must be resolved prior to putting EGR Boost into high volume production.
NHTSA has concluded that these implementation issues could be resolved
and this technology could be brought to production by MY 2013. Supporting this
conclusion, MEMA has previously suggested a 5 to 7 percent effectiveness for
cooled EGR systems, although without boosting.32
Two public sources indicate a 10 to 20 percent fuel consumption
effectiveness for a downsized DI engine with cooled EGR compared to a naturally
aspirated baseline engine33 and a 4 percent fuel consumption effectiveness for
cooled EGR compared to a conventional downsized DI turbocharged engine.34
Based on the data from these reports, NHTSA estimates the incremental reduction
in fuel consumption for EGR Boost to be 4 percent over a turbocharged and
downsized DI engine. Thus, if TRBDS precedes EGRB, adding the 12 percent gain
from TRBDS to the 4 percent gain from EGRB results in total fuel consumption
reduction of 16 percent. This is in agreement with the range suggested in the Lotus
report.
Regarding costs, the addition of EGR cooler and EGR valve were estimated in
NHTSA's MY 2011 rule to have an incremental RPE cost impact of approximately $173
based on confidential individual component cost data from 2008 Martec describing EGR
cooler costs of $75, EGR valve costs of $20 and associated piping costs of $20, totaling
$115 (non-RPE). For purposes of this NPRM, NHTSA found no information to indicate
that these estimates were inaccurate. To that end, NHTSA applied an indirect cost
3-35
-------
Draft Joint Technical Support Document
multiplier of 1.25, for a medium complexity technology, resulting in a compliance cost of
$144 for MY 2012 vehicles with time-based learning applied. However, given the lack
of public data on this technology, the agencies seek comment on these assumptions.
3.4.2.1.11 Diesel Engine Technologies
Diesel engines have several characteristics that give them superior fuel efficiency
compared to conventional gasoline, spark-ignited engines. Pumping losses are much
lower due to lack of (or greatly reduced) throttling. The diesel combustion cycle operates
at a higher compression ratio, with a very lean air/fuel mixture, and turbocharged light-
duty diesels typically achieve much higher torque levels at lower engine speeds than
equivalent-displacement naturally-aspirated gasoline engines. Additionally, diesel fuel
has a higher energy content per gallon.11 However, diesel fuel also has a higher carbon to
hydrogen ratio, which increases the amount of CO2 emitted per gallon of fuel used by
approximately 15% over a gallon of gasoline.
Diesel engines have emissions characteristics that present challenges to meeting
federal Tier 2 NOx emissions standards. It is a significant systems-engineering
challenge to maintain the fuel consumption advantage of the diesel engine while meeting
U.S. emissions regulations. Fuel consumption can be negatively impacted by emissions
reduction strategies depending on the combination of strategies employed. Emission
compliance strategies for diesel vehicles sold in the U.S. are expected to include a
combination of combustion improvements and aftertreatment. These emission control
strategies are being introduced on Tier 2 light-duty diesel vehicles today
To achieve U.S. Tier 2 emissions limits, roughly 45 to 65 percent more NOX
reduction is required compared to the Euro VI standards. Additionally, as discussed
below, there may be a fuel consumption penalty associated with diesel aftertreatment
since extra fuel is needed for the aftertreatment, and this extra fuel is not used in the
combustion process of the engine that provides torque to propel the vehicle.
Light-duty diesel emissions control systems capable of meeting Tier 2 Bin 5
emission standards are already in production. Several key advances in diesel technology
have made it possible to reduce emissions coming from the engine prior to aftertreatment.
These technologies include improved fuel systems (higher injection pressure and
multiple-injection capability), advanced controls and sensors to optimize combustion and
emissions performance, higher EGR levels and EGR cooling to reduce NOX, and
advanced turbocharging systems.
On the aftertreatment side, the traditional 3-way catalyst aftertreatment found on
gasoline-powered vehicles is ineffective due to the lean-burn combustion of a diesel. All
diesels will require a diesel particulate filter (DPF) or catalyzed diesel particulate filter
h Burning one gallon of diesel fuel produces about 15 percent more carbon dioxide than gasoline due to the
higher density and carbon to hydrogen ratio.
3-36
-------
(CDPF), a diesel oxidation catalyst (DOC), and a NOX reduction strategy to comply with
Tier 2 emissions standards. The most common NOX reduction strategies include the use
of lean NOX traps (LNT) or selective catalytic reduction (SCR), which are outlined
below.
3. Diesel Engine with Lean NOX Trap (LNT) Catalyst After-
Treatment
A lean NOX trap operates, in principle, by oxidizing NO to NO2 in the exhaust
and storing NO2 on alkali sorbent material. When the control system determines (via
mathematical model or a NOX sensor) that the trap is saturated with NOX, it switches the
engine into a rich operating mode or may in some cases inject fuel directly into the
exhaust stream to produce excess hydrocarbons that act as a reducing agent to convert the
stored NOX to NI and water, thereby "regenerating" the LNT and opening up more
locations for NOX to be stored. LNTs preferentially store sulfate compounds from the
fuel, which can reduce catalytic performance. The system must undergo periodic
desulfurization by operating at a net-fuel-rich condition at high temperatures in order to
retain NOX trapping efficiency.
NHTSA and EPA have concluded that the application of diesel engines on small
vehicles is not a viable or cost effective option. The two agencies have also concluded
that LNT-based diesel engines are best suited to smaller vehicle. Thus for purposes of
this NPRM the application of LNT-based diesel engines has not been included and cost
and effectiveness estimates were not generated.
4. Diesel Engine with Selective Catalytic Reduction (SCR) After-
Treatment
An SCR aftertreatment system uses a reductant (typically, ammonia derived from
urea) that is injected into the exhaust stream ahead of the SCR catalyst. Ammonia
combines with NOX in the SCR catalyst to form NI and water. The hardware
configuration for an SCR system is more complicated than that of an LNT, due to the
onboard urea storage and delivery system (which requires a urea pump and injector to
inject urea into the exhaust stream). While a rich engine-operating mode is not required
for NOX reduction, the urea is typically injected at a rate of approximately 3 percent of
the fuel consumed. Manufacturers designing SCR systems intend to align urea tank
refills with standard maintenance practices such as oil changes. As is the case with LNT-
based diesels, EPA and NHTSA project that SCR-based diesel engines will be available
within the next couple of years. Mercedes-Benz recently introduced two 2009 model
year vehicles R320 and GL320, both of which are certified to Tier 2, Bin 5 emission
standards. Based on public announcements from several other companies, an increased
number of product offerings from multiple companies are expected over the next few
years.
In order to maintain equivalent performance to comparable gasoline-engine
vehicles, an in-line 4-cylinder diesel engine, with displacement varying around 2.8 liters
3-37
-------
Draft Joint Technical Support Document
was assumed to replace a V6 gasoline base engine for Performance Compact,
Performance Midsize, Large Passenger Car, Minivan, and Midsize Truck for the CAFE
model. A V-6 diesel engine, with displacement varying around 4.0 liters to meet vehicle
performance requirements, was assumed to replace a V8 gasoline base engine for Large
Truck and Performance Large Car vehicle classes for the CAFE model. It was also
assumed that diesel engines for these classes would utilize SCR aftertreatment systems.
Confidential manufacturer and non-confidential comment data submitted in response to
NHTSA's past rulemaking for diesel engines showed a fuel consumption reduction in the
range of 16.7 percent to 26.7 percent or a 15% to 23% COi emission reduction over a
baseline gasoline engine.
NHTSA's MY 2011 CAFE final rule, which was supported by confidential
manufacturer data, estimated the fuel consumption reduction of a SCR-based diesel
engine to be between 19 to 25 percent over a baseline gasoline engine. NHTSA and EPA
have revisited these values and found them to be reasonable. EPA estimates a 15 to 26%
CC>2 reduction for light-duty diesels equipped with SCR.
Diesel engines are more costly than port-injected spark-ignition gasoline engines.
These higher costs result from:
Fuel systems (higher pressures and more responsive injectors);
Controls and sensors to optimize combustion and emissions performance;
Engine design (higher cylinder pressures require a more robust engine, but
higher torque output means diesel engines can have reduced displacement);
Turbocharger(s);
Aftertreatment systems, which tend to be more costly for diesels;
Due to a significant decrease in platinum group metal prices since NHTSA's MY
2011 CAFE final rule analysis, the agencies chose to re-analyze diesel costs. In EPA's
2008 Staff Report, costs were considered for two types of diesel systems: one using a
lean-NOx trap (LNT) along with a diesel particulate filter (DPF); and one using a
selective catalytic reduction (SCR) system along with a DPF. In that report, EPA
estimated direct manufacturing costs to range from $1,860 for the small car (LNT plus
DPF) to $2,710 for the large truck (SCR plus DPF). For comparison, the NESCCAF
study showed direct manufacturing costs of $1,500 to $1,950. More recently, NHTSA's
2011 CAFE final rule showed direct manufacturing costs of $2,670 for a 4-cylinder
engine using a LNT plus DPF system, $3,735 for a 6-cylinder engine using a SCR plus
DPF system, and $4,668 for an 8-cylinder engine using a SCR plus DPF system.
NHTSA noted that estimates in the MY 2011 CAFE final rule were higher than those
shown in the proposed rule due largely to the spike in platinum group metal prices that
had occurred in the months just prior to issuing the final rule.
3-38
-------
The following diesel engine costs were developed by first taking a look back at
EPA's 2008 Staff Report, which reveals a couple of factors that resulted in somewhat
misleading costs. First, the engine costs estimated there did not take into account the
downsizing that would occur when moving from a gasoline engine to a diesel engine
(provided equivalent performance was maintained). Second, the engine costs used in that
analysis were actually stated in terms of 2002 dollars rather than 2006 dollars in which
the report was meant to be stated. The agencies decided that an update to the engine-
related costs would provide a much better cost estimate for converting to diesel. This
was done by starting with the source for engine costs in the 2008 staff report which was
an October 2005 EPA Interim Report35 which, in turn, sourced estimates from a 2003
study done by FEV for EPA contained within a 2004 EPA Interim Technical Report.3
These direct manufacturing costs are reproduced in Table 3-6.
36
Table 3-6 Diesel Engine Direct Manufacturing Source Costs, Incremental to a Baseline Gasoline
Engine (2002 dollars)
Component(s)
Gasoline engine (baseline)
Diesel engine
Add high-pressure, common rail diesel fuel injection system
Delete gasoline fuel injection system
Add variable geometry turbocharger
Delete gasoline ignition system
Delete fuel pump and other changes to fuel system
Enhance powertrain mounting system
Other engine changes
Add air intercooler, ducts, and sensor
Larger battery and starter, add glow plugs
Delete exhaust gas oxygen sensor*
Add supplemental heater
Modify transmission
Enhance sound insulation package
Smaller radiator
Total
Large SUV
5LV8
4LV8
$980
-$245
$175
-$120
-$94
$87
$80
$80
$72
-$60
$50
$25
$25
-$13
$1,042
Midsize
2.4L 14
2.2L 14
$630
-$165
$126
-$75
-$75
$107
$70
$55
$50
-$30
$15
$25
$10
-$4
$739
Note: Table reproduced from EPA420-R-05-012, October 2005
Building on the direct manufacturing costs shown in Table 3-6, the agencies used
appropriate scaling to estimate the costs for replacing a baseline gasoline engine with a
diesel engine for the following four situations: a large car converted from a 4.5L V8
gasoline to a 3L V6 diesel; a medium/large MPV converted from a 3.2L V6 to a 2.8L 14
diesel; a small truck converted from a 3.2L V6 gasoline to a 2.8L 14 diesel; and a large
truck converted from a 5.6L V8 gasoline to a 4L V6 diesel. A small car conversion was
not considered since the diesel conversion for the small car was not considered to be a
viable or cost effective option. The results for the four base gasoline to diesel
conversions are shown in Table 3-7. Values from Table 3-6 have been updated to 2007
dollars using the GDP price deflator factor of 1.15 (see Appendix 3.A). Since the source
costs were developed in 2003, this analysis conservatively considers the costs shown in
Table 3-7 as being applicable to the 2012 model year.
3-39
-------
Draft Joint Technical Support Document
Table 3-7 Diesel Engine Direct Manufacturing Scaled-Costs in 2012, Incremental to Baseline
Gasoline Engine (2007 dollars)
Component(s)
Gasoline engine (baseline)
Diesel engine
Add high-pressure, common rail diesel
fuel injection system
Delete gasoline fuel injection system
Add variable geometry turbocharger
Delete gasoline ignition system
Delete fuel pump and other changes to
fuel system
Enhance powertrain mounting system
Other engine changes
Add air intercooler, ducts, and sensor
Larger battery and starter, add glow
plugs
Delete exhaust gas oxygen sensor*
Add supplemental heater
Modify transmission
Enhance sound insulation package
Smaller radiator
Engine downsize credit
Total
Large
car
4.5L V8
3LV6
$1,026
-$89
$173
-$138
-$108
$100
$86
$78
$70
$0
$37
$29
$20
-$15
-$185
$1,085
Med/large
MPV
3.2L V6
2.8L 14
$724
-$73
$145
-$112
-$86
$123
$80
$63
$57
$0
$17
$29
$11
-$10
-$390
$580
Small
truck
3.2L V6
2.8L 14
$724
-$73
$145
-$112
-$86
$123
$80
$63
$57
$0
$17
$29
$11
-$10
-$390
$580
Large
truck
5.6L V8
4LV6
$1,026
-$89
$201
-$138
-$108
$100
$86
$92
$70
$0
$57
$29
$29
-$15
-$185
$1,156
Notes
(see text
below)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
: Note: Oxygen sensor removals are included in aftertreatment costs.
The costs shown in Table 3-7 were scaled in the following ways:
1. Large car and large track calculated as 75% the cost of Table 3-6's large SUV and
25% of midsize car; medium/large MPV and small track calculated as equal to Table
3-6's midsize car. Values converted to 2007 dollars using GDP factor of 1.15.
2. The estimates generated by FEV for eliminating the gasoline fuel injection systems
were considerably larger than EPA & NHTSA believed was appropriate. Therefore,
to remain conservative, these costs were estimated, in 2007 dollars as follows: large
car and large truck were calculated using incremental costs of $8/injector, $20/fuel
rail, and $5 for a pressure damper or $8x8+$20+$5=$89; medium/large MPV and
small truck were calculated using incremental costs of $8/injector, $20/fuel rail, and
$5 for a pressure damper or $8x6+$20+$5=$73.
3. Large car calculated as the average of Table 3-6's large SUV and midsize car;
medium/large MPV and small truck calculated as equal to Table 3-6's midsize car,
and large track calculated as equal to Table 3-6's large SUV. Values converted to
2007 dollars using GDP factor of 1.15.
3-40
-------
4. Medium/large MPV and small track calculated as the average of Table 3-6's large
SUV and midsize car; Large car and large truck calculated as equal to Table 3-6's
large SUV. Values converted to 2007 dollars using GDP factor of 1.15.
5. Medium/large MPV and small track calculated as equal to Table 3-6's midsize car;
Large car and large track calculated as equal to Table 3-6's large SUV. Values
converted to 2007 dollars using GDP factor of 1.15.
6. Medium/large MPV and small track calculated as equal to Table 3-6's large SUV;
Large car and large track calculated as equal to Table 3-6's midsize car. Values
converted to 2007 dollars using GDP factor of 1.15.
7. Medium/large MPV and small track calculated as equal to Table 3-6's midsize car;
Large car and large track calculated as the average of Table 3-6's large SUV and
midsize car. Values converted to 2007 dollars using GDP factor of 1.15.
8. Medium/large MPV and small track calculated as equal to Table 3-6's midsize car;
Large car calculated as the average of Table 3-6's large SUV and midsize car; Large
truck calculated as equal to Table 3-6's large SUV. Values converted to 2007 dollars
using GDP factor of 1.15.
9. Medium/large MPV and small track calculated as equal to Table 3-6's midsize car;
Large car and large track calculated as the average of Table 3-6's large SUV and
midsize car. Values converted to 2007 dollars using GDP factor of 1.15.
10. Medium/large MPV and small track calculated as equal to Table 3-6's midsize car;
Large car and large track calculated as the average of Table 3-6's large SUV and
midsize car. Values converted to 2007 dollars using GDP factor of 1.15.
11. Medium/large MPV and small track calculated as equal to Table 3-6's midsize car;
Large car calculated as the average of Table 3-6's large SUV and midsize car; Large
truck calculated as equal to Table 3-6's large SUV. Values converted to 2007 dollars
using GDP factor of 1.15.
12. Values from Table 3-6 converted to 2007 dollars using GDP factor of 1.15.
13. Medium/large MPV and small track calculated as equal to Table 3-6's midsize car;
Large car calculated as the average of Table 3-6's large SUV and midsize car; Large
truck calculated as equal to Table 3-6's large SUV. Values converted to 2007 dollars
using GDP factor of 1.15.
14. Medium/large MPV and small track calculated as the average of Table 3-6's large
SUV and midsize car; Large car and large truck calculated as equal to Table 3-6's
large SUV. Values converted to 2007 dollars using GDP factor of 1.15.
15. Based on the approach presented in the turbocharging/downsizing section (see section
3.5.2.1.11), the savings associated with downsizing the gasoline engine were
calculated by estimating the cost in 2007 dollars of each cylinder at $51, each valve at
3-41
-------
Draft Joint Technical Support Document
$10, and each cam at $103.' Therefore, the large car and large track, which each lose
two cylinders (-$102), eight valves (-$82) and no cams realize a $185 savings. The
medium/large MPV and small truck would each lose two cylinders (-$102) and eight
valves (-$82) and two cams (-$205) for a savings of $390.
For the diesel aftertreatment systems, the approach taken is consistent with the
approach taken in EPA's 2007/2010 Highway Diesel rale and EPA's recent locomotive
and marine rule.37 For platinum group metal costs, monthly average prices as of March
2009 as reported by Johnson-Matthey were used.38 Those values were $l,085/troy ounce
for platinum and $l,169/troy ounce for rhodium. Aftertreatment devices were sized
according to the diesel engine displacement with a 1:1 ratio for both the SCR catalyst and
the DPF, and a 0.5:1 ratio for the DOC (i.e., the DOC is half the displacement of the
engine). The end result for aftertreatment devices, including a urea dosing unit, urea tank
and necessary brackets and heaters, are shown in Table 3-8. Also shown in Table 3-8 are
the savings associated with removal of the gasoline catalyst. Note that the gasoline
catalyst was sized according to the gasoline engine that served as the baseline engine.
Table 3-8 Diesel Aftertreatment Direct Manufacturing Costs in 2012 (2007 dollars)
Component(s)
Gasoline engine (baseline)
Diesel engine
DOC
DPF (includes a $20 pressure sensor for
OBD & sensing)
SCR system (includes a $50 NOx sensor for
OBD & sensing)
Removal of gasoline catalysts & sensors
Total
Large car
4.5L V8
3LV6
$277
$534
$904
-$401
$1,314
Med/large
MPV
3.2L V6
2.8L 14
$257
$503
$904
-$288
$1,376
Small truck
3.2L V6
2.8L 14
$257
$503
$914
-$298
$1,376
Large truck
5.6L V8
4LV6
$339
$668
$996
-$483
$1,520
The incremental costs to convert from a gasoline to a diesel engineTable 3-7
and Table 3-8 combinedare shown in Table 3-9.
Table 3-9 Direct Manufacturing Costs to Convert from a Gasoline to Diesel System in 2012 (2007
dollars)
1 These are the correct costs for the 2015MY in 2007$. But, they are used here inappropriately as 2012MY
values. Technically, 3 years of time-based learning should have been backed out to get 2012MY values.
So, the 2012MY costs are slightly underestimated by roughly $10.
3-42
-------
Component(s)
Gasoline engine (baseline)
Diesel engine
Engine-related costs
Aftertreatment
Total
Large car
4.5L V8
3LV6
$1,085
$1,314
$2,399
Med/large
MPV
3.2L V6
2.8L 14
$580
$1,376
$1,956
Small truck
3.2L V6
2.8L 14
$580
$1,376
$1,956
Large truck
5.6L V8
4LV6
$1,156
$1,520
$2,676
This analysis applies time-based learning to diesel systems and a 1.25 ICM for a
medium complexity rating. Therefore, the MY 2016 compliance costs are as shown in
Table 3-10.
Table 3-10 Compliance Costs to Convert from a Gasoline to Diesel System in 2016 (2007 dollars)
Component(s)
Gasoline engine (baseline)
Diesel engine
Total
Large car
4.5L V8
3LV6
$2,655
Med/large
MPV
3.2L V6
2.8L 14
$2,164
Small truck
3.2L V6
2.8L 14
$2,164
Large truck
5.6L V8
4LV6
$2,961
In the NHTSA analysis, and consistent with the above analysis, SCR-based diesel
costs for a 2.8L in-line 4-cylinder from a V-6 gasoline engine were estimated to be $2445
in MY 2012. A MY 2012 cost of $3345 was estimated for a conversion to 4.0L V-6
diesel from a V-8 gasoline engine. These MY 2012 estimates were obtained using the in-
line 4 cylinder cost of $1956 and V-6 cost of $2676, from Table 3-9, and multiplying
them by the medium complexity ICM markup of 1.25 and are eligible for time-based
learning cost reductions.
3.4.2.2 Transmission Technologies
NHTSA and EPA have also reviewed the transmission technology estimates used
in the MY 2011 CAFE final rule. In doing so NHTSA and EPA considered or
reconsidered all available sources and updated the estimates as appropriate. The section
below describes each of the transmission technologies considered for this rulemaking.
3.4.2.2.1 Improved Automatic Transmission Control (IATC) (Aggressive Shift Logic
and Early Torque Converter Lockup)
Calibrating the transmission shift schedule to upshift earlier and quicker, and to
lock-up or partially lock-up the torque converter under a broader range of operating
conditions can reduce fuel consumption and COi emissions. However, this operation can
result in a perceptible degradation in noise, vibration, and harshness (NVH). The degree
to which NVH can be degraded before it becomes noticeable to the driver is strongly
3-43
-------
Draft Joint Technical Support Document
influenced by characteristics of the vehicle, and although it is somewhat subjective, it
always places a limit on how much fuel consumption can be improved by transmission
control changes. Given that the Aggressive Shift Logic and Early Torque Converter
Lockup are best optimized simultaneously due to the fact that adding both of them
primarily requires only minor modifications to the transmission or calibration software,
these two technologies are combined in the modeling.
3.4.2.2.2 Aggressive Shift Logic
During operation, an automatic transmission's controller manages the operation of
the transmission by scheduling the upshift or downshift, and locking or allowing the
torque converter to slip based on a preprogrammed shift schedule. The shift schedule
contains a number of lookup table functions, which define the shift points and torque
converter lockup based on vehicle speed and throttle position, and other parameters such
as temperature. Aggressive shift logic (ASL) can be employed in such a way as to
maximize fuel efficiency by modifying the shift schedule to upshift earlier and inhibit
downshifts under some conditions, which reduces engine pumping losses and engine
friction. The application of this technology does require a manufacturer to confirm that
drivability, durability, and NVH are not significantly degraded.
The agencies have estimated the effectiveness of aggressive shift logic at 1 to 2
percent. The agencies have estimated the cost at $32 for a MY 2012 vehicle, regardless
of size (2007$).j With time-based learning applied, this cost would be $28 for a MY
2016 vehicle (2007$)
3.4.2.2.3 Early Torque Converter Lockup
A torque converter is a fluid coupling located between the engine and
transmission in vehicles with automatic transmissions and continuously-variable
transmissions (CVT). This fluid coupling allows for slip so the engine can run while the
vehicle is idling in gear (as at a stop light), provides for smoothness of the powertrain,
and also provides for torque multiplication during acceleration, and especially launch.
During light acceleration and cruising, the inherent slip in a torque converter causes
increased fuel consumption, so modern automatic transmissions utilize a clutch in the
torque converter to lock it and prevent this slippage. Fuel consumption can be further
reduced by locking up the torque converter at lower vehicle speeds, provided there is
sufficient power to propel the vehicle, and noise and vibration are not excessive.k If the
torque converter cannot be fully locked up for maximum efficiency, a partial lockup
strategy can be employed to reduce slippage. Early torque converter lockup is applicable
' This reflects a revisiting of component costs for NHTSA since the 2011 CAFE FRM.
k Although only modifications to the transmission calibration software are considered as part of this
technology, very aggressive early torque converter lock up may require an adjustment to damper stiffness
and hysteresis inside the torque converter.
3-44
-------
to all vehicle types with automatic transmissions. Some torque converters will require
upgraded clutch materials to withstand additional loading and the slipping conditions
during partial lock-up. As with aggressive shift logic, confirmation of acceptable
drivability, performance, durability and NVH characteristics is required to successfully
implement this technology.
Regarding the effectiveness of Improved Automatic Transmission Control, the
MY2011 CAFE final rule, which was supported by the 2002 NAS and NESCCAF reports
as well as confidential manufacturer data, estimated an effectiveness improvement of 1 to
2 percent for aggressive shift logic and 0.5 percent for early torque converter lockup.
These estimates are in agreement with the values stated in the NESCCAF report and
confidential manufacturer data. For the purpose of this NPRM, NHTSA and EPA
concluded that the combined estimated effectiveness is 1.5 to 2.5% reduction in COi
emissions and fuel consumption.
Regarding cost estimates, the agencies estimate the cost at $29, regardless of
vehicle size and including a low complexity ICM of 1.11 for a MY 2012 vehicle(2007$).
With time-based learning, this cost would be $25 for a MY 2016 vehicle (2007$). For
IATC (or the combination of aggressive shift logic and early torque converter lockup),
the cost, including the 1.11 ICM, would be $60 for a MY 2012 vehicle and $53 for a MY
2016 vehicle after time-based learning (2007$).
3.4.2.2.4 Automatic 6-, 7- and 8-Speed Transmissions (NAUTO)
Manufacturers can also choose to replace 4- and 5-speed transmission with 6-, 7-,
or 8-speed automatic transmissions. Additional ratios allow for further optimization of
engine operation over a wider range of conditions, but this is subject to diminishing
returns as the number of speeds increases. As additional planetary gear sets are added
(which may be necessary in some cases to achieve the higher number of ratios),
additional weight and friction are introduced. Also, the additional shifting of such a
transmission can be perceived as bothersome to some consumers, so manufacturers need
to develop strategies for smooth shifts. Some manufacturers are replacing 4- and 5-speed
automatics with 6-speed automatics, and 7- and 8-speed automatics have also entered
production, albeit in lower-volume applications in luxury and performance oriented cars.
As discussed in the MY 2011 CAFE final rule, confidential manufacturer data
projected that 6-speed transmissions could incrementally reduce fuel consumption by 0 to
5 percent from a baseline 4-speed automatic transmission, while an 8-speed transmission
could incrementally reduce fuel consumption by up to 6 percent from a baseline 4-speed
automatic transmission. The 2008 Martec report estimated a cost of $323 (RPE adjusted)
for converting a 4-speed to a 6-speed transmission and a cost of $638 (RPE adjusted) for
converting a 4-speed to an 8-speed transmission. GM has publicly claimed a fuel
economy improvement of up to 4 percent for its new 6-speed automatic transmissions.39
The 2008 EPA Staff Technical Report found a 4.5 to 6.5 percent fuel consumption
improvement for a 6-speed over a 4-speed automatic transmission.40 Based on this
3-45
-------
Draft Joint Technical Support Document
information, NHTSA estimated in the MY 2011 rule, that the conversion to a 6-,7- and 8-
speed transmission (NAUTO) from a 4 or 5-speed automatic transmission with IATC
would have an incremental fuel consumption benefit of 1.4 percent to 3.4 percent, for all
vehicle classes. From a baseline 4 or 5 speed transmission without IATC, the
incremental fuel consumption benefit would be approximately 3 to 6 percent, which is
consistent with the EPA Staff Report estimate.
NHTSA and EPA reviewed these effectiveness estimates and concluded that they
remain accurate. While the CAFE model follows the incremental approach discussed
above, the GHG model estimates the packaged effectiveness of 4.5 to 6.5 percent.
NHTSA and EPA also reviewed the cost estimates from the MY 2011 CAFE final
rule and the 2008 staff report (which assumed use of a Lepelletier gear set) and
concluded that some but not all 6-speed automatic transmissions would be equipped with
Lepelletier gear set, and as such, the estimates were revised to establish the cost for the 6-
speed transmission to be equally divided between application using Lepelletier, and
application of a standard planetary gear set 6-speed automatic transmission as estimated
in the 2008 Martec report (and the MY 211 CAFE final rule). As a result, the final
incremental cost estimate was $170, independent of vehicle type and size and including a
low complexity 1.11 ICM (2007$). In the CAFE model, due to the structure vehicle
classes used, an additional $102 (2007$) is included to account for performance vehicle
classes and for medium and large trucks. This is because for performance classes,
additional gear ratios, such as 7 and 8 speed transmissions may be utilized, and for
medium and large trucks heavier duty transmissions are required. These estimates
represent MY 2012 vehicle costs. With time-based learning, the $170 cost would lower
to $150 for a MY 2016 vehicle (2007$).
3.4.2.2.5 Dual Clutch Transmissions /Automated Manual Transmissions (DCTAM)
An Automated Manual Transmission (AMT) is mechanically similar to a
conventional manual transmission, but shifting and launch functions are automatically
controlled by the electronics. There are two basic types of AMTs, single-clutch and dual-
clutch (DCT). A single-clutch AMT is essentially a manual transmission with automated
clutch and shifting. Because of shift quality issues with single-clutch designs, DCT's
will likely be far more common in the U.S. and are the basis of the estimates that follow.
A DCT uses separate clutches (and separate gear shafts) for the even number gears and
odd-numbered gears. In this way, the next expected gear is pre-selected which allows for
faster and smoother shifting. For example, if the vehicle is accelerating in third gear, the
shaft with gears one, three and five has gear three engaged and is transmitting power.
The shaft with gears two, four, and six is idle, but has gear four engaged. When a shift is
required, the controller disengages the odd-gear clutch while simultaneously engaging the
even-gear clutch, thus making a smooth shift. If, on the other hand, the driver slows
down instead of continuing to accelerate, the transmission will have to change to second
3-46
-------
gear on the idling shaft to anticipate a downshift. This shift can be made quickly on the
idling shaft since there is no torque being transferred on it.
In addition to single-clutch and dual-clutch AMTs, there are also wet clutch and
dry clutch designs which are used for different types of vehicle applications. Wet clutch
AMTs offer a higher torque capacity that comes from the use of a hydraulic system that
cools the clutches. Wet clutch systems are less efficient than the dry clutch systems due
to the losses associated with hydraulic pumping. Additionally, wet AMT's have a higher
cost due to the additional hydraulic hardware required.
Overall, DCT's likely offer the greatest potential for effectiveness improvements
among the various transmission options presented in this report because they offer the
inherently lower losses of a manual transmission with the efficiency and shift quality
advantages of electronic controls. The lower losses stem from the elimination of the
conventional lock-up torque converter, and a greatly reduced need for high pressure
hydraulic circuits to hold clutches or bands to maintain gear ratios (in automatic
transmissions) or hold pulleys in position to maintain gear ratio (in Continuously
Variable Transmissions). However, the lack of a torque converter will affect how the
vehicle launches from rest, so a DCT will most likely be paired with an engine that offers
sufficient torque at low engine speeds to allow for adequate launch performance.
For the MY 2011 CAFE final rule, NHTSA estimated a 5.5 to 9.5 percent
improvement in fuel consumption over a baseline 4/5-speed automatic transmission for a
wet clutch DCT, which was assumed for all but the smallest of vehicle subclasses,
Subcompact and Compact cars and small LT. This results in an incremental effectiveness
estimate of 2.7 to 4.1 percent over a 6-speed automatic transmission with IATC. For
Subcompact and Compact Cars and small LT, which were assumed to use a dry clutch
DCT, NHTSA estimated an 8 to 13 percent fuel consumption improvement over a
baseline 4/5-speed automatic transmission, which equates to a 5.5 to 7.5 percent
incremental improvement over the 6-speed transmission. This is consistent with the EPA
staff report, which estimated that these transmissions offer an effectiveness of 9.5%-
14.5% over a 4-speed automatic transmission. EPA and NHTSA therefore conclude that
8 to 13 percent effectiveness is appropriate for this proposal.
In the MY 2011 CAFE final rule, the dry clutch DCT incremental cost was
previously estimated to be $68 for Subcompact and Compact cars relative to the 6-speed
automatic transmission. For larger vehicles such as Midsize and Large Car and Minivan
the wet clutch DCT incremental cost over 6-speed transmission was estimated at $218,
with both of these estimates assuming a 1.5 RPE. In the 2008 staff report, EPA estimated
the RPE (1.5 RPE factor) of a wet clutch DCT (called an AMT in the report) at $141
(2006$)
3-47
-------
Draft Joint Technical Support Document
The agencies revised these estimates to reflect updated sources and the ICM
markup of a medium complexity technology of 1.26. For this NPRM, the agencies
estimated MY 2012 costs of $74 for dry and $159 for wet DCT (2007$), which are only
slightly different from MY 2011 CAFE final rule. With time-based learning these costs
would be $65 and $139, respectively, for a MY 2016 vehicle (2007$).
3.4.2.2.6 Continuously Variable Transmission (CVT)
A Continuously Variable Transmission (CVT) is unique in that it does not use
gears to provide ratios for operation. Instead, the most common CVT design uses two V-
shaped pulleys connected by a metal belt. Each pulley is split in half and a hydraulic
actuator moves the pulley halves together or apart. This causes the belt to ride on either a
larger or smaller diameter section of the pulley which changes the effective ratio of the
input to the output shafts. Advantages of the CVT are that the engine can operate at its
most efficient speed-load point more of the time, since there are no fixed ratios.
However, CVT's are limited by engine power and cannot be applied to high torque
applications. Also, CVTs often have a wider range of ratios compared to conventional
automatic transmissions which can provide for more options in engine optimization.
While CVTs by definition are fully continuous, some automakers choose to emulate
conventional stepped automatic operation because some drivers are not used to the
sensation of the engine speed operating independently of vehicle speed.
Considering the confidential data together with independent review, NHTSA has
estimated the fuel consumption effectiveness for CVTs at 2.2 to 4.5 percent over a 4/5-
speed automatic transmission, which translates into a 0.7 to 2.0 incremental effectiveness
improvement over an automatic transmission with the IATC technology. EPA and
NHTSA reviewed these estimates and found them to be accurate.
The agencies adjusted the original estimates used in MY 2011 CAFE final rule to
account for medium complexity ICM of 1.25. This resulted in estimated incremental
compliance cost of $250 for the MY 2012 vehicles. In the Volpe model, this technology
was only applied to vehicles manufactured with unibody construction methods. CVTs
are an established and readily available technology so time based learning is applied.
3.4.2.2.7 6-Speed Manual Transmissions (6MAN)
Manual transmissions are entirely dependent upon driver input to shift gears: the
driver selects when to perform the shift and which gear to select. This is the most
efficient transfer of energy of all transmission layouts, because it has the lowest internal
gear losses, with a minimal hydraulic system, and the driver provides the energy to
actuate the clutch. From a systems viewpoint, however, vehicles with manual
3-48
-------
transmissions have the drawback that the driver may not always select the optimum gear
ratio for fuel economy. Nonetheless, increasing the number of available ratios in a
manual transmission can improve fuel economy by allowing the driver to select a ratio
that optimizes engine operation more often. Typically, this is achieved through adding
overdrive ratios to reduce engine speed at cruising velocities (which saves fuel through
reduced engine pumping losses) and pushing the torque required of the engine towards
the optimum level. However, if the gear ratio steps are not properly designed, this may
require the driver to change gears more often in city driving resulting in customer
dissatisfaction. Additionally, if gear ratios are selected to achieve improved launch
performance instead of to improve fuel economy, then no fuel saving effectiveness is
realized.
NHTSA's MY 2011 CAFE final rule estimated an effectiveness increase of 0.5
percent for replacing a 5-speed manual with a 6-speed manual transmission, which was
derived from confidential manufacturer data. NHTSA and EPA have found no evidence
to dispute this estimate and chosen to use 0.5 percent reduction in CC>2 emissions and fuel
consumption for replacing a 5-speed manual with a 6-speed manual transmission for this
proposal.
NHTSA updated costs to reflect the ICM low complexity markup of 1.11 which
resulted in a incremental compliance cost of $250 as compared to $338 for MY 2012,
with lower costs later MYs due to the application of time based learning factors. In the
2008 staff report, EPA estimated this RPE (1.5 RPE factor) at $107 (2006$) relative to a
5-speed manual transmission and for a MY 2015 vehicle. EPA updated this value to
2007$ and used a 1.11 ICM to get costs of $89 for a MY 2012 vehicle and $69 for a MY
2016 vehicle, both relative to a 5-speed manual transmission (2007$).
3.4.2.3 Hybrid and Electrification/Accessory Technologies
A Hybrid is a vehicle that combines two or more sources of propulsion energy,
where one uses a consumable fuel (like gasoline), and one is rechargeable (during
operation, or by another energy source). Hybrid technology is established in the U.S.
market and more manufacturers are adding hybrid models to their lineups. Hybrids
reduce fuel consumption through three major mechanisms:
The internal combustion engine can be optimized (through downsizing,
modifying the operating cycle, or other control techniques) to operate at or
near its most efficient point more of the time. Power loss from engine
downsizing can be mitigated by employing power assist from the secondary
power source.
Some of the energy normally lost as heat while braking can be captured and
stored in the energy storage system for later use.
3-49
-------
Draft Joint Technical Support Document
The engine is turned off when it is not needed, such as when the vehicle is
coasting or when stopped.
Hybrid vehicles utilize some combination of the three above mechanisms to
reduce fuel consumption and CC>2 emissions. A fourth mechanism to reduce petroleum
fuel consumption, available only to plug-in hybrids, is by substituting the petroleum fuel
energy with energy from another source, such as the electric grid. The effectiveness of
fuel consumption and COi reduction depends on the utilization of the above mechanisms
and how aggressively they are pursued. One area where this variation is particularly
prevalent is in the choice of engine size and its effect on balancing fuel economy and
performance. Some manufacturers choose not to downsize the engine when applying
hybrid technologies. In these cases, performance is vastly improved, while fuel
efficiency improves significantly less than if the engine was downsized to maintain the
same performance as the conventional version. While this approach has been used in
cars such as the Honda Accord Hybrid (now discontinued), it is more likely to be used for
vehicles like trucks where towing and/or hauling is an integral part of their performance
requirements. In these cases, if the engine is downsized, the battery can be quickly
drained during a long hill climb with a heavy load, leaving only a downsized engine to
carry the entire load. Because towing capability is currently a heavily-marketed truck
attribute, manufacturers are hesitant to offer a truck with downsized engine which can
lead to a significantly diminished towing performance with a low battery, and therefore
engines are traditionally not downsized for these vehicles.
Although hybrid vehicles using other energy storage concepts (flywheel,
hydraulic) have been developed, the systems currently in production in the U.S. for
passenger cars and light trucks use battery storage and electric drive systems. Hybrid
electric vehicles (HEV) are part of a continuum of vehicles using systems with differing
levels of electric drive and electric energy storage. This range of vehicles includes
relatively basic engine start/stop systems, HEV systems with varying degrees of electric
storage and electric drive system capability, plug-in hybrid electric vehicles (PHEV) with
differing degrees of all electric range and battery electric vehicles (EV) that rely entirely
on electric drive and battery electric energy storage.
Different HEV, PHEV and EV concepts utilize these mechanisms differently, so
they are treated separately for the purposes of this analysis. Below is a discussion of
battery energy storage and the major hybrid concepts that were determined to be
available in the near term.
3.4.2.3.1 Batteries for MHEV, HEV, PHEV and EV Applications
The design of battery secondary cells can vary considerably between MHEV,
HEV, PHEV and EV applications.
MHEV batteries: Due to their lower voltage (12-42 VDC) and reduced power and
energy requirements, MHEV systems will likely continue to use lead-acid batteries.
3-50
-------
MHEV battery designs differ from those of current starved-electrolyte (typical
maintenance free batteries) or flooded-electrolyte (the older style lead-acid batteries
requiring water "top-off) batteries used for starting, lighting and ignition (SLI) in
automotive applications. Standard SLI batteries are primarily designed to provide high-
current for engine start-up and then recharge immediately after startup via the vehicle's
charging system. Deeply discharging a standard SLI battery will greatly shorten its life.
MHEV applications are expected to use:
Extended-cycle-life flooded (ELF) lead-acid batteries
Absorptive glass matt, valve-regulated lead-acid (AGM/VRLA) batteries -
or-
Asymmetric lead-acid battery/capacitor hybrids (e.g. flooded
ultrabatteries)
MHEV systems using electrolytic double-layer capacitors are also under development.
HEV batteries: HEV applications operate in a narrow, short-cycling, charge-
sustaining state of charge (SOC). Energy capacity in HEV applications is somewhat
limited by the ability of the battery and power electronics to accept charge and by space
and weight constraints within the vehicle design. HEV battery designs tend to be
optimized for high power density rather than high energy density, with thinner cathode
and anode layers and more numerous current collectors and separators (Figure 3-1).
EV batteries: EV batteries tend to be optimized for high energy density and are
considerably larger than HEV batteries. PHEV battery designs are intermediate between
power-optimized HEV and energy-optimized EV battery cell designs.
PHEV batteries: PHEV batteries must provide both charge depleting operation
similar to an EV and charge sustaining operation similar to an HEV. Unlike HEV
applications, charge sustaining operation with PHEVs occurs at a relatively low battery
state of charge (SOC) which can pose a significant challenge with respect to attaining
acceptable battery cycle life. In the case of the GM Volt, this limits charge depleting
operation to a minimum SOC of approximately 30%.41 An alternative approach for
PHEV applications that has potential to allow extension of charge depletion to a lower
battery SOC is using energy-optimized lithium-ion batteries for charge depleting
operation in combination with the use of supercapacitors for charge sustaining
operation.42
3-51
-------
Draft Joint Technical Support Document
Figure 3-1: Schematic representation of power and energy optimized
prismatic-layered battery cells
Collector (-)
Cathode (-)
Separator
Anode (+)
Collector (+)
HEV Power-optimized Battery Cell
EV Energy-optimized Battery Cell
Power-split hybrid vehicles from Toyota, Ford and Nissan, integrated motor assist
hybrid vehicles from Honda and the GM 2-mode hybrid vehicles currently use nickel-
metal hydride (NiMH) batteries. Lithium-ion (Li-ion) batteries offer the potential to
approximately double both the energy and power density relative to current NiMH
batteries, enabling much more electrical-energy-intensive automotive applications such
as PHEVs and EVs. Li-ion batteries for high-volume automotive applications differ
substantially from those used in consumer electronics applications with respect to
cathode chemistry, construction and cell size. Li-ion battery designs currently under
development by CPI (LG-Chem) for the GM Volt PHEV and by AESC, GS-Yuasa and
A123 Systems (respectively) for the upcoming Nissan, Mitsubishi and Chrysler EVs use
large-format, layered-prismatic cells assembled into battery modules. The modules are
then combined into battery packs.
Cathodes for large-format, automotive Li-ion batteries are becoming increasingly
focused on two chemistries - LiMn2O4-spinel (CPI, GS-Yuasa, AESC) and LiFePO4
(A123 Systems).
In addition to the purely hybrid technologies, which decrease the proportion of
propulsion energy coming from the fuel by increasing the proportion of that energy
coming from electricity, there are other steps that can be taken to improve the efficiency
of auxiliary functions (e.g., power-assisted steering or air-conditioning) which also
reduce CO2 emissions and fuel consumption. These steps, together with the hybrid
technologies, are collectively referred to as "vehicle electrification" because they
3-52
-------
generally use electricity instead of engine power. In order to achieve consistency
between the two modeling techniques, and to improve the number and range of
technology offerings, the CAFE model was revised to include one additional mild hybrid
technology. The high voltage or improved efficiency alternator (HVIA) technology,
which was used in the 2011 rule, is no longer represented as a separate technology and
has instead been incorporated into this new mild hybrid technology, as discussed further
below.
3.4.2.3.2 Hybrid System Sizing and Cost Estimating Methodology
NHTSA and EPA reviewed estimates of cost and effectiveness for hybrid and
related electrical technologies and adjusted them as appropriate. Both agencies found the
hybrid technology cost estimating methodology that Ricardo and NHTSA developed
during the 2011 final rule to be reasonable and used it to estimate hybrid systems costs
and account for variation in component sizing across both the hybrid types and vehicle
subclasses. That method utilizes four pieces of data: (1) key component sizes for a
midsize car by hybrid system type; (2) normalized costs for each key component; (3)
component scaling factors that are applied to each vehicle class/subclass by hybrid
system type; and (4) vehicle characteristics for the subclasses which are used as the basis
for the scaling factors. During development of the methodology, NHTSA and Ricardo
made several assumptions:
1) Hybrid controls hardware varies with the level of functionality offered by the
hybrid technology. Assumed hybrid controls complexity for a 12V micro
hybrid (MHEV) and belt integrated starter generator (BISG) was 25 percent of
a strong hybrid controls system and the complexity for a Crank Integrated
Starter Generator (CISG) was 50 percent. These ratios were estimates based
on the directional need for increased functionality as system complexity
increases.
2) Li-Ion batteries for hybrid electric vehicles are currently entering production,
including a 2010 MY Mercedes and Hyundai. One estimate from Anderman
indicates that Li-ion market penetration will achieve 35 percent by 2015.43
However, as was discussed above, significant development effort is underway
by a number of battery producers which could impact cost and overcome other
technical concerns. Therefore it was assumed that mild hybrids (MHEV,
BISG, CISG) and strong hybrids (PSHEV, 2MHEV, PHEV) will use either
Li-Ion or NiMH batteries, depending on cost considerations. However, plug-
in hybrids will use Li-ion batteries only. Battery usage is discussed further
below.
3) The plug-in hybrid battery pack was sized for a mid-sized car by assuming:
the vehicle has a 20 mile all electric range and consumes an average of 300
W-hr per mile; the battery pack can be discharged down to 30 percent depth of
3-53
-------
Draft Joint Technical Support Document
discharge;1 and the capacity of a new battery pack is 20 percent greater than at
end of life (i.e., range on a new battery pack is 24 miles).
4) All hybrid systems included a DC/DC converter which was sized to
accommodate vehicle electrical loads appropriate for increased vehicle
electrification in the time frame considered.
5) High voltage wiring scaled with hybrid vehicle functionality and could be
represented as a fraction of strong hybrid wiring. These ratios were estimates
based on the directional need for increased functionality as system complexity
increases.
6) All hybrid systems included a supplemental heater to provide vehicle heating
when the engine is stopped; however, in this proposal, it is assumed that only
half of the vehicles will adapt this technology, as discussed further below.
Only the strong hybrids included electric air conditioning to enable engine
stop/start when vehicle air conditioning was requested by the operator.
Furthermore, NHTSA and Ricardo recognized that some strong hybrid systems
replaced a conventional transmission with a hybrid-specific transmission, resulting in a
cost offset (i.e., a cost credit) for the removal of a portion of the clutches and gear sets
within the transmission. In the MY 2011 rule, the transmission cost in Table 3-11 below
expressed hybrid transmission costs as a percentage of traditional automatic transmission
cost, as described in the 2008 Martec Report, at $850 direct manufacturing costs (non-
RPE/ICM). The method assumed that the mechanical aspect of a power-split
transmission with a reduced number of gear sets and clutches resulted in a cost savings of
50 percent ($425) over a conventional transmission with torque converter. For a 2-mode
hybrid, the mechanical aspects of the transmission are similar in complexity to a
conventional transmission, so no cost savings was appropriate. The plug-in hybrid
assumed a highly simplified transmission for electric motor drive, thus 25 percent of the
base vehicle transmission cost was applied (resulting in a $638 credit).
The NHTSA MY 2011 CAFE final rule discusses in detail how the hybrid cost
estimating methodology uses the information provided in the tables below to calculate
costs for each of the strong hybrid systems used in this proposal. It also includes a step-
by-step example for the midsize vehicle mild hybrid systems used in the MY 2011 CAFE
final rule.44 As in that analysis, it is important to understand that the CISG technology
replaces existing mild hybrid systems."1
NHTSA and EPA in reviewing the above made the following revisions. First,
NHTSA and EPA revalidated the component sizes that were estimated for a midsize car
for each type of hybrid system as shown in Table 3-11. However, NHTSA and EPA
added an additional component- front engine accessory drive (FEAD), because
hybridization often involves revision to the FEAD design such that certain devices (belts,
pulleys, idlers, etc) as well as other engine components (alternator, A/C compressor, and
1 The GM Volt operates between 30% DOD and 85% DOD. So there is 55% useable DOD, but charge
sustaining operation starts at 30% and cycles between 30 and 35% DOD.
m For the incremental CAFE model, before CISG is applied, the costs for MHEV and BISG are subtracted
if they were previously applied.
3-54
-------
starter) may no longer be needed and can thus be eliminated, or may be de-specified to
lower cost alternatives. This is applicable to CISG and the strong hybrid technologies,
and is intended to account for cost savings associated with items that changed or are no
longer required as a result of these technology applications.
Table 3-11. Component Sizes by Hybrid Type for a Midsize Car
Component
Primary Motor power, continuous (kW)
Secondary Motor power, continuous (kW)
Primary Inverter power, continuous (kW)
Secondary Inverter power, continuous (kW)
Controls complexity (relative to strong hybrid)
NiMH Battery Pack capacity (kW-hr)1
Li-Ion Battery Pack capacity (kW-hr)1
DC/DC Converter power (kW)
High Voltage Wiring (relative to strong hybrid)
Supplemental heating2
Mechanical Transmission (relative to baseline vehicle)
Electric AC
Blended Brakes
FEAD Credit
Charger power, continuous (kW)
Hybrid Type
MHEV
BISG
3
na
3
na
25%
na
na
0.7
na
50%
100%
No
No
No
na
CISG
11
na
11
na
50%
1
1
2
50%
50%
100%
No
Yes
Yes
na
PSHEV
45
30
45
30
100%
2
2
2
100%
50%
50%
Yes
Yes
Yes
na
2MHEV
45
45
45
45
100%
2
2
2
100%
50%
100%
Yes
Yes
Yes
na
PHEV
45
30
45
30
100%
na
15
2
100%
50%
25%
Yes
Yes
Yes
3
1 - Assumes the use of either NiMH or Li-Ion, and not both.
2 - Implemented through a reduction in component cost (50%)
Second, the costs estimates of the key components were revised. The MY 2011
CAFE final rule was developed at a time when economic conditions were significantly
different than those that currently exist, a time when many of the commodity materials
used in the hybrid systems were more expensive than today. These changes in economic
conditions were one of the factors leading to some of the cost revisions EPA and NHTSA
jointly discussed and made. Differences in estimates provided by confidential sources to
either EPA or NHTSA also played a part in the revisions. In addition, the agencies
applied the new ICM mark-up factors instead of the RPE that was used previously. An
appropriate ICM factor (1.45 for most mild and strong hybrid technologies) replaces the
previous RPE factor (1.5), as discussed in section 3.3.2.2.2. Specifically, the primary and
secondary inverter cost per kilowatt were revised downward from $10 to $7, the controls
cost was revised upward from $100 to $115, the DC/DC converter costs were revised
from $100 to $88, the blended brake system that was revised from $400 to $310, and
finally the fully learned, high volume production, cost per kilowatt hour (kW-hr) for
Nickel Metal Hydride (NiMH) battery was revised from $350 to $320 to match the
updated costs for Lithium Ion batteries discussed below.n
n In the MY 2011 final rule NHTSA had misstated this $350 per kW-hr cost as $50 per kW-hr due to a
typographical error.
3-55
-------
Draft Joint Technical Support Document
The cost for Lithium Ion (Li-Ion) batteries was also revised. As previously stated,
Li-Ion batteries are being implemented in series production in model year 2010. Battery
technology is changing rapidly in the marketplace today, as discussed above, and is
expected to continue along this path throughout the rulemaking period. OEMs are now
forming relationships with battery manufacturers in an effort to research and develop not
only new and improved battery technology, but also more efficient manufacturing
processes capable of supporting high volume production. Accordingly, as shown in
Table 3-12, the $600 per kW-hr used in the 2011 rule was revised downward to $320 per
kW-hr. The revision downward from $600/kW-hr in the 2011 CAFE final rule to
$320/kW-hr in this analysis was done based a study by Deutsche Bank that estimated Li-
Ion battery costs at 300-400 /kW-hr.45 This was converted to $500/kW-hr then learned
twice using volume-based learning to arrive at the $320/kW-hr applicable in the 2015
timeframe.
Li-ion batteries were originally restricted to plug-in hybrids only. Recent vehicle
introductions confirm either battery technology can be used in any mild or strong HEV
application. However, manufacturers are likely to consider cost highly in their selection
of battery technology. If Li-ion battery prices remain high, NiMH would be the default
battery technology for all hybrid electric vehicles. For plug-in hybrids Li-ion would
continue to be required because plug-in hybrids demand higher energy density than
NiMH can provide. Neither the CAFE nor OMEGA model predicts a high penetration of
plug-in technology in achieving the proposed standards.
Finally, the agencies assessed the cost savings associated with the FEAD credit
discussed above. This cost was not previously represented in the hybrid cost model. As
shown in Table 3-12 below, a $100 credit is proposed which offsets directly the costs of
the other components specified. This is the best approximation of the value of these
items, based on NHTSA and EPA engineering assessment.
Estimates of each key component are shown in Table 3-12 below along with the
sources of those estimates. The cost basis estimates assume fully learned, high-volume
(greater than 1.2 million units per annum) production, and the costs shown are direct
manufacturing costs that are not RPE or ICM adjusted. This table does not show a cost
applicable to the belt integrated starter generator system (BISG) since it is a fixed cost
that, like the automatic transmission pump cost, is not scaled by subclass as described
later.
3-56
-------
Table 3-12. Component Cost Basis at High Volumes and Data Sources
COMPONENT
Primary Motor ($/kW)
Secondary Motor ($/kW)
Primary Inverter ($/kW)
Secondary Inverter ($/kW)
Controls
NiMH Battery Pack ($/kW-hr.)
Li-ion Battery Pack ($/kW-hr.)
DC/DC Converter (Size: 2kW)
High Voltage Wiring
Supplemental Heating
Mechanical Transmission
Electric Air Conditioning
Blended Brakes
Charger
Automatic Transmission Pump
FEAD Credit
COST BASIS
$15
$15
$7
$7
$115
$320
$320
$88
$200
$42
$850
$450
$310
$100
$75
$(100)
DATA SOURCE
Martec 2008
Confidential Business Information
2011CAFEFRM
(with revision)
Deutsche Bank 2008
Confidential Business Information
Martec 2008
Martec 2008 (to 4-spd auto)
Confidential Business Information
Martec 2008
Confidential Business Information
Third, NHTSA and EPA also revised component size/scaling assumptions for
some vehicles (i.e., large trucks). NHTSA and EPA recognized that some manufacturers
may choose not to use supplemental cabin heating opting instead to continue engine
operation in the event heat demand occurs; therefore supplemental heating is specified for
only half of the vehicles. Table 3-12 above indicates the 50 percent application rate
implemented in the hybrid cost estimating methodology reducing the component cost
from $84 to $42.
EPA and NHTSA also reviewed the choice of a 3 kW DC/DC converter as a
component size input for a midsize vehicle, which represented a 250 amp current
capability. In retrospect this is a high specification for a midsize vehicle and revised the
estimate to a 2 kW DC/DC converter, as shown in Table 3-12 above, which would
represent a more reasonable 150 amp current capacity.
The scaling factor used for the primary and secondary motors and inverters on the
large truck and SUV vehicles was revised. As in the MY 2011 CAFE final rule, a linear
extrapolation was used from the midsize vehicle and extended it out to the largest of
vehicles, the large truck class. This resulted in projected component sizes that are larger
than those used on a commercially realized truck in this vehicle class, the Chevrolet
Tahoe two-mode HEV. Accordingly the scaling factors have been revised for this class
(and the agencies have verified scaling factors for the other classes). This more closely
approximates the motor and inverter sizes specified in the Tahoe application. For future
3-57
-------
Draft Joint Technical Support Document
analysis, the agencies are considering whether it may be more accurate to use one set of
scaling for passenger cars and another different set for light trucks.
Another revision involves the addition of a stand-alone higher voltage Start-Stop/
BISG mild hybrid system. NHTSA and EPA determined that by applying a cost increase
to the MHEV technology to allow for a voltage increase (lead acid batteries) and
efficiency improvements to the alternator, the system would then approximate the higher
voltage Start-Stop /BISG applied by EPA. Based on confidential sources, the estimates
provided were first converted to 2007 dollars and then reverse learned through two
cycles, since volume learning is applicable, to arrive at a non-RPE/ICM incremental
compliance cost to be $229. This cost is applicable to all classes that use higher voltage
Start-Stop/BISG and is not scaled by any vehicle attribute.
Component scaling factors for each type of hybrid system are as shown in Table
3-13 below.
Table 3-13. Component Scaling Factors applied to Vehicle Class for each Hybrid System
Component
Primary Motor
Secondary Motor
Primary Inverter
Secondary Inverter
Controls
NiMH Battery Pack
Li-Ion Battery Pack
DC/DC Converter
High Voltage Wiring
Supplemental heating
Mechanical Transmission
Electric AC
Blended Brakes
Charger
Hybrid Type
MHEV
Engine
displacement
na
CISC
Curb
weight
na
PSHEV
Curb
2MHEV
weight1
Engine
displacement
PHEV
Engine power
Curb weight
Primary motor power
na
na
Secondary motor power
Complexity
na
na
Curb weight
na
Curb weight
Curb weight3
na
Vehicle footprint
Vehicle footprint
Same for all vehicle classes
na
na
na
na
Vehicle footprint
Same for all vehicle classes
na
na
na
Same for all
vehicle classes
(1)
(2)
(3)
For all vehicle classes except for performance classes which use Engine Torque
Curb weight used as surrogate for vehicle road load
Curb weight used as surrogate for vehicle electrical load
3-58
-------
Regarding the market data file from the MY 2011 CAFE final rale, NHTSA and
EPA did not make any revisions to the average vehicle characteristics for each vehicle
subclass as shown in Table 3-14, which defines the average vehicle characteristics for
each vehicle subclass. These characteristics were used as the basis of the scaling factors
in the CAFE and OMEGA models. EPA used these vehicle specifications solely for the
purpose of scaling hybrid vehicle attributes in order to remain consistent with the
NHTSA methodology.
Table 3-14. Key Vehicle Characteristics For Each Vehicle Subclass for CAFE Model
Vehicle Subclass
Sub com pact Car
Compact Car
Midsize Car
Large Car
Performance Subcompact Car
Performance Compact Car
Performance Midsize Car
Performance Large Car
Minivan
Small Truck
Medium Truck
Large Truck
Curb
Weight
(Ibs)
2795
3359
3725
4110
3054
3516
3822
4189
4090
3413
4260
5366
Footprint
(ft2)
41
44
47
50
40
44
47
51
50
45
50
63
Engine
Disp. (L)
1.9
2.2
2.9
3.4
2.7
3.0
3.9
4.8
3.3
2.6
3.6
5.0
Engine
Power
(hp)
134
166
205
258
260
269
337
394
247
178
250
323
Torque
(ft-lb)
133
167
206
248
260
260
318
388
242
185
256
352
3.4.2.3.3 Electrical Power Steering (EPS)
Electric power steering (EPS) provides a potential reduction in COi emissions and
fuel consumption over hydraulic power steering because of reduced overall accessory
loads. This eliminates the parasitic losses associated with belt-driven power steering
pumps which consistently draw load from the engine to pump hydraulic fluid through the
steering actuation systems even when the wheels are not being turned. EPS is an enabler
for all vehicle hybridization technologies since it provides power steering when the
engine is off. EPS may be implemented on most vehicles with a standard 12V system.
Some heavier vehicles may require a higher voltage system which may add cost and
complexity.
In the 2011 final rule NHTSA estimated a 1 to 2 percent effectiveness based on
the 2002 NAS report, a Sierra Research report, and confidential manufacturer data.
3-59
-------
Draft Joint Technical Support Document
NHTSA and EPA reviewed these effectiveness estimates and found them to be accurate,
thus they have been retained for this proposal.
Regarding costs, in the MY 2011 CAFE final rule, NHTSA estimated EPS at
$105-120 at 1.5 RPE mark-up factor. NHTSA and EPA adjusted the EPS cost for the
current proposal based on a review of the specification of the system. Adjustments were
made to include potentially higher voltage or heavier duty system operation, such as
would be required on some hybrid trucks. Accordingly, higher costs were estimated for
systems with higher capability. After accounting for the differences in system capability
and applying the ICM markup of low complexity technology of 1.11, the estimated costs
for this rulemaking are $106 for a MY 2012 vehicle (2007$). With time-based learning,
this cost would be $94 for a MY 2016 vehicle (2007$). As EPS systems are in wide
spread usage today, time based learning is deemed applicable.
3.4.2.3.4 Improved Accessories (IACC)
The accessories on an engine, including the alternator, coolant and oil pumps are
traditionally mechanically driven. A reduction in COi emissions and fuel consumption
can be realized by driving them electrically, and only when needed ("on-demand").
Electric water pumps and electric fans can provide better control of engine
cooling. For example, coolant flow from an electric water pump can be reduced and the
radiator fan can be shut off during engine warm-up or cold ambient temperature
conditions which will reduce warm-up time, reduce warm-up fuel enrichment, and reduce
parasitic losses.
Indirect benefit may be obtained by reducing the flow from the water pump
electrically during the engine warm-up period, allowing the engine to heat more rapidly
and thereby reducing the fuel enrichment needed during cold starting of the engine.
Further benefit may be obtained when electrification is combined with an improved,
higher efficiency engine alternator. Intelligent cooling can more easily be applied to
vehicles that do not typically carry heavy payloads, so larger vehicles with towing
capacity present a challenge, as these vehicles have high cooling fan loads.0
NHTSA and EPA jointly reviewed the estimates of 1 to 2 percent effectiveness
estimates used in MY 2011 rule and found them to be accurate for Improved Electrical
Accessories. NHTSA and EPA also reevaluated the assumptions from the MY 2011
CAFE final rule. For its analysis, NHTSA simply adjusted the 2011 CAFE final rule
costs to reflect the new ICM markup of low complexity value of 1.11, resulting in a cost
estimate for this rulemaking of $128 at MY 2012 (2007$). Improved accessory systems
0 In the CAFE model, IACC refers solely to improved engine cooling however EPA has included a high
efficiency alternator in this category, as well as improvements to the cooling system.
3-60
-------
are in production currently and thus timer based learning is applied. For its analysis,
EPA considered both an accessory upgrade to a 12 Volt system and a combined
accessory upgrade with conversion to a 42 Volt system. The accessory upgrade itself is
estimated to cost $86 in MY 2012 and $76 in MY 2016 (with time-based learning; both
in 2007$). The conversion to a 42 Volt electrical system is estimated to cost $97 for a
MY 2012 vehicle and $86 for a MY 2016 vehicle (again with time-based learning; both
in 2007$).
3.4.2.3.5 12VMicro Hybrid (MHEV)
12V Micro-Hybrid (MHEV) systems are the most basic of hybrid systems and
offer only the ability to turn the engine off when the vehicle is stopped or potentially
during deceleration (i.e. idle stop). Their low cost and adaptability to existing
powertrains and platforms can make them attractive for some applications. The
conventional belt-driven alternator is replaced with a belt-driven, enhanced power
starter-alternator and a redesigned front-end accessory drive system. A conventional
12V gear-reduction starter is retained to ensure reliable cold-weather starting. Also,
during idle-stop, some functions such as power steering and automatic transmission
hydraulic pressure are lost; so electric power steering and an auxiliary transmission
pump are needed.
Although EPA doesn't use the 12 V version of this stop-start technology, NHTSA
and EPA jointly reviewed the assumptions. The effectiveness estimates for this
technology ranged from 2.0 to 4.0 percent dependent on whether the vehicle is equipped
with a 4, 6 or 8 cylinder engine, with the 4 cylinder engine having the lowest range and
the 8 cylinder having the highest. The estimates reflect the limited capability of 12 volt
systems; sources citing higher estimates typically involve higher voltage systems that
have increased capability.
For this proposal, the system specifications assumed in the 2011 CAFE final
rule46 were applied (i.e., use of a 3 kW motor and a DC/DC converter) and the hybrid
technology cost method produced costs like those in 2011 rule; however, the use of new
ICM markups resulted in costs ranging from $288 for Subcompact subclass to $410 for
the Large Performance subclass. This technology is not applied to the Large Truck and
SUV subclass due to the heavy duty utility requirements of these vehicles. Time based
learning is considered applicable, and a schematic of the MHEV system is shown in
Figure 3-2.
Figure 3-2 Schematic of MHEV type System [Husted, 2003]
3-61
-------
Draft Joint Technical Support Document
Motor/
Generator
Controller
Inverter
Battery
Ending
/Trans.
Final Drive
Wheel
O
3.4.2.3.6 Higher Voltage Stop-Start/Belt Integrated Starter Generator (BISG)
Higher Voltage Stop-Start and Belt Mounted Integrated Starter Generator
(BISG) systems are similar to a micro-hybrid system, offering idle-stop functionality
except that they utilize larger electric machine and a higher capacity battery, typically
42 volts or above, thus enabling a limited level of regenerative braking which is
generally not applicable to 12 volt systems. The larger electric machine and battery
also enables a limited degree of power assist, which MHEV cannot provide. However,
because of the limited torque capacity of the belt driven design, these systems have a
smaller electric machine, and thus less capability than crank integrated or stronger
hybrid systems. These systems replace the conventional alternator with a belt-driven
starter/alternator and may add electric power steering and an auxiliary automatic
transmission pump. The limited electrical requirements of these systems allow the use
of lead-acid batteries or supercapacitors for energy storage.
The effectiveness estimates used for these technologies range from 3.0 to 7.5
percent dependent on vehicle subclass. The CAFE model, which applies this
effectiveness incrementally to the prior 12 Volt technology, uses estimates of 4 to 6
percent. The estimates are based on those found in the 2002 NAS and 2004 NESCAFE
reports, and confidential manufacturer data.
EPA estimates an incremental compliance cost range of $549 (small car) to $682
(large truck) for a MY 2012 vehicle and including a medium complexity ICM of 1.25
(2007$). With volume-based learning applied, these become $351 (small car) and $437
(large truck) for a MY 2016 vehicle (2007$). The cost estimate in the CAFE model is
incremental to the 12 Volt micro hybrid systems as just noted and therefore is adjusted
upwards to $286 to reflect the additional battery capacity, wiring upgrades, and a larger
optimized electric machine only, as discussed above. The $286 reflects volume based
learning factors and the ICM medium complexity markup of 1.25.
3-62
-------
3.4.2.3.7 Integrated Motor Assist (IMA)/Crank Integrated Starter Generator (CISC)
IMA is a system developed and marketed by Honda47 and is similar to CISG.
They both utilize a thin axial electric motor bolted to the engine's crankshaft and
connected to the transmission through a torque converter or clutch. The axial motor is
motor/generator that typically operates above 100 volts (but lower than the stronger
hybrid systems discussed below, which typically operate at around 300 volts) and can
provide sufficient torque for launch as well as generate sufficient current to provide
significant levels of brake energy recovery. The motor/generator also acts as the starter
for the engine and can replace a typical accessory-driven alternator. Current IMA/CISG
systems typically do not launch the vehicle on electric power alone, although some
commercially available systems can cruise on electric power and dual-clutch
IMA/CISG systems capable of all-electric drive are under development. IMA and
CISG could be applied to all classes of vehicles. A schematic of the Honda IMA
system is shown in Figure 3-3.
Figure 3-3 Schematic of Honda IMA System [Husted, 2003]
o
[
Moior/
(jeneralor
ins.
NHTSA did not have an equivalent technology to CISG in the 2011 final rule (the
ISO technology used in the 2011 rule was envisioned to be less capable than the CISG
technology defined here). For the CISG technology NHTSA estimated a net
effectiveness range of 16 to 20 percent, relative to the baseline vehicle and across all
vehicle subclasses in this analysis. The CAFE model therefore applies an incremental
effectiveness of approximately 8.6 to 8.9 percent relative to the BISG technology and
dependent on vehicle subclass except for large truck and SUV in which case BISG does
not apply. Note that the net effectiveness assumptions used in this proposal do not
include engine downsizing, or any other effectiveness gains from engine, transmission, or
vehicle technologies added to the vehicle by the modeling process.
EPA relied on a combination of certification data (comparing vehicles available
with and without a hybrid system and backing out other components where appropriate)
and manufacturer-supplied information to determine that the effectiveness of these
3-63
-------
Draft Joint Technical Support Document
systems in terms of CCh reduction is 30% for small cars, 25% for large cars, and 20% for
minivans and small trucks similar to the range estimated by NHTSA for the respective
vehicle classes.. The effectiveness for small cars assumes engine downsizing to maintain
approximately equivalent performance. The large car, minivan, and small truck
effectiveness values assume less engine downsizing in order to improve vehicle
performance and/or maintain towing and hauling performance.
EPA has estimated these costs, relative to a base engine/transmission, as $4,459
(small car) to $6,924 (large truck) for a MY 2012 vehicle and including a high
complexity ICM of 1.45 (2007$). With volume-based learning, these costs become
$2,854 and $4,431, respectively, for a MY 2016 vehicle (2007$). NHTSA has estimated
the MY 2012 compliance cost, relative to a BISG baseline, to range from $2,791 for
Subcompact vehicles to $3,611 for Midsize Light Truck and Minivan vehicles. These
include a high complexity ICM markup factor of 1.45 for this technology. As this
technology is still in limited production use, volume based learning is applied, resulting
in lower costs in later years.
3.4.2.3.8 Power Split Hybrid
The Power Split hybrid (PSHEV) has the ability to move the vehicle on electric
power only. It replaces the vehicle's transmission with a single planetary gear and a
motor/generator. A second, more powerful motor/generator is directly connected to the
vehicle's final drive. The planetary gear splits engine power between the first
motor/generator and the final drive. The first motor/generator uses power from the
engine to either charge the battery or supply power to the wheels. The speed of the first
motor/generator determines the relative speed of the engine to the wheels. In this way,
the planetary gear allows the engine to operate independently of vehicle speed, much like
a CVT. The Toyota Prius and the Ford Hybrid Escape are two examples of power split
hybrid vehicles.
In addition to providing the functions of idle engine stop, subsequent restart and
regenerative braking, this hybrid system allows for pure EV operation. The power split
system provides very good fuel consumption in city driving. During highway cycles, the
hybrid functions of regenerative braking, engine start/stop and optimal engine operation
cannot be applied as often as in city driving, and so the effectiveness in fuel consumption
is less. Additionally, it is less efficient at highway speeds due to the fact that the first
motor/generator must be spinning at a relatively high speed and therefore incurs losses.
Newer designs incorporate a gear-reduction motor to provide improved high speed
efficiency and improved matching of motor torque to engine torque.
The Power Split hybrid also reduces the cost of the transmission, replacing a
conventional multi-speed unit with a single planetary gear. The electric components are
bigger than those in mild hybrid and IMA/CISG configurations so the costs are
correspondingly higher.
3-64
-------
EPA and NHTSA used a combination of manufacturer-supplied information and a
comparison of vehicles available with and without a hybrid system from EPA's fuel
economy test data to determine that the effectiveness is 19-36% for the classes to which it
is applied. The estimate would depend on whether engine downsizing is also assumed.
As an example, See Table 3-15 and Table 3-16, which assume some level of engine
downsizing. In the CAFE incremental model, the range of effectiveness used was 23 to
33 percent as engine downsizing is not assumed (and accounted for elsewhere).
Table 3-15 Large Car Power Split Certification Data
Nissan Altima
3.5LCVT
HEV 2.5L PS
Net % difference
Toyota Camry
3.0L5-auto
HEV 2.4L PS
Net % difference
Lexus GS
4.3L 6-auto
HEV 3. 5L PS
Net % difference
Tailpipe CO2
City
444
317
404
222
493
355
Hwy
306
254
286
234
355
317
55/45 comb.
386
286
-26%
355
228
-36%
423
341
-19%
Table 3-16 Small Truck Power Split Certification Data
Ford Escape 4X4
3.0L4-auto
HEV 2.3L PS
Net % difference
Ford Escape 4X2
3.0L4-auto
HEV 2.3L PS
Net % difference
Toyota Highlander 4X4
3.3L5-auto
HEV 3. 3L PS
Net % difference
Tailpipe CO2
City
467
277
444
247
493
286
Hwy
386
306
370
286
370
329
55/45 comb.
423
286
-32%
404
261
-35%
423
306
-28%
Using the hybrid cost estimating methodology with updated component costs,
EPA established a compliance cost range of $4,481 (small cars) to $8,145 (large trucks)
relative to the base engine and transmission for a MY 2012 vehicle (all in 2007$). With
time-based learning, these costs become $3,967 and $7,210, respectively, for a MY 2016
vehicle (2007$). Using the same methodology, NHTSA established overall PSHEV
system costs with electric power steering (EPS) and improved accessories (IACC)
3-65
-------
Draft Joint Technical Support Document
ranging from $5,509 to $11,534 for MY 2012, which leads to incremental costs in the
CAFE model ranging from $1,600 to $6,723 depending on the subclass (all in 2007$).
The estimates were determined with a 1.45 ICM for a high complexity technology.
Time-based learning is applicable to the power split technology.
3.4.2.3.9 2-Mode Hybrid
The 2-Mode Hybrid (2MHEV) uses an adaptation of a conventional stepped-ratio
automatic transmission which replaces some of the transmission clutches with two
electric motor/generators allowing the transmission to act like a CVT. The
motor/generators control the ratio of engine speed to vehicle speed. The clutches allow
the motors to be bypassed improving the transmission's torque capacity and the
efficiency for improved fuel economy at highway speeds. This type of system is used in
the Chevrolet Tahoe Hybrid.
In addition to providing the hybrid functions of engine stop and subsequent restart
and regenerative braking, the 2MHEV allows for pure EV operation. The two
motor/generators allow the engine to be run in efficient operating zones. The 2-mode
system provides significantly reduced COi emissions and fuel consumption. The primary
motor/generator is comparable in size to that in the PSHEV system, but the secondary
motor/generator is larger. The 2-mode system cost is greater than that for the power split
system due to the additional transmission complexity and secondary motor sizing.
The battery pack for 2MHEV is assumed to be 300V NiMH as is used in current
2MHEV systems today however GM recently announced an agreement with Hitachi to
provide Li-ion battery packs for 2012 and later 2MHEVs.
For this proposal the CAFE model considered a range of 23 to 33 percent with a
midpoint of 28 percent, assuming no engine downsizing to preserve the utility nature of
medium and large trucks (e.g., maintaining full towing capability even in situations with
low battery charge) and EPA estimates COi emissions reduction effectiveness to be 25
percent for large trucks (LDT3 and LDT4 categories) based on vehicle certification data.
EPA estimates an effectiveness of 40 percent for smaller vehicles.
EPA and NHTSA estimated MY 2012 costs using the updated component costs
and scaling factors in the hybrid cost estimating methodology discussed above. EPA
estimated the compliance cost to range from $5,469 for large car to $7,236 for large
truck/SUV relative to the base engine/transmission (all in 2007$). NHTSA determined
incremental cost estimates ranging from $3,521 to $5,779 for light-duty truck
applications in the CAFE model. These include the 1.45 ICM markup value for high
complexity technology. Given that the CAFE and OMEGA models classify vehicles
differently, these numbers varied slightly in the modeling to reflect the different
classification methods of the OMEGA and CAFE models, as described in the individual
NHTSA and EPA PRIAs. Volume based learning is applicable to this technology.
3-66
-------
3.4.2.3.10 Plug-In Hybrid
Plug-In Hybrid Electric Vehicles (PHEVs) are very similar to Hybrid Electric
Vehicles, but with three significant functional differences. The first is the addition of a
means to charge the battery pack from an outside source of electricity (e.g. the electric
grid). Second, a PHEV would have a larger battery pack with more energy storage, and a
greater capability to be discharged. Finally, a PHEV would have a control system that
allows the battery pack to be significantly depleted during normal operation.
Table 3-17 below, illustrates how PHEVs compare functionally to both hybrid
electric vehicles (HEV) and electric vehicles (EV). These characteristics can change
significantly within each class/subclass, so this is simply meant as an illustration of the
general characteristics. In reality, the design options are so varied that all these vehicles
exist on a continuum with HEVs on one end and EVs on the other.
Table 3-17 Conventional, HEVs, PHEVs, and EVs Compared
Attribute
Drive Power
Engine Size
Electric Range
Battery Charging
Increasing Electrification
Conventional
Engine
Full Size
None
None
HEV
Blended
Engine/Electric
Full Size or Smaller
None to Very Short
On-Board
PHEV
Blended
Engine/Electric
Smaller or Much
Smaller
Short to Medium
Grid/On-Board
EV
Electric
No Engine
Medium to Long
Grid Only
Deriving some of their propulsion energy from the electric grid provides several
advantages for PHEVs. PHEVs offer a significant opportunity to replace petroleum used
for transportation energy with domestically-produced electricity. The reduction in
petroleum usage does, of course, depend on the amount of electric drive the vehicle is
capable of under its duty cycle. PHEVs also provide electric utilities the possibility to
increase electric generation during "off-peak" periods overnight when there is excess
generation capacity and electricity prices are lower. Utilities like to increase this "base
load" because it increases overall system efficiency and lowers average costs. Utilities
are also investigating the use of PHEV and EV batteries as a source of grid storage
capacity to provide ancillary services for grid stabilization purposes. PHEVs can lower
localized emissions of criteria pollutants and air toxics especially in urban areas by
operating on electric power. The emissions from the power generation occur outside the
urban area at the power generation plant which provides health benefits for residents of
the more densely populated urban areas by moving emissions of ozone precursors out of
the urban air shed. Unlike most other alternative fuel technologies, PHEVs can initially
use an existing infrastructure for refueling (charging and liquid refueling) so investments
in infrastructure may be reduced.
3-67
-------
Draft Joint Technical Support Document
In analyzing the impacts of grid-connected vehicles like PHEVs and EVs, the
emissions from the electrical generation can be accounted for if a full upstream and
downstream analysis is desired. While this issue is being studied on an on-going basis,
upstream CC>2 emissions are not unique to grid-connected technologies and so are not
included in this analysis.
PHEVs will be considerably more costly than conventional vehicles and some
other advanced technologies. To take advantage of their capability, consumers would
have to be willing to charge the vehicles nightly, and would need access to electric power
where they park their vehicles. For many urban dwellers who may park on the street, or
in private or public lots or garages, charging may not be practical. Charging may be
possible at an owner's place of work, but that would increase grid loading during peak
hours which would eliminate some of the benefits to utilities of off-peak charging vs. on-
peak. Oil savings will still be the same in this case assuming the vehicle can be charged
fully.
The effectiveness potential of PHEVs depends on many factors, the most
important being the energy storage capacity designed into the battery pack. To estimate
the fuel consumption and tailpipe COi reduction potential of PHEVs, EPA has developed
an in-house vehicle energy model (PEREGRIN) to estimate the fuel consumption/CO2
emissions reductions of PHEVs. This model is based on the PERE (Physical Emission
Rate Estimator) physics-based model used as a fuel consumption input for EPA's
MOVES mobile source emissions model.
The PHEV small car, large car, minivan and small trucks were modeled using
parameters from a midsize car similar to today's hybrids and scaled to each vehicle's
weight. The large truck PHEV was modeled separately assuming no engine downsizing.
PHEVs can have a wide variation in the All Electric Range (AER) that they offer. Some
PHEVs are of the "blended" type where the engine is on during most of the vehicle
operation, but the proportion of electric energy that is used to propel the vehicle is
significantly higher than that used in a PSHEV or 2MHEV. Each PHEV was modeled
with enough battery capacity for a 20-mile-equivalent AER and a power requirement to
provide similar performance to a hybrid vehicle. 20 miles was selected because it offers
a good compromise for vehicle performance, weight, battery packaging and cost. Given
expected near-term battery capability, a 20 mile range represents the likely capability that
will be seen in PHEVs in the near-to-mid term.
To calculate the total energy use of a PHEV, the PHEV can be thought of as
operating in two distinct modes, electric (EV) mode, and hybrid (HEV) mode. At the
tailpipe, the COi emissions during EV operation are zero. The EV mode fuel economy
can then be combined with the HEV mode fuel economy using the Utility Factor
calculation in SAE J1711 to determine a total MPG value for the vehicle. (See Table
3-18)
3-68
-------
Table 3-18 Sample Calculation of PHEV Gasoline-Equivalent CO2 Reduction
EV energy comb (0.55 city / 0.45 hwy)
EV range (from PEREGRIN)
SAEJ1711 utility factor
HEV mode comb FE (0.55 city / 0.45 hwy)
Total UF-adjusted FE (UF*FCEV + (1-UF)*FCHEV)
Baseline FE
Percent FE gain
Percent CO2 reduction
Midsize Car
0.252 kwh/mi
20 miles
0.30
49.1 mpg
70.1 mpg
29.3 mpg
139%
-58%
Large Truck
0.429 kwh/mi
20 miles
0.30
25.6 mpg
36.6 mpg
19.2 mpg
90%
-47%
Calculating a total fuel consumption and tailpipe COi reduction based on model
outputs and the Utility Factor calculations results in a 58% reduction for small cars, large
cars, minivans, and small trucks. For large trucks, the result is a 47% reduction. The
lower improvement is due to less engine downsizing in the large truck class. The CAFE
model does not apply this technology to medium and large trucks and SUVs.
Using the hybrid cost estimating model and updated component costs, EPA and
NHTSA determined MY 2012 incremental cost estimates for the CAFE model ranging
from $11,527 for a subcompact car to $19,265 for a midsize performance car. This
includes the 1.64 ICM markup value for very high complexity technology. In the
OMEGA model, and relative to a IMA/ISA/BSG, these values range from $6,922 for
small car to $12,467 for a large truck, and relative to power-split hybrid, these values
range from $5,423 for small car to $9,643 for large truck (all values for MY 2016 in
2007$). Volume-based learning is applied to plug-in hybrid costs since this is a newly
emerging technology.
3.4.2.4 Vehicle Technologies
3.4.2.4.1 Mass Reduction
Reducing a vehicle's mass, or down weighting the vehicle, decreases fuel
consumption by reducing the energy demand that is needed to overcome forces resisting
motion, and rolling resistance. Manufacturers employ a systematic approach to mass
reduction were the net mass reduction is the addition of a direct component or system
mass reduction plus the additional mass reduction taken from indirect ancillary systems
and components, as a result of full vehicle optimization, effectively compounding the
mass reductions. Many manufacturers have already announced proposed future products
plans reducing the weight of a vehicle body through the use of high strength steel body-
in-white, composite body panels, magnesium alloy front and rear energy absorbing
structures reducing vehicle weight sufficiently to allow a smaller, lighter and more
efficient engine.
3-69
-------
Draft Joint Technical Support Document
Use of a smaller, lighter engine with lower torque-output subsequently allows the
use of a smaller, lighter-weight transmission and drive line components. The
compounded weight reductions of the body, engine and drivetrain reduce stresses on the
suspension components, steering components, wheels, tires and brakes, allowing further
reductions in the weight of these subsystems. The reductions in weight for unsprung
masses such as brakes, control arms, wheels and tires further reduce stresses in the
suspension mounting points allowing still furthering the compounding effect of mass
reductions.
Down-weighting through material substitution is broadly applicable across all
vehicle subsystems including the engine, exhaust system, transmission, chassis,
suspension, brakes, body, closure panels, glazing, seats and other interior components,
engine cooling systems and HVAC systems. Weight reductions of 10 percent without the
compounding effects account for a 2 to 4 percent fuel consumption and CO2 emissions
reduction. When the mass reduction compounding techniques are used, this value is
increased to a 6 to 7 percent fuel consumption and CO2 emissions reduction.48'49'50 It is
estimated that up to 1.25 kilograms of secondary weight savings can be achieved for
every kilogram of weight saved on a vehicle when all subsystems are redesigned to take
into account the initial primary weight savings.51'52
Down-weighting/material substitution can be accomplished by proven methods
such as:
Smart Design: Computer aided engineering (CAE) tools can be used to better
optimize load paths within structures by reducing stresses and bending moments
applied to structures (Figure 3-4 through Figure 3-6). This allows better
optimization of the sectional thicknesses of structural components to reduce mass
while maintaining or improving the function of the component. Smart designs
also integrate separate parts in a manner that reduces mass by combining
functions or the reduced use of separate fasteners. In addition, some "body on
frame" vehicles are redesigned with a lighter "unibody" construction.
Material Substitution: Substitution of lower density and/or higher strength
materials into a design in a manner that preserves or improves the function of the
component. This includes substitution of high-strength steels, aluminum,
magnesium or composite materials for components currently fabricated from mild
steel.
Reduced Powertrain Requirements: Reducing vehicle weight sufficiently allows
for the use of a smaller, lighter and more efficient engine while maintaining or
increasing performance. Approximately half of the reduction is due to these
reduced powertrain output requirements from reduced engine power output and/or
displacement, changes to transmission and final drive gear ratios. The subsequent
reduced rotating mass (e.g. transmission, driveshafts/halfshafts, wheels and tires)
via weight and/or size reduction of components are made possible by reduced
torque output requirements.
3-70
-------
Using smart design, material substitution and space efficient designs have already
been integrated into vehicles as new platforms cycle into production although net vehicle
weight and performance has continued to increase (Figure 3-4).
Roof Rail 2
Roof Rail 1
IP Beam
Front Header
Roof Bow
(14 positions)
Rocker
Inner
Front Seat
Cross-member
(8 positions)
Kick down
cross-member
Figure 3-4: Example of CAE optimization of the design and position of the roof rails and front seat
cross-member for the "Future Generation Passenger Compartment."53 The thin lines represent
relative positions for roof bows and front cross-members considered during iterations of the CAE
analysis of the vehicle structure. The different colors given to the structural components represent
differences in sectional material thickness within this particular CAE modeling run.
3-71
-------
Draft Joint Technical Support Document
Material = 500800
Thickness 1.0mm
Baseline Design
Mass = 13.9 kg
Material = 500800
Thickness 0.6mm
Thickiiess= 0.7mm
I
41%
Thickiiess= O.Smin
Optimized Design
Mass = 8.2 kg
Figure 3-5: Example of CAE optimization of a roof-rail design for the "Future Generation Passenger
Compartment" that reduces component mass by 41 %.53
Baseline Design
Mass = 12.58 kg
i
57%
Optimized Design
Mass = 5.38 kg
Figure 3-6: Example of CAE optimization of a door-beam design for the "Future Generation
Passenger Compartment" that reduces component mass by 57% and reduces part count and
assemby.
53
3-72
-------
ML
Figure 3-7: Lightweight, tubular front door intrusion beams (2007 Civic)
Automotive companies have largely used weight savings in some vehicle
subsystems to offset or mitigate weight gains in other subsystems from increased
feature content (sound insulation, entertainment systems, improved climate
control, panoramic roof, etc.).
Lightweight designs have also been used to improve vehicle performance
parameters by increased acceleration performance or superior vehicle handling
and braking.
Many manufacturers have already announced proposed future products plans
reducing the weight of a vehicle body through the use of high strength steel body-in-
white, composite body panels, magnesium alloy front and rear energy absorbing
structures reducing vehicle weight sufficiently to allow a smaller, lighter and more
efficient engine. Use of a smaller, lighter engine with lower torque-output subsequently
allows the use of a smaller, lighter-weight transmission and drive line components. The
compounded weight reductions of the body, engine and drivetrain reduce stresses on the
suspension components, steering components, wheels, tires and brakes, allowing further
reductions in the weight of these subsystems. The reductions in weight for unsprung
masses such as brakes, control arms, wheels and tires further reduce stresses in the
suspension mounting points allowing still further reductions in weight.
3-73
-------
Draft Joint Technical Support Document
NHTSA in the CAFE model has used a tiered approach using material
substitution to reduce mass without considering other means of vehicle mass reduction.
MSI, a 1.5 percent mass reduction, is applicable in any year at refresh or redesign. MS2,
a 3.5 to 8.5 percent mass reduction is only available in MY 2014 and beyond in an
attempt to mimic the product plans' applications of down-weighting/material substitution
in later years. For smaller vehicles MS2 was limited to a 3.5 percent mass reduction of
down-weighting/material substitution on smaller vehicles and for larger cars and trucks
up to 8.5 percent mass reduction.
In the CAFE model, MS2 is cumulative to MSI, as it is only applied after MSI;
therefore the maximum weight reduction that can occur for smaller subclass vehicles is
5%, while large cars, truck, and SUVs could experience up to a 10% weight reduction.
Restricting weight reduction on smaller vehicles to lower limits, and vice versa for larger
vehicles, is intended to mitigate or minimize the potential safety consequences from the
modeled down-weighting/material substitution. Making the MS2 technology unavailable
until MY 2014 is intended to recognize the lead time required to implement platform
redesigns that would be necessary for these levels of weight reduction.
NHTSA and EPA estimate that a 10 percent reduction in mass results in a 6.5
percent reduction in fuel consumption while maintaining equivalent vehicle performance
(i.e. 0-60 mph time, towing capacity, etc.) which is consistent with estimates in the 2002
NAS report
NHTSA and EPA have reviewed three studies of down-weighting/material
substitution and the associated cost. The first, NAS2, estimated that vehicle weight could
be reduced for approximately $1.50 per pound. (3-4% reductions in fuel consumption,
without engine downsizing, from a 5% reduction in vehicle weight at a cost of $210-
$350. This translates into $1.50 per pound, assuming a 3800 pound base vehicle and
using the midpoint cost.) Additionally, Sierra Research estimates a 10% reduction, with
compounding, could be accomplished for a cost of $1.01 per pound54. Finally, MIT55
estimated that the weight of a vehicle could be reduced by 14%, with no compounding,
for a cost of $1.36 per pound. The final cost estimate is $1.32 per pound and is based on
the average of the three referenced studies. Applying an ICM factor of 1.11 for a low
complexity technology results in a compliance cost of $1.48 per pound. For the vehicle
mass reduction technologies, neither volume-based nor time-based cost reductions are
applied since many of the materials under consideration are commodity based.
3.4.2.4.2 Low Drag Brakes (LDB)
Low drag brakes reduce the sliding friction of disc brake pads on rotors when the
brakes are not engaged because the brake pads are pulled away from the rotating disc
either by mechanical or electric methods. While most passenger cars have already
adopted this technology with the standardization of electronic brake control, there are
indications that this technology is still available for body-on-frame trucks.
3-74
-------
NHTSA's MY 2011 CAFE final rale estimated the effectiveness of LDB to be up
to 1 percent, based on confidential manufacturer data. NHTSA and EPA have reviewed
this estimate and believe it to be applicable for the NPRM.
NHTSA and EPA have reviewed the cost estimates from the MY 2011 CAFE
final rule and have not found any evidence to suggest that this cost is not a valid cost and
have chosen to maintain this estimate. However, the agencies adjusted the costs to apply
the ICM of 1.11, for a low complexity technology. The compliance cost for LDB is
estimated to be $63 for MY 2012 through MY 2016 vehicles as no learning effects are
applied to this technology..
3.4.2.4.3 Low Rolling Resistance Tires (ROLL)
Tire rolling resistance is the frictional loss associated mainly with the energy
dissipated in the deformation of the tires under load and thus influences fuel economy
and COi emissions. Other tire design characteristics (e.g., materials, construction, and
tread design) influence durability, traction (both wet and dry grip), vehicle handling, and
ride comfort in addition to rolling resistance. A typical low rolling resistance tires
attributes would include: increased tire inflation pressure, material changes, and tire
construction with less hysteresis, geometry changes (e.g., reduced aspect ratios), and
reduction in sidewall and tread deflection. These changes would generally be
accompanied with additional changes to suspension tuning and/or suspension design.
EPA and NHTSA estimate a 1 to 2 percent increase in effectiveness with a 10
percent reduction in rolling resistance, which was based on the 2002 NAS report findings
and consistent with NHTSA's MY 2011 CAFE final rale estimate. NHTSA and EPA
still believe that this NAS effectiveness estimate is valid for this NPRM. Lower rolling
resistance tires are widely available today.
Based on NHTSA' MY 2011 CAFE final rule and the 2006 NAS/NRC report, the
agencies have estimated the cost for low rolling resistance tires to be $6 per vehicle.56
This is based on a cost of $1 per tire as estimated by NAS/NRC 2006 report, which is $5
per vehicle, including the spare tire. When applying the ICM low complexity markup
factor, of 1.11, results in compliance cost of $6 per vehicle for MY 2012 through MY
2016 vehicles (2007$).p Due to the commodity based nature of this technology cost
learning is not applied.
p Note that the costs developed for low rolling resistance tires for this analysis do not include the increase in
lifetime costs that would be expected at each tire replacement. Instead, the analysis includes only the
upfront increase in costs. The agencies intend to include the lifetime costs in the final analysis.
3-75
-------
Draft Joint Technical Support Document
3.4.2.4.4 Front or Secondary Axle Disconnect for Four-Wheel Drive Systems (SAX)
Energy is required to continually drive the front, or secondary, axle in a four
wheel drive system even when the system is not required during most operating
conditions. This energy loss directly results in increased fuel consumption and COi
emissions. Many part-time four-wheel drive systems use some type of front axle
disconnect to provide shift-on-the-fly capabilities,. The front axle disconnect is normally
part of the front differential assembly. As part of a shift-on-the-fly four-wheel drive
system, the front axle disconnect serves two basic purposes. First, in two-wheel-drive
mode, it disengages the front axle from the front driveline so the front wheels do not turn
the front driveline at road speed, saving wear and tear. Second, when shifting from two-
to four-wheel drive "on the fly" (while moving), the front axle disconnect couples the
front axle to the front differential side gear only when the transfer case's synchronizing
mechanism has spun the front driveshaft up to the same speed as the rear driveshaft.
Four-wheel drive systems that have a front axle disconnect typically do not have either
manual- or automatic-locking hubs. To isolate the front wheels from the rest of the front
driveline, front axle disconnects use a sliding sleeve to connect or disconnect an axle
shaft from the front differential side gear. NHTSA and EPA are not aware of any
manufacturer offering this technology in the US today on unibody frame vehicles;
however, it is possible this technology could be introduced by manufacturers within the
rulemaking time period.
Based on confidential manufacturer data, NHTSA's MY 2011 CAFE final rule
estimated an effectiveness improvement of 1 to 1.5 percent for axle disconnect. NHTSA
and EPA believe this to still be an accurate estimate for this rulemaking.
Regarding costs, the agencies believe the incremental compliance cost from
NHTSA's MY 2011 final rule to be accurate. However, this was revised by applying the
new ICM factor markup, for a low complexity technology, of 1.11. Thus, the compliance
cost estimate for this NPRM, is $95 for a MY 2012 vehicle (2007$). With time-based
learning this cost would be $84 for a MY 2016 vehicle (2007$).
3.4.2.4.5 Aerodynamic Drag Reduction (AERO)
Many factors affect a vehicle's aerodynamic drag and the resulting power
required to move it through the air. While these factors change with air density and the
square and cube of vehicle speed, respectively, the overall drag effect is determined by
the product of its frontal area and drag coefficient. Reductions in these quantities can
therefore reduce fuel consumption and COi emissions. Although frontal areas tend to be
relatively similar within a vehicle class (mostly due to market-competitive size
requirements), significant variations in drag coefficient can be observed. Significant
changes to a vehicle's aerodynamic performance may need to be implemented during a
3-76
-------
redesign (e.g. changes in vehicle shape). However, shorter-term aerodynamic reductions,
with a somewhat lower effectiveness, may be achieved through the use of revised exterior
components (typically at a model refresh in mid-cycle) and add-on devices that currently
being applied. The latter list would include revised front and rear fascias, modified front
air dams and rear valances, addition of rear deck lips and underbody panels, and lower
aerodynamic drag exterior mirrors.
The MY 2011 CAFE final rule estimated that a fleet average of 10 to 20 percent
total aerodynamic drag reduction is attainable (with a caveat for "high-performance"
vehicles described below) which equates to incremental reductions in fuel consumption
and COi emissions of 2 percent and 3 percent for cars and trucks. These numbers are
generally supported by confidential manufacturer data and public technical literature and
therefore NHTSA and EPA are retaining these estimates, as confirmed by joint review,
for the purposes of this proposal.
In the MY 2011 CAFE final rule, NHTSA estimated a range from $60 to $116
(1.5 RPE); $40 to $75 (without RPE). .NHTSA and EPA reviewed these estimates and
updated these estimates to reflect the new ICM markup factor of 1.11 for a low
complexity technology. In addition, given its low complexity and the facility of its
implementation during design and refresh cycles, NHTSA and EPA estimate closer to the
lower end of the range is more appropriate. Thus, the cost estimate used in this
rulemaking is $48, including the 1.11 ICM markup value, for a low complexity
technology (2007$). This compliance cost is for a MY 2012 vehicle. With time-based
learning, this cost would be $42 for a MY 2016 vehicle (2007$)
3.4.2.5 Technologies considered but not included in the final rule
analysis
NHTSA and EPA have identified six technologies that will not be available in the
time frame considered under this rulemaking (though electric vehicles are available
today). These technologies while considered were not made available in the CAFE and
OMEGA models. They are: electric vehicles, camless valve actuation (CVA), lean burn
gasoline direct injection (LBDI), homogeneous charge compression ignition (HCCI), and
electric assist turbocharging and full series hydraulic hybrids (HHV). NHTSA and EPA
will continue to monitor the industry and system suppliers for progress on these
technologies, and should they become more available, consider them for use in future
rulemaking activity.
3.4.2.5.1 Electric Vehicles
The recent intense interest in Hybrid vehicles and the development of Hybrid
vehicle battery and motor technology has helped make Electric Vehicle technology more
viable than it has ever been. Electric Vehicles (EVs) require much larger batteries than
either HEVs or PHEVs, but the batteries must be of a high-energy and lower-power
3-77
-------
Draft Joint Technical Support Document
design to deliver an appropriate amount of power over the useful charge of the battery.
These high-energy batteries are generally less expensive per kilowatt-hour than high-
power batteries required for hybrids, but the size of the battery pack still incurs a
considerable cost.
Electric motor and power electronics designs are very similar to HEV and PHEV
designs, but they must be larger, more powerful, and more robust since they provide the
only motive power for the vehicle. On the other hand, the internal combustion engine,
fuel system, and possibly the transmission can all be removed for significant weight,
complexity and cost savings.
As for PHEVs, two electric vehicles were modeled, a small car and a large car,
using the same model (PEREGRIN) and similar assumptions. EVs are only considered
for these two classes because the larger, heavier vehicles would require too much battery
capacity to be practical in the short-to-mid term and the agencies do not see any serious
development activities for these vehicle types in the market.
We chose to model the EVs with a range of 150 miles on the urban driving cycle
because this range offers a good compromise in capability and battery cost, weight and
size with expected technology in the near- to mid-term. Using the same methodology as
used for PHEVs to calculate gasoline-equivalent fuel consumption, the results obtained
are shown in Table 3-19, below.
Table 3-19 Electric Vehicle Gasoline-Equivalent CO2 Reduction
EV energy comb (0.55 city / 0.45 hwy)
City cycle EV range
Highway cycle EV range
Baseline FE
Tailpipe CO2 reduction
Small Car
0.202 kwh/mi
150 miles
166 miles
35.5
100%
Large Car
0.244 kwh/mi
150 miles
162 miles
25.3 mpg
100%
3.4.2.5.2 Camless Valve Actuation
Camless valve actuation relies on electromechanical actuators instead of
camshafts to open and close the cylinder valves. When electromechanical actuators are
used to replace cams and coupled with sensors and microprocessor controls, valve timing
and lift can be optimized over all conditions. An engine valvetrain that operates
independently of any mechanical means provides the increased flexibility for intake and
exhaust timing and lift optimization. With it comes increased ability to vary valve
overlap, the rapid response required to change between combustion operating modes
(such as HCCI and GDI), intake valve throttling, cylinder deactivation, and elimination
3-78
-------
of the camshafts (reduced friction and rotating mass). This level of control can enable
even further incremental reductions in fuel consumption and COi emissions.
This technology has been under research for many decades and although progress
is being made, NHTSA and EPA have not found evidence to support that the technology
can be successfully implemented within the 2012 through 2016 timeframe of these
regulations. Thus NHTSA and EPA have not estimated cost or effectiveness at this time.
3.4.2.5.3 Lean-Burn Gasoline Direct Injection Technology
Direct injection, especially with diesel-like "spray-guided" injection systems,
enables operation with excess air in a stratified or partially-stratified fuel-air mixture, as a
way of reducing the amount of intake throttling. Also, with higher-pressure fuel injection
systems, the fuel may be added late enough during the compression stroke so as to delay
the onset of autoignition, even with higher engine compression ratios or with boosted
intake pressure. Taken together, an optimized "lean-burn" direct injection gasoline
engine may achieve high engine thermal efficiency which approaches that of a diesel
engine. European gasoline direct-injection engines have implemented stratified-charge
lean-burn GDI, although at higher NOx emissions levels than are allowed at under U.S.
Federal Tier 2 emissions standards. Fuel system improvements, changes in combustion
chamber design and repositioning of the injectors have allowed for better air/fuel mixing
and combustion efficiency. There is currently a shift from wall-guided injection to spray
guided injection, which improves injection precision and targeting towards the spark
plug, increasing lean combustion stability. Combined with advances in NOX after-
treatment, lean-burn GDI engines may eventually be a possibility in North America.
EPA and NHTSA's current assessment is that the availability of ultra-low sulfur
(less than 15 ppm sulfur) gasoline is a key technical requirement for lean-burn GDI
engines to meet EPA's Tier 2 NOX emissions standards, therefore the technology was not
applied in within EPA or NHTSA analyses.
3.4.2.5.4 Homogeneous Charge Compression Ignition
Gasoline homogeneous charge compression ignition (HCCI), also referred to as
controlled autoignition (CAI), is an alternate engine operating mode that does not rely on
a spark event to initiate combustion. The principles are more closely aligned with a
diesel combustion cycle, in which the compressed charge exceeds a temperature and
pressure necessary for spontaneous autoignition although it differs from diesel by having
a homogenous fuel/air charge rather than being a diffusion controlled combustion event.
The subsequent combustion event is much shorter in duration with higher thermal
efficiency.
3-79
-------
Draft Joint Technical Support Document
An HCCI engine has inherent advantages in its overall efficiency for two main
reasons:
The engine is operated with a higher compression ratio, and with a shorter
combustion duration, resulting in a higher thermodynamic efficiency, and
The engine can be operated virtually unthrottled, even at light loads,
Combined, these effects have shown an increase in engine brake efficiency
(typically 25-28%) to greater than 35% at the high end of the HCCI operating range.57
Criteria pollutant emissions are very favorable during HCCI operation. Lower peak in-
cylinder temperatures (due to high dilution) keep engine-out NOx emissions to a
minimum - realistically below Tier 2 levels without aftertreatment - and particulates are
low due to the homogeneous nature of the premixed charge.
Due to the inherent difficulty in maintaining combustion stability without
encountering engine knock, HCCI is difficult to control, requiring feedback from in-
cylinder pressure sensors and rapid engine control logic to optimize combustion timing,
especially considering the transient nature of operating conditions seen in a vehicle. Due
to the highly dilute conditions under which gasoline-HCCI combustion is stable, the
range of engine loads achievable in a naturally-aspirated engine is somewhat limited.
Because of this, it is likely that any commercial application would operate in a "dual-
mode" strategy between HCCI and spark ignition combustion modes, in which HCCI
would be utilized for best efficiency at light engine loads and spark ignition would be
used at higher loads and at idle. This type of dual-mode strategy has already been
employed in diesel HCCI engines in Europe and Asia (notably the Toyota Avensis D-Cat
and the Nissan light-duty "MK" combustion diesels).
Until recently, gasoline-HCCI technology was considered to still be in the
research phase. However, most manufacturers have made public statements about the
viability of incorporating HCCI into light-duty passenger vehicles, and have significant
vehicle demonstration programs aimed at producing a viable product within the next 5-10
years.
There is widespread opinion as to the CCh reduction potential for HCCI in the
literature. Based on confidential manufacturer information, EPA and NHTSA believe
that a gasoline HCCI / GDI dual-mode engine might achieve 10-12% reduction in COi,
compared to a comparable SI engine. Despite its promise, application of HCCI in light
duty vehicles is not yet ready for the market. It is not anticipated to be seen in volume for
at least the next 5-10 years, which is concurrent with many manufacturers' public
estimates. NHTSA also noted in its MY 2011 CAFE final rule that the technology will
not be available within the time frame considered based on a review of confidential
product plan information.
3-80
-------
3.4.2.5.5 Electric Assist Turbocharging
The Alliance commented in NHTSA's previous rulemaking that global
development of electric assist turbocharging has not demonstrated the fuel efficiency
effectiveness of a 12V EAT up to 2kW power levels since the 2004 NESCCAF study,
and stated that it saw remote probability of its application over the next decade. While
hybrid vehicles lower the incremental hardware requirements for higher-voltage, higher-
power EAT systems, NHTSA and EPA agree that significant developmental work is
required to demonstrate effective systems and that implementation in significant volumes
will not occur in the 2012 to 2016 time frame considered in this rulemaking. Thus, this
technology was not included in the NPRM.
3.42.5.6 Full Series Hydraulic Hybrid
A Full Series Hydraulic Hybrid Vehicle (HHV) is somewhat similar in concept to
a full series electric hybrid vehicle, except that the energy is stored in the form of
compressed nitrogen gas and the power is transmitted in the form of hydraulic fluid.
Series HHV technology currently under development by EPA is capable of a 40%
decrease in tailpipe CCh emissions in the small car, large car, minivan, and small truck
classes. In the large truck class, a 30% CCh reduction is possible. The large truck benefit
is somewhat lower than the other classes because it is assumed that a large truck requires
a larger engine to maintain towing and hauling performance after the energy in the high
pressure hydraulic accumulator is exhausted. This technology is still under development
and not yet commercialized however there are technology demonstration vehicles in
service with UPS in daily package delivery service.
3.5 Cost and effectiveness tables
The tables representing the Volpe model input files for incremental technology
costs by vehicle subclass are presented below. The tables have been divided into
passenger cars, performance passenger cars, and light trucks to make them easier to read.
3-81
-------
Draft Joint Technical Support Document
3.5.1 NHTSA cost and effectiveness tables
Table 3-20 Technology Incremental Cost Estimates for Volpe model, Passenger Cars
VEHICLE TECHNOLOGY ICM COSTS PER VEHICLE (2007$) BY VEHICLE
SUBCLASS - PASSENGER CARS
Nominal Baseline Engine (For Cost Basis)
Low Friction Lubricants
Engine Friction Reduction
VVT - Coupled Cam Phasing (CCP) on SOHC
Discrete Variable Valve Lift (DVVL) on SOHC
Cylinder Deactivation on SOHC
VVT - Intake Cam Phasing (ICP)
VVT - Dual Cam Phasing (DCP)
Discrete Variable Valve Lift (DVVL) on DOHC
Continuously Variable Valve Lift (CVVL)
Cylinder Deactivation on DOHC
Cylinder Deactivation on OHV
VVT - Coupled Cam Phasing (CCP) on OHV
Discrete Variable Valve Lift (DVVL) on OHV
Conversion to DOHC with DCP
Stoichiometric Gasoline Direct Injection (GDI)
Combustion Restart
Turbocharging and Downsizing
Exhaust Gas Recirculation (EGR) Boost
Conversion to Diesel following CBRST
Conversion to Diesel following TRBDS
6-Speed Manual/Improved Internals
Improved Auto. Trans. Controls/Externals
Continuously Variable Transmission
6/7/8-Speed Auto. Trans with Improved Internals
Dual Clutch or Automated Manual Transmission
Electric Power Steering
Improved Accessories
12V Micro-Hybrid
Belt mounted Integrated Starter Generator
Crank mounted Integrated Starter Generator
Power Split Hybrid
2-Mode Hybrid
EFR
CCPS
DVVLS
DEACS
ICP
DCP
DVVLD
CVVL
DEACD
DEACO
CCPO
DVVLO
CDOHC
SGDI
CBRST
TRBDS
EGRB
DSLC
DSLT
6MAN
IATC
CVT
NAUTO
DCTAM
EPS
IACC
MHEV
BISG
CISG
PSHEV
2MHEV
Subcompact
Car
Inline 4
3
50
45
142
n.a.
45
38
142
277
n.a.
n.a.
45
142
276
251
118
644
144
n.a.
n.a.
250
60
250
170
73
106
128
288
286
2,791
1,600
n.a.
Compact
Car
Inline 4
3
50
45
142
n.a.
45
38
142
277
n.a.
n.a.
45
142
276
251
118
644
144
n.a.
n.a.
250
60
250
170
73
106
128
311
286
3,107
2,133
n.a.
Midsize
Car
Inline 4
3
50
45
142
n.a.
45
38
142
277
n.a.
n.a.
45
142
276
251
118
644
144
n.a.
n.a.
250
60
250
170
158
106
128
342
286
3,319
2,742
n.a.
Large
Car
V6
3
75
90
205
0-56
90
83
205
509
0-56
170
45
0-56
436
326
118
512
144
1,572 -
1,627
916-971
250
60
250
170
158
106
128
367
286
3,547
3,261
n.a.
3-82
-------
Plug -in Hybrid
Mass Reduction (1.5%)
Mass Reduction (3.5 to 8.5%)
Low Rolling Resistance Tires
Low Drag Brakes
Secondary Axle Disconnect
Aero Drag Reduction
PHEV
MSI
MS2
ROLL
LDB
SAX
AERO
11,520-
11,527
1.5
1.5
6
n.a.
87
48
14,135-
14,142
1.5
1.5
6
n.a.
87
48
16,136-
16,215
1.5
1.5
6
n.a.
87
48
n.a.
1.5
1.5
6
63
87
48
Table 3-21. Technology Incremental Cost Estimates for Volpe model, Performance Passenger cars
VEHICLE TECHNOLOGY ICM COSTS PER VEHICLE (2007$) BY VEHICLE
SUBCLASS - PERFORMANCE PASSENGER CARS
Nominal Baseline Engine (For Cost Basis)
Low Friction Lubricants
Engine Friction Reduction
VVT - Coupled Cam Phasing (CCP) on SOHC
Discrete Variable Valve Lift (DVVL) on SOHC
Cylinder Deactivation on SOHC
VVT - Intake Cam Phasing (ICP)
VVT - Dual Cam Phasing (DCP)
Discrete Variable Valve Lift (DVVL) on DOHC
Continuously Variable Valve Lift (CVVL)
Cylinder Deactivation on DOHC
Cylinder Deactivation on OHV
VVT - Coupled Cam Phasing (CCP) on OHV
Discrete Variable Valve Lift (DVVL) on OHV
Conversion to DOHC with DCP
Stoichiometric Gasoline Direct Injection (GDI)
Combustion Restart
Turbocharging and Downsizing
Exhaust Gas Recirculation (EGR) Boost
Conversion to Diesel following CBRST
Conversion to Diesel following TRBDS
6-Speed Manual/Improved Internals
Improved Auto. Trans. Controls/Externals
Continuously Variable Transmission
6/7/8-Speed Auto. Trans with Improved Internals
Dual Clutch or Automated Manual Transmission
Electric Power Steering
LUB
EFR
CCPS
DVVLS
DEACS
ICP
DCP
DVVLD
CVVL
DEACD
DEACO
CCPO
DVVLO
CDOHC
SGDI
CBRST
TRBDS
EGRB
DSLC
DSLT
6MAN
IATC
CVT
NAUTO
DCTAM
EPS
Perform.
Subcomp.
Car
Inline 4
3
50
45
142
n.a.
45
38
142
277
n.a.
n.a.
45
142
276
251
118
644
144
n.a.
n.a.
250
60
250
170 - 272
73
106
Perform.
Compact
Car
V6
3
75
90
205
0-56
90
83
205
509
0-56
170
45
0-56
436
326
118
512
144
1,572-
1,627
916-971
250
60
250
170 - 272
158
106
Perform.
Midsize
Car
V6
3
75
90
205
0-56
90
83
205
509
0-56
170
45
0-56
436
326
118
512
144
1,572-
1,627
916-971
250
60
250
170 - 272
158
106
Perform.
Large
Car
V8
3
101
90
293
0-56
90
82
293
555
0-56
190
45
0-56
552
353
118
1,098
144
2,331 -
2,377
1,090 -
1,145
250
60
250
170 - 272
158
106
3-83
-------
Draft Joint Technical Support Document
Improved Accessories
12V Micro-Hybrid
Belt mounted Integrated Starter Generator
Crank mounted Integrated Starter Generator
Power Split Hybrid
2-Mode Hybrid
Plug-in Hybrid
Mass Reduction (1.5%)
Mass Reduction (3.5 to 8.5%)
Low Rolling Resistance Tires
Low Drag Brakes
Secondary Axle Disconnect
Aero Drag Reduction
IACC
MHEV
BISG
CISG
PSHEV
2MHEV
PHEV
MSI
MS2
ROLL
LDB
SAX
AERO
128
314
286
2,839
3,661
n.a.
14,891 -
14,993
1.5
1.5
n.a.
n.a.
87
n.a.
128
372
286
3,335
5,106-
5,287
n.a.
19,085 -
19,265
1.5
1.5
n.a.
n.a.
87
n.a.
128
372
286
3,149
3,838 -
4,018
n.a.
16,612-
16,714
1.5
1.5
n.a.
n.a.
87
n.a.
128
410
286
3,571
6,543 -
6723
n.a.
n.a.
1.5
1.5
n.a.
63
87
n.a.
Table 3-22 . Technology Incremental Cost Estimates for Volpe model, Light Trucks
VEHICLE TECHNOLOGY ICM COSTS PER VEHICLE (2007$) BY VEHICLE
SUBCLASS - LIGHT TRUCKS
Nominal Baseline Engine (For Cost Basis)
Low Friction Lubricants
Engine Friction Reduction
VVT - Coupled Cam Phasing (CCP) on SOHC
Discrete Variable Valve Lift (DVVL) on SOHC
Cylinder Deactivation on SOHC
VVT - Intake Cam Phasing (ICP)
VVT - Dual Cam Phasing (DCP)
Discrete Variable Valve Lift (DVVL) on DOHC
Continuously Variable Valve Lift (CVVL)
Cylinder Deactivation on DOHC
Cylinder Deactivation on OHV
VVT - Coupled Cam Phasing (CCP) on OHV
Discrete Variable Valve Lift (DVVL) on OHV
Conversion to DOHC with DCP
Stoichiometric Gasoline Direct Injection (GDI)
Combustion Restart
Turbocharging and Downsizing
Exhaust Gas Recirculation (EGR) Boost
Conversion to Diesel following CBRST
LUB
EFR
CCPS
DVVLS
DEACS
ICP
DCP
DVVLD
CVVL
DEACD
DEACO
CCPO
DVVLO
CDOHC
SGDI
CBRST
TRBDS
EGRB
DSLC
Minivan
LT
V6
3
75
90
205
0-56
90
83
205
509
0-56
170
45
0-56
436
326
118
512
144
1,572-
1,627
Small
LT
Inline 4
3
50
45
142
n.a.
45
38
142
277
n.a.
n.a.
45
142
276
251
118
644
144
n.a.
Midsize
LT
V6
3
75
90
205
0-56
90
83
205
509
0-56
170
45
0-56
436
326
118
512
144
1,572 -
1,627
Large
LT
V8
3
101
90
293
0-56
90
83
293
555
0-56
190
45
0-56
552
353
118
1,098
144
2,331 -
2,387
3-84
-------
Conversion to Diesel following TRBDS
6-Speed Manual/Improved Internals
Improved Auto. Trans. Controls/Externals
Continuously Variable Transmission
6/7/8-Speed Auto. Trans with Improved Internals
Dual Clutch or Automated Manual Transmission
Electric Power Steering
Improved Accessories
12V Micro-Hybrid
Belt mounted Integrated Starter Generator
Crank mounted Integrated Starter Generator
Power Split Hybrid
2-Mode Hybrid
Plug-in Hybrid
Mass Reduction (1.5%)
Mass Reduction (3.5 to 8.5%)
Low Rolling Resistance Tires
Low Drag Brakes
Secondary Axle Disconnect
Aero Drag Reduction
DSLT
6MAN
IATC
CVT
NAUTO
DCTAM
EPS
IACC
MHEV
BISG
CISG
PSHEV
2MHEV
PHEV
MSI
MS2
ROLL
LDB
SAX
AERO
916-971
250
60
250
170
158
106
128
367
286
3,547
3,261
n.a.
n.a.
1.5
1.5
6
63
87
48
n.a.
250
60
250
170
73
106
128
325
286
3,141
2,377 -
2,384
3,521
14,589
1.5
1.5
6
n.a.
87
48
916-971
250
60
250
170 - 272
158
106
128
376
286
3,611
3,282 -
3,462
4,663 -
4,764
n.a.
1.5
1.5
6
63
87
48
1,090 -
1,145
250
60
250
170 - 272
158
106
n.a.
n.a.
n.a.
5,124
n.a.
5,678 -
5,779
n.a.
1.5
1.5
6
63
87
48
Table 3-23. Technology Incremental Effectiveness Estimates for Volpe model,
Passenger Cars
VEHICLE TECHNOLOGY INCREMENTAL FUEL CONSUMPTION REDUCTION
(-%) BY VEHICLE SUBCLASS - PASSENGER CAR
Nominal Baseline Engine (For Cost Basis)
Low Friction Lubricants
Engine Friction Reduction
VVT - Coupled Cam Phasing (CCP) on
SOHC
Discrete Variable Valve Lift (DVVL) on
SOHC
Cylinder Deactivation on SOHC
VVT - Intake Cam Phasing (ICP)
VVT - Dual Cam Phasing (DCP)
Discrete Variable Valve Lift (DVVL) on
DOHC
Continuously Variable Valve Lift (CVVL)
Cylinder Deactivation on DOHC
Cylinder Deactivation on OHV
LUB
EFR
CCPS
DVVLS
DEACS
ICP
DCP
DVVLD
CVVL
DEACD
DEACO
Subcompact
Car
Inline 4
0.5
1.0-2.0
1.0-3.0
1.0-3.0
n.a.
1.0-2.0
2.0 - 3.0
1.0-3.0
1.5-3.5
n.a.
n.a.
Compact
Car
Inline 4
0.5
1.0-2.0
1.0-3.0
1.0-3.0
n.a.
1.0-2.0
2.0 - 3.0
1.0-3.0
1.5-3.5
n.a.
n.a.
Midsize
Car
Inline
4
0.5
1.0-2.0
1.0-3.0
1.0-3.0
n.a.
1.0-2.0
2.0-3.0
1.0-3.0
1.5-3.5
n.a.
n.a.
Large
Car
V6
0.5
1.0-2.0
1.0-3.0
1.0-3.0
2.5-3.0
1.0-2.0
2.0 - 3.0
1.0-3.0
1.5-3.5
0-0.5
3.9-5.5
3-85
-------
Draft Joint Technical Support Document
VVT - Coupled Cam Phasing (CCP) on OHV
Discrete Variable Valve Lift (DVVL) on
OHV
Conversion to DOHC with DCP
Stoichiometric Gasoline Direct Injection
(GDI)
Combustion Restart
Turbocharging and Downsizing
Exhaust Gas Recirculation (EGR) Boost
Conversion to Diesel following CBRST
Conversion to Diesel following TRBDS
6-Speed Manual/Improved Internals
Improved Auto. Trans. Controls/Externals
Continuously Variable Transmission
6/7/8-Speed Auto. Trans with Improved
Internals
Dual Clutch or Automated Manual
Transmission
Electric Power Steering
Improved Accessories
12V Micro-Hybrid
Belt mounted Integrated Starter Generator
Crank mounted Integrated Starter Generator
Power Split Hybrid
2-Mode Hybrid
Plug-in Hybrid
Mass Reduction (1.5%)
Mass Reduction (3.5 to 8.5%)
Low Rolling Resistance Tires
Low Drag Brakes
Secondary Axle Disconnect
Aero Drag Reduction
CCPO
DVVLO
CDOHC
SGDI
CBRST
TRBDS
EGRB
DSLC
DSLT
6MAN
IATC
CVT
NAUTO
DCTAM
EPS
IACC
MHEV
BISG
CISG
PSHEV
2MHEV
PHEV
MSI
MS2
ROLL
LDB
SAX
AERO
1.0-1.5
0.5-2.5
1.0-2.5
2.0-3.0
2.0 - 2.5
4.2 - 4.8
3.5-4.0
n.a.
n.a.
0.5
1.5-2.5
0.7 - 2.0
1.4-3.4
2.7-4.1
1.0-2.0
1.0-2.0
2.0 - 3.0
4.0 - 6.0
8.6-8.9
6.3 - 12.4
n.a.
45.2 - 47.7
1.0
2.3
1.0-2.0
n.a.
1.0- 1.5
2.0-3.0
1.0-1.5
0.5-2.5
1.0-2.5
2.0 - 3.0
2.0 - 2.5
4.2 - 4.8
3.5-4.0
n.a.
n.a.
0.5
1.5-2.5
0.7 - 2.0
1.4-3.4
2.7-4.1
1.0-2.0
1.0-2.0
2.0 - 3.0
4.0 - 6.0
8.6-8.9
6.3 - 12.4
n.a.
45.2-
47.7
1.0
2.3
1.0-2.0
n.a.
1.0-1.5
2.0 - 3.0
1.0-1.5
0.5-2.5
1.0-2.5
2.0-3.0
2.0-2.5
4.2-4.8
3.5-4.0
n.a.
n.a.
0.5
1.5-2.5
0.7-2.0
1.4-3.4
2.7-4.1
1.0-2.0
1.0-2.0
2.0-3.0
4.0-6.0
8.6-8.9
6.3-
12.4
n.a.
45.2-
47.7
1.0
3.9
1.0-2.0
n.a.
1.0-1.5
2.0-3.0
1.0-1.5
0.5-2.5
1.0-2.5
2.0 - 3.0
2.0 - 2.5
1.8-1.9
3.5-4.0
10.8-
11.7
5.3-6.9
0.5
1.5-2.5
0.7 - 2.0
1.4-3.4
2.7-4.1
1.0-2.0
1.0-2.0
2.5-3.5
3.5 -5.5
8.7-8.9
6.3-
12.4
n.a.
n.a.
1.0
5.6
1.0-2.0
0.5 - 1.0
1.0-1.5
2.0 - 3.0
3-86
-------
Table 3-24. Component Technology Incremental Effectiveness Estimates for Volpe
model, Performance Cars
VEHICLE TECHNOLOGY INCREMENTAL FUEL CONSUMPTION REDUCTION
(-%) BY VEHICLE SUBCLASS - PERFORMANCE PASSENGER CAR
Nominal Baseline Engine (For Cost Basis)
Low Friction Lubricants
Engine Friction Reduction
VVT - Coupled Cam Phasing (CCP) on
SOHC
Discrete Variable Valve Lift (DVVL) on
SOHC
Cylinder Deactivation on SOHC
VVT - Intake Cam Phasing (ICP)
VVT - Dual Cam Phasing (DCP)
Discrete Variable Valve Lift (DVVL) on
DOHC
Continuously Variable Valve Lift (CVVL)
Cylinder Deactivation on DOHC
Cylinder Deactivation on OHV
VVT - Coupled Cam Phasing (CCP) on OHV
Discrete Variable Valve Lift (DVVL) on
OHV
Conversion to DOHC with DCP
Stoichiometric Gasoline Direct Injection
(GDI)
Combustion Restart
Turbocharging and Downsizing
Exhaust Gas Recirculation (EGR) Boost
Conversion to Diesel following CBRST
Conversion to Diesel following TRBDS
6-Speed Manual/Improved Internals
Improved Auto. Trans. Controls/Externals
Continuously Variable Transmission
6/7/8-Speed Auto. Trans with Improved
Internals
Dual Clutch or Automated Manual
Transmission
Electric Power Steering
Improved Accessories
12V Micro-Hybrid
Belt mounted Integrated Starter Generator
Crank mounted Integrated Starter Generator
Power Split Hybrid
2-Mode Hybrid
LUB
EFR
CCPS
DVVLS
DEACS
ICP
DCP
DVVLD
CVVL
DEACD
DEACO
CCPO
DVVLO
CDOHC
SGDI
CBRST
TRBDS
EGRB
DSLC
DSLT
6MAN
IATC
CVT
NAUTO
DCTAM
EPS
IACC
MHEV
BISG
CISG
PSHEV
2MHEV
Perform.
Subcomp.
Car
Inline 4
0.5
1.0-2.0
1.0-3.0
1.0-3.0
n.a.
1.0-2.0
2.0-3.0
1.0-3.0
1.5-3.5
n.a.
n.a.
1.0-1.5
0.5-2.5
1.0-2.5
2.0-3.0
2.0-2.5
4.2-4.8
3.5-4.0
n.a.
n.a.
0.5
1.5-2.5
0.7-2.0
1.4-3.4
2.7-4.1
1.0-2.0
1.0-2.0
2.0-3.0
4.0-6.0
8.6-8.9
6.3 - 12.4
n.a.
Perform.
Compact
Car
V6
0.5
1.0-2.0
1.0-3.0
1.0-3.0
2.5-3.0
1.0-2.0
2.0-3.0
1.0-3.0
1.5-3.5
0.0-0.5
3.9-5.5
1.0-1.5
0.5-2.5
1.0-2.5
2.0-3.0
2.0-2.5
1.8-1.9
3.5-4.0
10.8-
11.7
5.3-6.9
0.5
1.5-2.5
0.7-2.0
1.4-3.4
2.7-4.1
1.0-2.0
1.0-2.0
2.5-3.5
3.5-5.5
8.7-8.9
6.3-
12.4
n.a.
Perform.
Midsize
Car
V6
0.5
1.0-2.0
1.0-3.0
1.0-3.0
2.5-3.0
1.0-2.0
2.0-3.0
1.0-3.0
1.5-3.5
0.0-0.5
3.9-5.5
1.0-1.5
0.5-2.5
1.0-2.5
2.0-3.0
2.0-2.5
1.8-1.9
3.5-4.0
10.8-
11.7
5.3-6.9
0.5
1.5-2.5
0.7-2.0
1.4-3.4
2.7-4.1
1.0-2.0
1.0-2.0
2.5-3.5
3.5-5.5
8.7-8.9
6.3-
12.4
n.a.
Perform.
Large
Car
V8
0.5
1.0-2.0
1.0-3.0
1.0-3.0
2.5-3.0
1.0-2.0
2.0-3.0
1.0-3.0
1.5-3.5
0.0-0.5
3.9-5.5
1.0-1.5
0.5-2.5
1.0-2.5
2.0-3.0
2.0-2.5
1.8-1.9
3.5-4.0
10.8-
11.7
5.3-6.9
0.5
1.5-2.5
n.a.
1.4-3.4
2.7-4.1
1.0-2.0
1.0-2.0
3.0-4.0
3.0-5.0
8.7-8.9
6.3-
12.4
n.a.
3-87
-------
Draft Joint Technical Support Document
Plug-in Hybrid
Mass Reduction (1.5%)
Mass Reduction (3.5 to 8.5%)
Low Rolling Resistance Tires
Low Drag Brakes
Secondary Axle Disconnect
Aero Drag Reduction
PHEV
MSI
MS2
ROLL
LDB
SAX
AERO
45.2-
47.7
1.0
2.3
n.a.
n.a.
1.0-1.5
n.a.
45.2-
47.7
1.0
2.3
n.a.
n.a.
1.0-1.5
n.a.
45.2-
47.7
1.0
3.9
n.a.
n.a.
1.0-1.5
n.a.
n.a.
1.0
5.6
n.a.
0.5-
1.0
1.0-1.5
n.a.
Table 3-25. Component Technology Incremental Effectiveness Estimates for Volpe model, Light
Trucks
VEHICLE TECHNOLOGY INCREMENTAL FUEL CONSUMPTION
REDUCTION (-%) BY VEHICLE SUBCLASS - LIGHT TRUCKS
Nominal Baseline Engine (For Cost Basis)
Low Friction Lubricants
Engine Friction Reduction
VVT - Coupled Cam Phasing (CCP) on
SOHC
Discrete Variable Valve Lift (DVVL) on
SOHC
Cylinder Deactivation on SOHC
VVT - Intake Cam Phasing (ICP)
VVT - Dual Cam Phasing (DCP)
Discrete Variable Valve Lift (DVVL) on
DOHC
Continuously Variable Valve Lift (CVVL)
Cylinder Deactivation on DOHC
Cylinder Deactivation on OHV
VVT - Coupled Cam Phasing (CCP) on
OHV
Discrete Variable Valve Lift (DVVL) on
OHV
Conversion to DOHC with DCP
Stoichiometric Gasoline Direct Injection
(GDI)
Combustion Restart
Turbocharging and Downsizing
Exhaust Gas Recirculation (EGR) Boost
Conversion to Diesel following CBRST
Conversion to Diesel following TRBDS
6-Speed Manual/Improved Internals
Improved Auto. Trans. Controls/Externals
Continuously Variable Transmission
LUB
EFR
CCPS
DVVLS
DEACS
ICP
DCP
DVVLD
CVVL
DEACD
DEACO
CCPO
DVVLO
CDOHC
SGDI
CBRST
TRBDS
EGRB
DSLC
DSLT
6MAN
IATC
CVT
Minivan
LT
V6
0.5
1.0-2.0
1.0-3.0
1.0-3.0
2.5-3.0
1.0-2.0
2.0-3.0
1.0-3.0
1.5-3.5
0-0.5
3.9-5.5
1.0-1.5
0.5-2.5
1.0-2.5
2.0-3.0
2.0-2.5
1.8-1.9
3.5-4.0
10.8-
11.7
5.3-6.9
0.5
1.5-2.5
0.7-2.0
Small
LT
Inline 4
0.5
1.0-2.0
1.0-3.0
1.0-3.0
n.a.
1.0-2.0
2.0-3.0
1.0-3.0
1.5-3.5
n.a.
n.a.
1.0-1.5
0.5-2.5
1.0-2.5
2.0-3.0
2.0-2.5
4.2-4.8
3.5-4.0
n.a.
n.a.
0.5
1.5-2.5
0.7-2.0
Midsize
LT
V6
0.5
1.0-2.0
1.0-3.0
1.0-3.0
2.5-3.0
1.0-2.0
2.0-3.0
1.0-3.0
1.5-3.5
0.0-0.5
3.9-5.5
1.0-1.5
0.5-2.5
1.0-2.5
2.0-3.0
2.0-2.5
1.8- 1.9
3.5-4.0
10.8-
11.7
5.3-6.9
0.5
1.5-2.5
0.7-2.0
Large
LT
V8
0.5
1.0-2.0
1.0-3.0
1.0-3.0
2.5-3.0
1.0-2.0
2.0 - 3.0
1.0-3.0
1.5-3.5
0.0 - 0.5
3.9-5.5
1.0-1.5
0.5-2.5
1.0-2.5
2.0 - 3.0
2.0 - 2.5
1.8- 1.9
3.5-4.0
10.8-
11.7
5.3-6.9
0.5
1.5-2.5
n.a.
3-88
-------
6/7/8-Speed Auto. Trans with Improved
Internals
Dual Clutch or Automated Manual
Transmission
Electric Power Steering
Improved Accessories
12V Micro-Hybrid
Belt mounted Integrated Starter Generator
Crank mounted Integrated Starter Generator
Power Split Hybrid
2-Mode Hybrid
Plug-in Hybrid
Mass Reduction (1.5%)
Mass Reduction (3.5 to 8.5%)
Low Rolling Resistance Tires
Low Drag Brakes
Secondary Axle Disconnect
Aero Drag Reduction
NAUTO
DCTAM
EPS
IACC
MHEV
BISG
CISG
PSHEV
2MHEV
PHEV
MSI
MS2
ROLL
LDB
SAX
AERO
1.4-3.4
2.7-4.1
1.0-2.0
1.0-2.0
2.5-3.5
3.5-5.5
8.7-8.9
6.3-
12.4
n.a.
n.a.
1.0
5.6
1.0-2.0
0.5-1.0
1.0-1.5
2.0-3.0
1.4-3.4
2.7-4.1
1.0-2.0
1.0-2.0
2.0-3.0
4.0-6.0
8.6-8.9
6.3-
12.4
3.0-7.3
45.2-
47.7
1.0
3.9
1.0-2.0
n.a.
1.0-1.5
2.0-3.0
1.4-3.4
2.7-4.1
1.0-2.0
1.0-2.0
2.5-3.5
3.5-5.5
8.7-8.9
6.3-
12.4
3.0-7.2
n.a.
1.0
3.9
1.0-2.0
0.5-1.0
1.0-1.5
2.0-3.0
1.4-3.4
2.7-4.1
1.0-2.0
n.a.
n.a.
n.a.
14.1-
16.3
n.a.
4.1-9.5
n.a.
1.0
5.6
1.0-2.0
0.5 - 1.0
1.0-1.5
2.0 - 3.0
3.5.2 EPA cost and effectiveness tables
The tables below show costs and effectiveness values for each particular
technology described throughout this chapter (those considered by EPA in its model).
The costs are shown as incremental to a particular baseline as noted in the tables. In
general, the costs are meant to reflect the cost of a given technology relative to a baseline
vehicle or engine with fixed timing 4 valves per cylinder, port fuel injection and a 4
speed automatic transmission. Where this is not the case is made clear in the tables. The
same is true of the effectiveness tables below. As such, the costs and effectiveness values
shown in the tables below will differ from the values contained in the NHTSA tables due
to the different ways that the agency's models handles them. However, the underlying
cost and effectiveness values used by each agency in their respective analyses are
consistent.
3-89
-------
Draft Joint Technical Support Document
Table 3-26 EPA's Incremental Piece Costs for Engine Technologies Marked up to include both Direct and
Indirect Costs in 2016 (2007 Dollars per Vehicle)
TECHNOLOGY
c«
-------
Convert to V8 DOHC
Gasoline HCCI dual-
mode
Diesel - Lean NOx trap
Diesel - urea SCR
V80HV
GDI - stolen
Base gasoline engine
Base gasoline engine
n/a
$253
$509
$375
$2,655
$509
$375
$2,164
$509
$375
$2,164
$509
$659
$2,961
Table 3-27 EPA's Incremental Piece Costs for Transmission Technologies Marked up to include both Direct and
Indirect Costs in 2016 (2007 Dollars per Vehicle)
TECHNOLOGY
Aggressive shift logic
Early torque converter lockup
5-speed automatic
6-speed automatic
6-speed DCT - dry clutch
6-speed DCT - wet clutch
6-speed manual
CVT
INCREMENTAL TO
Base trans
Base trans
4-speed auto trans
4-speed auto trans
6-speed auto trans
6-speed auto trans
5-speed manual trans
4-speed auto trans
VEHICLE CLASS
Small
Car
$28
$25
$90
$150
$65
$139
$79
$192
Large
Car
$28
$25
$90
$150
$65
$139
$79
$224
Minivan
$28
$25
$90
$150
$65
$139
$79
$224
Small
Truck
$28
$25
$90
$150
$65
$139
$79
n/a
Large
Truck
$28
$25
$90
$150
$65
$139
$79
n/a
Table 3-28 EPA's Incremental Piece Costs for Hybrid Technologies Marked up to include both Direct and
Indirect Costs in 2016 (2007 Dollars per Vehicle)
TECHNOLOGY
Stop-Start
IMA/ISA/BSG
(includes engine
downsize)
2-Mode hybrid
electric vehicle
Power-split
hybrid electric
vehicle
Plug-in hybrid
electric vehicle
Plug-in hybrid
electric vehicle
Full electric
vehicle
INCREMENTAL
TO
Base engine &
trans
Base engine &
trans
Base engine &
trans
Base engine &
trans
IMA/ISA/BSG
hybrid
Power-split
hybrid
Base engine &
trans
VEHICLE CLASS
Small Car
$351
$2,854
$4,232
$3,967
$6,922
$5,423
$27,628
Large Car
$398
$3,612
$5,469
$5,377
$9,519
$7,431
n/a
Minivan
$398
$3,627
$5,451
$5,378
$9,598
$7,351
n/a
Small Truck
$398
$3,423
$4,943
$4,856
$9,083
$7,128
n/a
Large Truck
$437
$4,431
$7,236
$7,210
$12,467
$9,643
n/a
3-91
-------
Draft Joint Technical Support Document
Table 3-29 EPA's Incremental Piece Costs for Accessory Technologies Marked up to include both Direct and
Indirect Costs in 2016 (2007 Dollars per Vehicle)
TECHNOLOGY
Improved high efficiency
alternator &
electrification of
accessories
Upgrade to 42 volt
electrical system
Electric power steering
(12 or 42 volt)
INCREMENTAL
TO
Base accessories
1 2 volt electrical
system
Base power
steering
VEHICLE CLASS
Small Car
$76
$86
$94
Large Car
$76
$86
$94
Minivan
$76
$86
$94
Small Truck
$76
$86
$94
Large Truck
$76
$86
$94
Table 3-30 EPA's Incremental Piece Costs for Vehicle Technologies Marked up to include both Direct and
Indirect Costs in 2016 (2007 Dollars per Vehicle)
TECHNOLOGY
Aero drag reduction (20%
on cars, 10% on trucks)
Low rolling resistance
tires
Low drag brakes (ladder
frame only)
Secondary axle disconnect
(unibody only)
Front axle disconnect
(ladder frame only)
INCREMENTAL
TO
Base vehicle
Base tires
Base brakes
Base vehicle
Base vehicle
VEHICLE CLASS
Small Car
$42
$6
n/a
$514
n/a
Large Car
$42
$6
n/a
$514
n/a
Minivan
$42
$6
n/a
$514
n/a
Small Truck
$42
$6
$63
$514
$84
Large Truck
$42
$6
$63
n/a
$84
Table 3-31 through Table 3-35 summarize the estimates for the CC>2 reduction
estimates of various technologies which can be applied to cars and light-duty trucks.
3-92
-------
Table 3-31 Engine Technology Effectiveness
Technology
Low friction lubricants - incremental to base engine
Engine friction reduction - incremental to base engine
Absolute CO2 Reduction (% from baseline vehicle)
Small Car
0.5
1-3
Large
Car
0.5
1-3
Minivan
0.5
1-3
Small
Truck
0.5
1-3
Large
Truck
0.5
1-3
Overhead Cam Branch
VVT - intake cam phasing
VVT - coupled cam phasing
VVT - dual cam phasing
Cylinder deactivation (includes imp. oil pump,
if applicable)
Discrete VVLT
Continuous VVLT
2
3
3
n.a.
4
5
1
4
4
6
3
6
1
2
2
6
3
4
1
3
2
6
4
5
2
4
4
6
4
5
Overhead Valve Branch
Cylinder deactivation (includes imp. oil
pump, if applicable)
VVT - coupled cam phasing
Discrete VVLT
Continuous VVLT (includes conversion to
Overhead Cam)
n.a.
3
4
5
6
4
4
6
6
2
3
4
6
3
4
5
6
4
4
5
Camless valvetrain (electromagnetic)
Gasoline Direct Injection-stoichiometric (GDI-S)
Gasoline Direct Injection-lean burn (incremental to
GDI-S)
Gasoline HCCI dual-mode (incremental to GDI-S)
Turbo+downsize (incremental to GDI-S)
Diesel - Lean NOx trap[]*
Diesel -urea SCR []*
5-15
1-2
8-10
10-12
5-7
15-26
15-26
5-15
1-2
9-12
10-12
5-7
15-26
15-26
5-15
1-2
9-12
10-12
5-7
15-26
15-26
5-15
1-2
9-12
10-12
5-7
15-26
15-26
5-15
1-2
10-14
10-12
5-7
15-26
15-26
Table 3-32 Transmission Technology Effectiveness
Technology
5-speed automatic (from 4-speed auto)
Aggressive shift logic
Early torque converter lockup
6-speed automatic (from 4-speed auto)
6-speed AMT (from 4-speed auto)
6-speed manual (from 5-speed manual)
CVT (from 4-speed auto)
Absolute CO2 Reduction (% from baseline vehicle)
Small
Car
2.5
1-2
0.5
4.5-6.5
9.5-14.5
0.5
6
Large
Car
2.5
1-2
0.5
4.5-6.5
9.5-14.5
0.5
6
Minivan
2.5
1-2
0.5
4.5-6.5
9.5-14.5
0.5
6
Small
Truck
2.5
1-2
0.5
4.5-6.5
9.5-14.5
0.5
n.a.
Large
Truck
2.5
1-2
0.5
4.5-6.5
9.5-14.5
0.5
n.a.
3-93
-------
Draft Joint Technical Support Document
Table 3-33 Hybrid Technology Effectiveness
Technology
Stop-Start with 42 volt system
IMA/IS A/BSG (includes engine downsize)
2-Mode hybrid electric vehicle
Power-split hybrid electric vehicle
Full-Series hydraulic hybrid
Plug-in hybrid electric vehicle
Full electric vehicle (EV)
Absolute CO2 Reduction (% from baseline vehicle)
Small
Car
7.5
30
n.a.
35
40
58
100
Large
Car
7.5
25
40
35
40
58
100
Minivan
7.5
20
40
35
40
58
n.a.
Small
Truck
7.5
20
40
35
40
58
n.a.
Large
Truck
7.5
20
25
n.a.
30
47
n.a.
Table 3-34 Accessory Technology Effectiveness
Technology
Improved high efficiency alternator & electrification of
accessories (12 volt)
Electric power steering (12 or 42 volt)
Improved high efficiency alternator & electrification of
accessories (42 volt)
Absolute CO2 Reduction (% from baseline vehicle)
Small
Car
1-2
1.5
2-4
Large
Car
1-2
1.5-2
2-4
Minivan
1-2
2
2-4
Small
Truck
1-2
2
2-4
Large
Truck
1-2
2
2-4
Table 3-35 Other Vehicle Technology Effectiveness
Technology
Aero drag reduction (20% on cars, 10% on trucks)
Low rolling resistance tires (10%)
Low drag brakes (ladder frame only)
Secondary axle disconnect (unibody only)
Front axle disconnect (ladder frame only)
Absolute CO2 Reduction (% from baseline vehicle)
Small
Car
3
1-2
n.a.
1
n.a.
Large
Car
3
1-2
n.a.
1
n.a.
Minivan
3
1-2
n.a.
1
n.a.
Small
Truck
2
1-2
1
1
1.5
Large
Truck
2
n.a.
1
n.a.
1.5
3-94
-------
Appendix 3.A
Gross Domestic Product Adjustments Used to Adjust 2002-2006 Dollars
to 2007 Dollars
Throughout this analysis, the incremental cost estimates have been presented in
terms of 2007 dollars. However, most of the data sources relied upon as the basis for the
estimates have presented costs in terms of 2006 dollars. To convert the 2006 dollars to
2007 dollars, the "chain-type" Gross Domestic Product (GDP) deflator was used. Table
3.A-1 shows the GDP deflator data used for this analysis.
Table 3.A-1 Price Indexes for Gross Domestic Product
LINE
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Gross domestic product
Personal consumption
expenditures
Durable goods
Nondurable goods
Services
Gross private domestic
investment
Fixed investment
Nonresidential
Structures
Equipment and software
Residential
Change in private inventories
Net exports of goods and
services
Exports
Goods
Services
Imports
Goods
Services
Government consumption
expenditures and gross
investment
Federal
National defense
Nondefense
State and local
2002
104.193
103.542
95.766
102.089
106.018
101.64
101.66
99.513
110.03
95.956
107.24
99.273
98.661
100.762
96.341
95.288
101.878
105.507
105.631
105.792
105.345
105.435
2003
106.409
105.597
92.366
104.145
109.379
103.191
103.313
99.591
113.872
94.912
112.372
101.429
100.64
103.333
99.685
98.063
108.291
109.849
110.094
110.751
108.898
109.712
2004
109.462
108.392
90.696
107.626
112.929
106.686
106.845
100.896
120.912
94.6
120.587
104.997
104.323
106.632
104.526
102.915
113.049
114.754
115.322
115.932
114.218
114.431
2005
113.039
111.581
89.984
111.606
116.7
111.381
111.638
103.829
135.177
94.534
129.268
108.814
107.536
111.855
111.154
109.568
119.496
121.47
120.834
121.944
118.744
121.862
2006
116.676
114.675
88.772
114.984
120.752
116.102
116.38
107.277
151.822
94.594
136.897
112.618
111.131
116.156
115.932
114.171
125.257
127.239
125.806
127.381
122.803
128.109
2007
119.819
117.659
87.154
118.407
124.712
117.735
117.995
108.739
157.662
94.87
138.884
116.586
115.062
120.211
120.168
118.326
129.928
132.941
130.076
131.874
126.636
134.671
2008
122.5
121.596
86.071
124.666
128.752
118.509
118.727
110.513
163.298
95.485
135.587
122.788
121.147
126.684
132.698
131.408
139.217
139.234
134.289
136.574
129.869
142.219
Source: Bureau of Economic Analysis, U.S. Department of Commerce,
http://www.bea.gov/national/nipaweb/Table View. asp?SelectedTable=4
Using Line 1 in the table, 2006 dollars can be adjusted to 2007 dollars by
multiplying the ratio 119.8/116.7, or 1.027, by any value expressed in 2006 dollars. For
example, $1 in 2006 dollars would be $1 x 1.027, or $1.03 in 2007 dollars.
3-95
-------
Draft Joint Technical Support Document
References
1 "EPA Staff Technical Report: Cost and Effectiveness Estimates of Technologies Used to Reduce Light-
duty Vehicle Carbon Dioxide Emissions," EPA420-R-08-008, March 2008.
2 "Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards," National Research
Council, National Academy of Sciences, 2002.
3 "Reducing Greenhouse Gas Emissions from Light-Duty Motor Vehicles," Northeast States Center for a
Clean Air Future, September 2004.
4 "Staff Report: Initial Statement of Reasons for Proposed Rulemaking," California Environmental
Protection Agency, Air Resources Board, Regulations to Control Greenhouse Gas Emissions from Motor
Vehicles, August 6, 2004.
5 "Technology to Improve the Fuel Economy of Light Duty Trucks to 2015," Energy and Environmental
Analysis, Inc., May 2006.
6 "Variable Costs of Fuel Economy Technologies," prepared for The Alliance of Automobile
Manufacturers, June 1, 2008; and, "Variable Costs of Fuel Economy Technologies," prepared for The
Alliance of Automobile Manufacturers, June 1, 2008, Amended December 10, 2008.
7 National Highway Traffic Safety Administration, Request for Product Plan Information, 70 FR 51466,
August 30, 2005.
8 U.S. Environmental Protection Agency, "Draft Report - Light-Duty Technology Cost Analysis Pilot
Study," Contract No. EP-C-07-069, Work Assignment 1-3, September 3, 2009.
9 Memorandum from Don Kopinski, U.S. EPA, to docketEPA-HQ-OAR-2009-0472, dated September 11,
2009.
10 Memorandum from Don Kopinski, U.S. EPA, to docketEPA-HQ-OAR-2009-0472, dated September 11,
2009.
11 NHTSA's 2011 CAFE final rule at 74 FR 14196.
12 "EPA Staff Technical Report: Cost and Effectiveness Estimates of Technologies Used to Reduce Light-
duty Vehicle Carbon Dioxide Emissions," EPA420-R-08-008, March 2008.
13 National Income and Product Accounts Table 1.1.4: Price Indexes for Gross Domestic Product;
downloaded on 6/25/2009 At 12:00:25 PM, Last Revised June 25, 2009; Bureau of Economic Analysis,
U.S. Department of Commerce, www.bea.gov.
14 "EPA Staff Technical Report: Cost and Effectiveness Estimates of Technologies Used to Reduce Light-
Duty Vehicle Carbon Dioxide Emissions," EPA420-R-08-008, March 2008.
15 U.S. Environmental Protection Agency, "Draft Report - Light-Duty Technology Cost Analysis Pilot
Study," Contract No. EP-C-07-069, Work Assignment 1-3, September 3, 2009.
3-96
-------
16 Rogozhin, Alex, Michael Gallaher, and Walter McManus, "Automobile Industry Retail Price Equivalent
and Indirect Cost Multipliers," EPA 420-R-09-003, February 2009, http://epa.gov/otaq/ld-
hwy/420r09003.pdf.
17 "Memorandum: Documentation of the Development of Indirect Cost Multipliers for Three Automotive
Technologies," Gloria Helfand and Todd Sherwood, Office of Transportation and Air Quality, U.S.
Environmental Protection Agency, August 2009.
18 "Final Regulatory Analysis: Control of Emissions from Nonroad Diesel Engines," EPA420-R-04-007,
May 2004; "Regulatory Impact Analysis: Control of Emissions of Air Pollution from Locomotive Engines
and Marine Compression-Ignition Engines Less than 30 Liters per Cylinder," EPA420-D-08-001a, May
2008; NHTSA 2011 CAFE final rule at 74 FR 14196.
19 "Learning Curves in Manufacturing," Linda Argote and Dennis Epple, Science, February 23, 1990, Vol.
247, pp. 920-924.
20 NHTSA 2011 CAFE final rule at 74 FR 14196; "EPA Staff Technical Report: Cost and Effectiveness
Estimates of Technologies Used to Reduce Light-Duty Vehicle Carbon Dioxide Emissions," EPA420-R-
08-008, March 2008; "Final Regulatory Analysis: Control of Emissions from Nonroad Diesel Engines,"
EPA420-R-04-007, May 2004; "Regulatory Impact Analysis: Control of Emissions of Air Pollution from
Locomotive Engines and Marine Compression-Ignition Engines Less than 30 Liters per Cylinder,"
EPA420-D-08-001a, May 2008.
21 "Impact of Friction Reduction Technologies on Fuel Economy," Fenske, G. Presented at the March 2009
Chicago Chapter Meeting of the 'Society of Tribologists and Lubricated Engineers' Meeting, March 18th,
2009. Available at: http://www.chicagostle.org/program/2008-
2009/Impact%20of%20Friction%20Reduction%20Technologies%20on%20Fuel%20Economy%20-
%20with%20VGs%20removed.pdf (last accessed July 9, 2009).
22 "Light-Duty Automotive Technology and Fuel Economy Trends: 1975 Through 2007", EPA420-S-07-
001, September 2007. Available at http://www.epa.gov/oms/cert/mpg/fetrends/fetrends-archive.htm (last
accessed July 9, 2009).
23 Paul Whitaker, Ricardo, Inc., "Gasoline Engine Performance And Emissions - Future Technologies and
Optimization," ERC Symposium, Low Emission Combustion Technologies for Future 1C Engines,
Madison, WI, June 8-9, 2005. Available at
http://www.erc.wisc.edu/symposiums/2005_Symposium/June%208%20PM/Whitaker_Ricardo.pdf (last
accessed Nov. 9, 2008).
24 Stefan Trampert, FEV Motorentechnik GmbH, "Engine and Transmission Development Trends - Rising
Fuel Cost Pushes Technology," Symposium on International Automotive Technology, Pune, India, January
2007.
2 Stefan Wolff, Dirk Abendroth, Werner Weigl, Claus-Peter Linner, Rupert Neudecker, Michael
Schneider, Wolfgang Huber, and Andreas Rau, BMW, "Introducing The Automatic Start-Stop (ASS)
Function In Series Models," 7th Stuttgart Automotive Vehicle and Engine Symposium, Organised by
FKFS, Mar 2007, Vol. 1.
26 O.K. Fraidl, P.E. Kapus, and H. Friedl, AVL List GmbH, "Future Gasoline Engine Technologies for 130
g/Km CO2," VKM-THD 11th Symposium on the Working Process of Combustion Engines, TU Graz,
Sept. 2007.
27 "Development and Optimization of the Ford 3.5L V6 EcoBoost Combustion System," Yi,J., Wooldridge,
S., Coulson, G., Hilditch, J. Iyer, C.O., Moilanen, P., Papaioannou, G., Reiche, D. Shelby, M.,
3-97
-------
Draft Joint Technical Support Document
VanDerWege, B., Weaver, C. Xu, Z., Davis, G., Hinds, B. Schamel, A. SAE Technical Paper No. 2009-01-
1494, 2009.
28 David Woldring and Tilo Landenfeld of Bosch, and Mark J. Christie of Ricardo, "DI Boost: Application
of a High Performance Gasoline Direct Injection Concept," SAE 2007-01-1410. Available at
http://www.sae.org/technical/papers/2007-01-1410 (last accessed Nov. 9, 2008)
29 Yves Boccadoro, Loiic Kermanac'h, Laurent Siauve, and Jean-Michel Vincent, Renault Powertrain
Division, "The New Renault TCE 1.2L Turbocharged Gasoline Engine," 28th Vienna Motor Symposium,
April 2007.
30 Tobias Heiter, Matthias Philipp, Robert Bosch, "Gasoline Direct Injection: Is There a Simplified, Cost-
Optimal System Approach for an Attractive Future of Gasoline Engines?" AVL Engine & Environment
Conference, September 2005.
31 U.S. Environmental Protection Agency, "Draft Report - Light-Duty Technology Cost Analysis Pilot
Study," Contract No. EP-C-07-069, Work Assignment 1-3, September 3, 2009.
32 Docket No. NHTSA-2008-0089-0193.1
33 Cairns et al, Lotus, "Low Cost Solutions for Improved Fuel Economy in Gasoline Engines," Global
Powertrain Congress September 27-29, 2005, vol. 33. Available at http://www.gpc-
icpem.org/pages/publications.html (last accessed Nov. 9, 2008).
34 Tim Lake, John Stokes, Richard Murphy, and Richard Osborne of Ricardo and Andreas Schamel of
Ford-Werke, "Turbocharging Concepts for Downsized DI Gasoline Engines," VKA/ika Aachen
Colloquium 2003. Available at http://cat.inist.fr/?aModele=afficheN&cpsidt=16973598 (last accessed
Nov. 9, 2008).
35 "Interim Report: New Powertrain Technologies and Their Projected Costs," October 2005, EPA420-R-
05-012.
36 "Cost and Fuel Economy Comparison of Diesel and Gasoline Powertrains in Passenger Cars and Light
Trucks," submitted by FEV Engine Technology, Inc., April 23, 2003, contained as Appendix I within EPA
Interim Technical Report EPA420-R-04-002.
37 EPA's 2007/2010 diesel heavy-duty highway final rule at66 FR 5002; EPA's Locomotive and Marine
final rule at 73 FR 37096.
38
http: //www .platinum, matthey. com/.
39 General Motors, news release, "From Hybrids to Six-Speeds, Direct Injection And More, GM's 2008
Global Powertrain Lineup Provides More Miles with Less Fuel" (released Mar. 6, 2007). Available at
http://www.gm.com/experience/fuel_economy/news/2007/adv_engines/2008-powertrain-lineup-082707.jsp
(last accessed Sept. 18, 2008).
40 "EPA Staff Technical Report: Cost and Effectiveness Estimates of Technologies Used to Reduce Light-
duty Vehicle Carbon Dioxide Emissions" Environmental Protection Agency, EPA420-R-08-008, March
2008, at page 17.
41 "Latest Chevrolet Volt Battery Pack and Generator Details and Clarifications." Lyle Dennis interview
of Rob Peterson (GM) regarding the all-electric drive range of the GM Volt, August 29, 2007. Accessed on
the Internet on June 30, 2009 at: http://gm-volt.com/2007/08/29/latest-Chevrolet-volt-battery-pack-and-
generator-details-and-clarifications/
3-98
-------
42 "Active Combination of Ultracapacitorsand Batteries for PHEV ESS." Bohn, T. U.S. Department of
Energy 2009 Vehicle Technologies Merit Review, May 20, 2009
43 Anderman, Advanced Automotive Battery Conference, May 2008. Proceedings available for purchase at
http://www.advancedautobat.com/Proceedings/index.html (last accessed October 17, 2008).
44 74 FR 14291 (Mar. 30, 2009)
45 "Electric Cars: Plugged In, Batteries must be included," Deutsche Bank Global Markets Research
Company, June 9, 2008.
46 74 FR 14293 (Mar. 30, 2009)
47 http://automobiles.honda.com/insight-hybrid/features.aspx?Feature=ima (last accessed on September 10,
2009).
48 "Factor of Two: Halving the Fuel Consumption of New U.S. Automobiles by 2035," Cheah, L., Evans,
C., Bandivadekar, A., Heywood, J., Massachusetts Institute of Technology, Publication No. LFEE 2007-04
RP, October 2007.
49 «pY 2007 Progress Report for Lightweighting Materials", Davis, P., Sullivan, R., Carpenter, J., U.S.
Department of Energy Office of Vehicle Technologies, December 2008. Accessed on the Internet on April
30, 2009 at: http://wwwl.eere.energy.gov/vehiclesandfuels/resources/vt_lm_fy07.html
50 "Benefit Analysis: Use of Aluminum Structures in Conjunction with Alternative Powertrain
Technologies in Automobiles," Bull, M. Chavali, R., Mascarin, A., Aluminum Association Research
Report, May 2008. Accessed on the Internet on April 30, 2009 at:
http://www.autoaluminum.org/downloads/IBIS-Powertrain-Study.pdf
51 "Future Generation Passenger Compartment-Validation (ASP 241)" Villano, P.J., Shaw, J.R.,
Polewarczyk, J., Morgans, S., Carpenter, J.A., Yocum, A.D., in "Lightweighting Materials -FY 2008
Progress Report," U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, Vehicle
Technologies Program, May 2009.
52 "Preliminary Vehicle Mass Estimation Using Empirical Subsystem Influence Coefficients," Malen, D.E.,
Reddy, K. Auto-Steel Partnership Report, May 2007. Accessed on the Internet on May 30, 2009 at:
http://www.a-sp.org/database/custom/Mass%20Compounding%20-%20Final%20Report.pdf
53 "Future Generation Passenger Compartment - Final Report," Auto-Steel Partnership Report, June 2007.
Accessed on the Internet on May 30, 2009 at: http://www.a-sp.org/database/custom/FGPC%20Phase
%201 %20Final%20Report.pdf
54 "Basic Analysis of the Cost and Long-Term Impact of the Energy Independence and Security Act Fuel
Economy Standards" Sierra Research Report No. SR2008-04-01
55 "Factor of Two: Halving the Fuel Consumption of New U.S. Automobiles by 2035" L. Cheah, C. Evans,
A. Bandivadekar and J. Heywood, Publication No. LFEE 2007-04 RP, October 2007
56 "Tires and Passenger Vehicle Fuel Economy," Transportation Research Board Special Report 286,
National Research Council of the National Academies, 2006.
57 "An HCCI Engine Power Plant for a Hybrid Vehicle," Sun, R., R. Thomas and C. Gray, Jr., SAE
Technical Paper No. 2004-01-0933, 2004.
3-99
-------
Economic Assumptions Used in the Agencies' Analyses
CHAPTER 4: Economic Assumptions Used in the Agencies' Analyses
4.1 How the Agencies Use the Economic Assumptions in their Analyses
Improving new vehicles' fuel efficiency and reducing greenhouse gas (GHG) emissions
provides direct benefits to their buyers and users by reducing fuel consumption and fuel costs
throughout those vehicles' lifetimes, stimulating increased vehicle use through the fuel economy
rebound effect, and increasing vehicles' driving range so that they require less frequent refueling.
At the same time, the reduction in fuel use that results from requiring higher fuel economy and
reducing GHGs also produces wider benefits to the U.S. economy by lowering the cost of
economic externalities that result from U.S. petroleum consumption and imports, including
reducing the price of petroleum, lowering the potential costs from disruption in the flow of oil
imports, and possibly reducing federal outlays to secure imported oil supplies and cushion the
U.S. economy against their potential interruption. Reducing fuel consumption and GHGs also
lowers the economic costs of environmental externalities resulting from fuel production and use,
including reducing the impacts on human health impacts from emissions of criteria air pollutants,
and reducing future economic damages from potential changes in the global climate caused by
greenhouse gas emissions.
These benefits are partly offset by the increase in fuel use that results from added vehicle
use due to the fuel economy rebound effect, as well as by added costs from the increased
congestion, crashes, and noise caused by increased vehicle use. They might also be offset by any
loss in the utility that new vehicles provide to their buyers (and subsequent owners) as a
consequence of reductions in their performance, carrying capacity, or comfort that manufacturers
may implement as part of their strategies to comply with higher fuel economy requirements and
GHG standards. Nevertheless, the total economic benefits from requiring higher fuel economy
and reducing GHGs are likely to be substantial, and the EPA and NHTSA have attempted to
develop detailed estimates of the economic benefits from adopting more stringent standards.
This section discusses the common economic values used by both NHTSA and EPA.
These economic inputs incorporate a range of forecast information, economic estimates, and
input parameters. This section describes the sources that EPA and NHTSA have relied upon for
this information, the rationale underlying each assumption, and the agencies' proposed estimates
of specific parameter values. These common economic values are then used as inputs into each
Agencies' respective modeling and other analyses of the economic benefits and costs of the EPA
and NHTSA programs. While the underlying economic input values are common to both
Agencies, it is the differences in the way each Agency assesses its program, that result in
differing proposed benefits estimates. This issue is discussed further in Section l.C of the
preamble to the joint rulemaking.
4-1
-------
Draft Joint Technical Support Document
4.2 What Economic Assumptions Do the Agencies Use?
4.2.1 Potential Opportunity Costs of Improved Fuel Economy and Reduced GHG
Emissions
An important concern is whether achieving the fuel economy improvements required by
alternative fuel economy or CC>2 emissions standards would require manufacturers to
compromise the performance, carrying capacity, safety, or comfort of their vehicle models. If it
did so, the resulting sacrifice in the value of these attributes to consumers would represent an
additional cost of achieving the required improvements in fuel economy, and thus of
manufacturers' compliance with stricter standards. Although the exact values that vehicle buyers
attach to vehicle attributes such as fuel economy, performance, passenger- and cargo-carrying
capacity, and other dimensions of vehicle utility are difficult to infer from their purchasing
decisions and vehicle prices, changing vehicle attributes can significantly affect the overall
utility that vehicles provide to their owners, and thus their value to potential buyers.
EPA and NHTSA have approached this potential problem by developing proposed cost
estimates for fuel economy-improving technologies that include any additional manufacturing
costs that would be necessary to maintain the reference fleet levels of performance, comfort,
capacity, or safety of any light-duty vehicle model to which those technologies are applied. In
doing so, the agencies primarily followed the precedent established by the 2002 NAS Report,
although EPA and NHTSA have extensively updated their assumptions as necessary for the
purposes of the current rulemaking. The NAS study estimated "constant performance and
utility" costs for fuel economy technologies, and EPA and NHTSA have used these as the basis
for their continuing efforts to refine the technology costs they employ in analyzing
manufacturer's costs complying with alternative passenger car and light truck fuel efficiency and
GHG standards for MYs 2012-2016. Although the agencies have revised its estimates of
manufacturers' costs for some technologies significantly for use in this rulemaking, these revised
estimates are still intended to represent costs that would allow manufacturers to maintain the
performance, carrying capacity, and utility of vehicle models while improving their fuel
economy.
EPA and NHTSA acknowledge the difficulty of estimating technology costs that include
costs for the accompanying changes in vehicle design that are necessary to maintain
performance, capacity, and utility. We believe that our proposed cost estimates for fuel
economy-improving technologies should be generally sufficient to prevent significant reductions
in consumer welfare provided by vehicle models to which manufacturers apply those
technologies. However, it is possible that the agencies' technology cost estimates do not include
adequate allowance for the necessary efforts by manufacturers to maintain vehicle performance,
carrying capacity, and utility while improving fuel economy. If this is the case, the true
economic costs of achieving higher fuel economy would include the opportunity costs to vehicle
owners of any sacrifices in vehicles' performance, carrying capacity, and utility that
accompanied increases in their fuel economy. In that event, the agencies' estimated technology
costs would underestimate the true economic costs of achieving it.
Recognizing this possibility, EPA and NHTSA may consider in the future whether there
are credible methods available to estimate explicitly the changes in vehicle buyers' welfare from
4-2
-------
Economic Assumptions Used in the Agencies' Analyses
the combination of higher prices for new vehicle models, increases in their fuel economy, and
any accompanying changes in vehicle attributes such as performance, passenger- and cargo-
carrying capacity, or other dimensions of utility. The net change in buyer's welfare that results
from the combination of these changes would provide an estimate of the true economic costs for
improving fuel economy, which would fall on vehicle purchasers and subsequent owners rather
than on manufacturers. As EPA discusses in its DRIA in Section 8.1.2, though, EPA believes
that the state of the art for developing these estimates may not be sufficient to provide credible
estimates as indicated by wide variation in modeling methods, few assessments of the accuracy
or consistency of results, and inconsistent results when comparisons have been made. NHTSA,
in its PRIA in Section XX, also discusses its difficulties with implementing vehicle choice
models that, among other features, could estimate changes in consumer welfare. Although EPA
and NHTSA have continued to rely on estimates of what it believes are "constant performance
and utility" technology costs for this proposed analysis of the costs of complying with
alternative fuel efficiency standards for MY 20120-2016 cars and light trucks, the agencies seek
comment on alternative ways to deal with this important issue.
4.2.2 The On-Road Fuel Economy "Gap"
Actual fuel economy levels achieved by vehicles in on-road driving fall significantly
short of their levels measured under the laboratory-like test conditions used under the CAFE
program to establish its published fuel economy ratings for different models. In analyzing the
fuel savings from passenger car and light truck fuel efficiency and GHG standards, the agency
adjusts the actual fuel economy performance of each passenger car and light truck model
downward from its rated value to reflect the expected size of this on-road fuel economy "gap."
In December 2006, EPA adopted changes to its regulations on fuel economy labeling, which
were intended to bring vehicles' rated fuel economy levels closer to their actual on-road fuel
economy levels.1
Comparisons of on-road and CAFE fuel economy levels developed by EPA as part of its
Final Rule implementing new fuel economy labeling requirements for new vehicles indicate that
actual on-road fuel economy for light-duty vehicles averages 20 percent lower than published
fuel economy levels.2 For example, if the overall EPA fuel economy rating of a light truck is 20
mpg, the on-road fuel economy actually achieved by a typical driver of that vehicle is expected
to be 16 mpg (20*.80). In its analysis supporting the Final Rule establishing CAFE standards for
MY 2011, NHTSA employed EPA's revised estimate of this on-road fuel economy gap in its
analysis of the fuel savings resulting from alternative fuel efficiency standards.
An analysis conducted by NHTSA confirmed that EPA's estimate of a 20 percent gap
between test and on-road fuel economy is well-founded. The agency used data on the number of
passenger cars and light trucks of each model year that were in service (registered for use) during
each calendar year from 2000 through 2006, average fuel economy for passenger cars and light
trucks produced during each model year, and estimates of average miles driven per year by cars
and light trucks of different ages during each calendar year over that period. These data were
combined to develop estimates of the usage-weighted average fuel economy that the U.S.
passenger car and light truck fleets would have achieved during each year from 2000 through
2006 under test conditions.
4-3
-------
Draft Joint Technical Support Document
Table 4-1 compares NHTSA's estimates of fleet-wide average fuel economy under test
conditions for 2000 through 2006 to the Federal Highway Administration's (FHWA) published
estimates of actual on-road fuel economy achieved by passenger cars and light trucks during
each of those years. As it shows, FHWA's estimates of actual fuel economy for passenger cars
ranged from 21 percent to 23 percent lower than NHTSA's estimates of its fleet-wide average
value under test conditions over this period. Similarly, FHWA's estimates of actual fuel
economy for light trucks ranged from 16 percent to 18 percent lower than NHTSA's estimates of
average light truck fuel economy under test conditions. Thus, these results appear to confirm
that the 20 percent on-road fuel economy discount or gap represents a reasonable estimate for
use in evaluating the fuel savings likely to result from alternative fuel efficiency standards for
MY 2012-2016 vehicles.
We are aware of two potential issues involved in these estimates. One, the estimates of
total annual car and truck VMT are developed by the states and submitted to FHWA. Each state
uses its own definition of a car and a truck. For example, some states classify minivans as cars
and some as trucks. Thus, there are known inconsistencies with these estimates when evaluated
separately for cars and trucks. Also, total gasoline consumption can be reasonably estimated
from excise tax receipts, but separate estimates for cars and trucks are not available. We are not
aware of the precise methodology used to develop the distinct on-road fuel economy estimates
for cars and trucks developed by FHWA. We do not believe that they are based on direct
measurements from substantial numbers of vehicles, as no such test programs were found by
EPA during its fuel economy labeling rule in 2006. Also, the year-to-year consistency for both
car and truck fuel economy implies some methodology other than direct measurement. For this
reason, NHTSA and EPA are not proposing to use distinct on-road fuel economy gaps for cars
and trucks, but one common value of 20% for both vehicle classes for purposes of estimating the
fuel savings of the proposed standards.
Table 4-1 Estimated Fleet-Wide Fuel Economy of Passenger Cars and Light Trucks
Compared to Reported Fuel Economy
YEAR
2000
2001
2002
2003
2004
2005
2006
Avg.,
2000-
2006
PASSENGER CARS
NHTSA
Estimated
Test MPG
28.2
28.2
28.3
28.4
28.5
28.6
28.8
28.4
FHWA
Reported
Actual MPG
21.9
22.1
22.0
22.2
22.5
22.1
22.5
22.2
Percent
Difference
-22.2%
-21.7%
-22.3%
-21.9%
-21.1%
-22.8%
-21.8%
-22.0%
LIGHT-DUTY TRUCKS
NHTSA
Estimated
Test MPG
20.8
20.8
20.9
21.0
21.0
21.1
21.2
21.0
FHWA
Reported
Actual MPG
17.4
17.6
17.5
17.2
17.2
17.7
17.8
17.5
Percent
Difference
-16.3%
-15.5%
-16.2%
-18.0%
-18.3%
-16.3%
-16.2%
-16.7%
4-4
-------
Economic Assumptions Used in the Agencies' Analyses
4.2.3 Fuel Prices and the Value of Saving Fuel
Projected future fuel prices are a critical input into the preliminary economic analysis of
alternative fuel efficiency and GHG standards, because they determine the value of fuel savings
both to new vehicle buyers and to society. EPA and NHTSA relied on the most recent fuel price
projections from the U.S. Energy Information Administration's (EIA) Annual Energy Outlook
(AEO) for this analysis. Specifically, we used the AEO 2009 (April 2009 release) Reference
Case forecasts of inflation-adjusted (constant-dollar) retail gasoline and diesel fuel prices, which
represent the EIA's most up-to-date estimate of the most likely course of future prices for
petroleum products.3 While NHTSA relied on the forecasts of fuel prices presented in AEO
2008 High Price Case in the MY 2011 CAFE final rule, we noted at the time that we were
relying on that estimate primarily because volatility in the oil market appeared to have overtaken
the Reference Case, and that we anticipated that the Reference Case forecast would be
significantly higher in the next AEO. In fact, EIA's AEO 2009 Reference Case forecast projects
higher retail fuel prices in most future years than those forecast in the High Price Case from
AEO 2008. EPA and NHTSA are thus confident that the AEO 2009 Reference Case is an
appropriate forecast for projected future fuel prices.
Federal government agencies generally use EIA's projections in their assessments of
future energy-related policies. The retail fuel price forecasts presented in AEO 2009 span the
period from 2009 through 2030. Measured in constant 2007 dollars, the Reference Case forecast
of retail gasoline prices during calendar year 2020 is $3.62 per gallon, rising gradually to $3.82
by the year 2030 (these values include federal, state and local taxes). However, valuing fuel
savings over the maximum lifetimes of passenger cars and light trucks used in this analysis
requires fuel price forecasts that extend through 2050, approximately the last year during which a
significant number of MY 2016 vehicles will remain in service/ To obtain fuel price forecasts
for the years 2031 through 2050, the agency assumes that retail fuel prices will continue to
increase after 2030 at the average annual rates projected for 2020-2030 in the AEO 2009 Revised
Reference Case.b This assumption results in a projected retail price of gasoline that reaches
$4.25 in 2050.
The value of fuel savings resulting from improved fuel economy and GHG emissions to
buyers of light-duty vehicles is determined by the retail price of fuel, which includes federal,
state, and any local taxes imposed on fuel sales. Total taxes on gasoline, including federal, state,
and local levies averaged $0.42 per gallon during 2006, while those levied on diesel averaged
$0.50. Because fuel taxes represent transfers of resources from fuel buyers to government
agencies, however, rather than real resources that are consumed in the process of supplying or
a The agency defines the maximum lifetime of vehicles as the highest age at which more than 2 percent of those
originally produced during a model year remain in service. In the case of light-duty trucks, for example, this age has
typically been 36 years for recent model years.
b This projection uses the rate of increase in fuel prices for 2020-2030 rather than that over the complete forecast
period (2009-2030) because there is extreme volatility in the forecasts for the years 2009 through approximately
2020. Using the average rate of change over the complete 2009-2030 forecast period would result in projections of
declining fuel prices after 2030.
4-5
-------
Draft Joint Technical Support Document
using fuel, their value must be deducted from retail fuel prices to determine the value of fuel
savings resulting from more stringent fuel efficiency and GHG standards to the U.S. economy as
a whole.
In estimating the economy-wide or "social" value of fuel savings from improved fuel
efficiency standards and GHG emission standards, EPA and NHTSA follow the assumptions
used by EIA in AEO 2009 that state and local gasoline taxes will keep pace with inflation in
nominal terms, and thus remain constant when expressed in constant 2007 dollars.0 In contrast,
EIA assumes that federal gasoline taxes will remain unchanged in nominal terms, and thus
decline throughout the forecast period when expressed in constant 2007 dollars. These differing
assumptions about the likely future behavior of federal and state/local fuel taxes are consistent
with recent historical experience, which reflects the fact that federal as well as most state motor
fuel taxes are specified on a cents-per-gallon basis, and typically require legislation to change.
The projected value of total taxes is deducted from each future year's forecast of retail gasoline
and diesel prices reported in AEO 2009 to determine the economic value of each gallon of fuel
saved during that year as a result of improved fuel economy. Subtracting fuel taxes results in a
projected value for saving gasoline of $3.22 per gallon during 2020, rising to $3.45 per gallon by
the year 2030.
EIA is widely-recognized as an impartial and authoritative source of analysis and
forecasts of U.S. energy production, consumption, and prices. The agency has published annual
forecasts of energy prices and consumption levels for the U.S. economy since 1982 in its AEOs.
These forecasts have been widely relied upon by federal agencies for use in regulatory analysis
and for other purposes. Since 1994, EIA's annual forecasts have been based upon the agency's
National Energy Modeling System (NEMS), which includes detailed representation of supply
pathways, sources of demand, and their interaction to determine prices for different forms of
energy.
From 1982 through 1993, EIA's forecasts of world oil prices - the primary determinant
of prices for gasoline, diesel, and other transportation fuels derived from petroleum -
consistently overestimated actual prices during future years, often very significantly. Of the total
of 119 forecasts of future world oil prices for the years 1985 through 2005 that EIA reported in
its 1982-1993 editions of AEO, 109 overestimated the subsequent actual values for those years,
on average exceeding their corresponding actual values by 75 percent.
Since that time, however, EIA's forecasts of future world oil prices show a more mixed
record for accuracy. The 1994-2005 editions of AEO reported 91 separate forecasts of world oil
prices for the years 1995-2005, of which 33 have subsequently proven too high while the
remaining 58 have underestimated actual prices. The average absolute error (i.e., regardless of
c By "social" value of fuel savings, the agencies mean that we value fuel savings at pre-tax prices, since this
measures the value of the resources that are saved when less fuel is produced and consumed. The social value of
saving fuel includes this plus the reductions in economic and environmental externalities that result from consuming
less fuel, importing less petroleum, and so on.
4-6
-------
Economic Assumptions Used in the Agencies' Analyses
its direction) of these forecasts has been 21 percent, but over- and underestimates have tended to
offset one another, so that on average EIA's more recent forecasts have underestimated actual
world oil prices by 7 percent. Although both its overestimates and underestimates of future
world oil prices for recent years have often been large, the most recent editions of AEO have
significantly underestimated petroleum prices during those years for which actual prices are now
available.
However, EPA and NHTSA does not regard EIA's recent tendency to underestimate
future prices for petroleum and refined products or the high level of current fuel prices as
adequate justification to employ forecasts that differ from the Reference Case forecast presented
in the AEO 2009 Reference Case. This is particularly the case because this forecast has been
revised upward significantly since the initial release of AEO 2008, which in turn represented a
major upward revision from EIA's fuel price forecast reported previously in AEO 2007.
Comparing different forecasts of world oil prices also shows that EIA's Reference Case forecast
reported in AEO 2009 is among the highest of all six publicly-available forecasts of world oil
prices over the 2010-30 time horizon.4 Because world petroleum prices are the primary
determinant of retail prices for refined petroleum products such as transportation fuels, this
suggests that the Reference Case forecast of U.S. fuel prices reported in AEO 2009 is likely to be
the highest of those projected by major forecasting services.
4.2.4 Vehicle Survival and Use Assumptions
The agencies' analyses of fuel savings and related benefits from adopting higher fuel
efficiency standards and establishing GHG emission standards for MY 2012-2016 passenger cars
and light trucks is based on estimates of the resulting changes in fuel use over their entire
lifetimes in the U.S. vehicle fleet. The first step in estimating lifetime fuel consumption by
vehicles produced during a model year is to calculate the number of those vehicles that is
expected to remain in service during each future year after they are produced and sold.d This
number is calculated by multiplying the number of vehicles originally produced during a model
year by the proportion expected to remain in service at the age they will have reached during
each subsequent year, often referred to as a "survival rate."
In 2008, NHTSA updated its previous estimates of car and light truck survival rates using
the most current registration data for vehicles produced during recent model years, in order to
ensure that its forecasts of the number of these vehicles remaining in use during future years
reflect recent increases in the durability and expected life spans of cars and light trucks.5 Both
d Vehicles are defined to be of age 1 during the calendar year corresponding to the model year in which they are
produced; thus for example, model year 2000 vehicles are considered to be of age 1 during calendar year 2000, age
1 during calendar year 2001, and to reach their maximum age of 26 years during calendar year 2025. NHTSA
considers the maximum lifetime of vehicles to be the age after which less than 2 percent of the vehicles originally
produced during a model year remain in service. Applying these conventions to vehicle registration data indicates
that passenger cars have a maximum age of 26 years, while light trucks have a maximum lifetime of 36 years. See
Lu, S., NHTSA, Regulatory Analysis and Evaluation Division, "Vehicle Survivability and Travel Mileage
Schedules," DOT HS 809 952, 8-11 (January 2006). Available at http://www-nrd.nhtsa.dot.gov/Pubs/809952.pdf
(last accessed July 27, 2009).
4-7
-------
Draft Joint Technical Support Document
agencies used these updated survival estimates in their analyses of fuel savings and other benefits
from adopting improved fuel efficiency and GHG standards for MY 2012-2106 cars and light
trucks. The proportions of passenger cars and light trucks expected to remain in service at each
age up to their maximum lifetimes (26 and 36 years, respectively) are shown in Table 4-2 and
Table 4-3.e
The next step in estimating fuel use is to calculate the total number of miles that cars and
light trucks produced in each model year affected by the proposed fuel efficiency and GHG
standards will be driven during each year of their lifetimes. To estimate total miles driven, the
number of cars and light trucks projected to remain in use during each future year is multiplied
by the average number of miles they are expected to be driven at the age they will have reached
in that year.
Updated estimates of average annual miles driven by age were developed by NHTSA
from the Federal Highway Administration's 2001 National Household Transportation Survey,
and these also differ from the estimates of annual mileage employed in past NHTSA analyses/
Table 4-2 and Table 4-3 also report NHTSA's updated estimates of average car and light truck
use. The total number of miles driven by passenger cars or light trucks produced during a model
year, during each year of its lifetime, is estimated by multiplying these age-specific estimates of
average car and light truck use by the number of vehicles projected to remain in service during
that year.
As Table 4-2 and Table 4-3 also show, the resulting survival-weighted mileage over the
26-year maximum lifetime of passenger cars is 161,847 miles, while that over the 36-year
maximum lifetime of light trucks is 190,066 miles. Fuel savings and other benefits resulting
from improved fuel efficiency and GHG standards for passenger cars and light trucks are
calculated over their respective lifetimes and total expected mileage. It should be noted,
however, that survival-weighted mileage is extremely low (less than 1,000 miles per year) after
age 20 for cars and after age 25 for light trucks, and thus has little impact on lifetime fuel savings
or other benefits from higher fuel economy, particularly after discounting those benefits to their
present values.
The survival and annual mileage estimates reported in the tables reflect the convention
that vehicles are defined to be of age 1 during the calendar year that coincides with their model
year Thus for example, model year 2012 vehicles will be considered to be of age 1 during
e The maximum age of cars and light trucks was defined as the age when the number remaining in service has
declined to approximately two percent of those originally produced. Based on an examination of recent registration
data for previous model years, typical maximum ages appear to be 26 years for passenger cars and 36 years for light
trucks.
See also NHTSA, "Vehicle Survival and Travel Mileage Schedules," Office of Regulatory Analysis and
Evaluation, January 2006, pp. 15-17. The original source of information on annual use of passenger cars and light
trucks by age used in this analysis is the 2001 National Household Travel Survey (NHTS), jointly sponsored by the
Federal Highway Administration, Bureau of Transportation Statistics, and National Highway Traffic Safety
Administration.
4-8
-------
Economic Assumptions Used in the Agencies' Analyses
calendar year 2012. This convention is used in order to account for the fact that vehicles
produced during a model year typical are first offered for sale in June through September of the
preceding calendar year (for example, sales of a model year typically begin in June through
September of the previous calendar year, depending on manufacturer). Thus virtually all of the
vehicles produced during a model year will be in use for some or all of the calendar year
coinciding with their model year, and they are considered to be of age 1 during that year.8
g As an illustration, virtually the entire production of model year 2012 cars and light trucks will have been sold by
the end of calendar year 2012, so those vehicles are defined to be of age 1 during calendar year 2012. Model year
2012 vehicles are subsequently defined to be of age 2 during calendar year 2013, age 3 during calendar year 2014,
and so on. One complication arises because registration data are typically collected for July 1 of each calendar year,
so not all vehicles produced during a model year will appear in registration data until the calendar year when they
have reached age 2 (and sometimes age 3) under this convention.
-------
Draft Joint Technical Support Document
Table 4-2 Survival Rates and Unadjusted Annual Miles Traveled (VMT) by Age for
Passenger Cars
Vehicle
Age
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Estimated Survival
Fraction
0.9950
0.9900
0.9831
0.9731
0.9593
0.9413
0.9188
0.8918
0.8604
0.8252
0.7866
0.7170
0.6125
0.5094
0.4142
0.3308
0.2604
0.2028
0.1565
0.1200
0.0916
0.0696
0.0527
0.0399
0.0301
0.0227
Estimated Annual
VMT
14,231
13,961
13,669
13,357
13,028
12,683
12,325
11,956
11,578
11,193
10,804
10,413
10,022
9,633
9,249
8,871
8,502
8,144
7,799
7,469
7,157
6,866
6,596
6,350
6,131
5,940
Estimated Passenger Car Lifetime VMT
Survival-Weighted
Annual VMT
14,160
13,821
13,438
12,998
12,497
11,938
11,324
10,662
9,961
9,237
8,499
7,466
6,138
4,907
3,831
2,934
2,214
1,652
1,220
896
656
478
348
253
185
135
161,847
4-10
-------
Economic Assumptions Used in the Agencies' Analyses
Table 4-3 Survival Rates and Unadjusted Annual Vehicle-Miles Traveled (VMT)
by Age for Light Trucks
VEHICLE AGE
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
ESTIMATED
SURVIVAL
FRACTION
0.9950
0.9741
0.9603
0.9420
0.9190
0.8913
0.8590
0.8226
0.7827
0.7401
0.6956
0.6501
0.6042
0.5517
0.5009
0.4522
0.4062
0.3633
0.3236
0.2873
0.2542
0.2244
0.1975
0.1735
0.1522
0.1332
0.1165
0.1017
0.0887
0.0773
0.0673
0.0586
0.0509
0.0443
0.0385
0.0334
ESTIMATED
ANNUAL VMT
16,085
15,782
15,442
15,069
14,667
14,239
13,790
13,323
12,844
12,356
11,863
11,369
10,879
10,396
9,924
9,468
9,032
8,619
8,234
7,881
7,565
7,288
7,055
6,871
6,739
6,663
6,648
6,648
6,648
6,648
6,648
6,648
6,648
6,648
6,648
6,648
Estimated Lifetime Light Truck VMT
SURVIVAL-
WEIGHTED
ANNUAL VMT
16,004
15,374
14,829
14,195
13,479
12,691
11,845
10,960
10,053
9,145
8,252
7,391
6,573
5,735
4,971
4,281
3,669
3,131
2,665
2,264
1,923
1,635
1,393
1,192
1,026
887
774
676
590
514
447
390
338
294
256
222
190,066
4-11
-------
Draft Joint Technical Support Document
4.2.4.1 Adjusting Vehicle Use for Future Fuel Prices
The estimates of average annual miles driven by passenger cars and light trucks reported in
Table 4-2 and Table 4-3 reflect the historically low gasoline prices that prevailed at the time the 2001
NHTS was conducted. To account for the effect on vehicle use of subsequent increases in fuel prices,
the estimates of annual vehicle use derived from the NHTS were adjusted to reflect the forecasts of
future gasoline prices reported in the AEO 2009 Reference Case. This adjustment accounts for the
difference between the average price per gallon of fuel forecast for each year over the expected
lifetimes of model year 2012-2016 passenger cars and light trucks, and the average price that prevailed
when the NHTS was conducted in 2001. The elasticity of annual vehicle use with respect to fuel cost
per mile corresponding to the 10 percent fuel economy rebound effect used in this analysis (i.e., an
elasticity of -0.10) was applied to the percent difference between each future year's fuel prices and
those prevailing in 2001 to adjust the estimates of vehicle use derived from the NHTS to reflect the
effect of higher future fuel prices.
4.2.4.2 Ensuring Consistency with Growth in Total Vehicle Use
The estimates of annual miles driven by passenger cars and light trucks at each age were
also adjusted to reflect projected future growth in average vehicle use. Increases in the average
number of miles cars and trucks are driven each year have been an important source of historical
growth in total car and light truck use, and are expected to represent an important source future
growth in total light-duty vehicle travel as well. As an illustration of the importance of growth in
average vehicle use, the total number of miles driven by passenger cars increased 35 percent
from 1985 through 2005, equivalent to a compound annual growth rate of 1.5 percent.6 During
that time, however, the total number of passenger cars registered for in the U.S. grew by only
about 0.3 percent annually.11 Thus growth in the average number of miles automobiles are driven
each year accounted for the remaining 1.2 percent (= 1.5 percent - 0.3 percent) annual growth in
total automobile use.1 Further, the AEO 2009 Reference Case forecasts of total car and light
truck use and of the number of cars and light trucks in use suggest that their average annual use
will continue to increase gradually from 2010 through 2030.
In order to develop reasonable estimates of future growth in average car and light truck
use, NHTSA calculated the rate of growth in the mileage schedules shown in Table 4-4 and
Table 4-5 that would be necessary for total car and light truck travel to increase at the rate
forecast in the AEO 2009 Reference Case. This rate was calculated in a manner that is also
consistent with future changes in the overall size and age distributions of the U.S. passenger car
and light truck fleets that are implied by the agency's adjusted forecasts of total car and light
truck sales, together with the survival rates reported in Table 4-2 and Table 4-3. The growth rate
in average annual car and light truck use produced by this calculation is approximately 1.1
h A slight increase in the fraction of new passenger cars remaining in service beyond age 10 has accounted for a
small share of growth in the U.S. automobile fleet. The fraction of new automobiles remaining in service to various
ages was computed from R.L. Polk vehicle registration data for 1977 through 2005 by the agency's Center for
Statistical Analysis.
1 See supra note [2 above here]
4-12
-------
Economic Assumptions Used in the Agencies' Analyses
percent per year.j This rate was applied to the mileage figures reported in Table 4-2 and Table
4-3 to estimate annual mileage by age during each calendar year of the expected lifetimes of MY
2012-2016 cars and light trucks.
Table 4-4 and Table 4-5 report the results of applying the adjustments for both future fuel
prices and annual growth in car and light truck use to the figures reported previously. While the
adjustment for future fuel prices reduces average mileage at each age from the values shown
previously, the adjustment for expected future growth in average vehicle use increases it. As
Table 4-4 and Table 4-5 show, the net effect of these two adjustments is to increase expected
lifetime mileage significantly; for passenger cars, this figure rises to 190,971 miles from the
161,847 miles reported previously in Table 4-2 (or by 18 percent), while expected lifetime
mileage for light trucks increases from the 190,066 miles reported previously in Table 4-3 to
221,199 miles (16 percent).
Separate adjustments for projected fuel prices and growth in car and light truck use were
made for each calendar year from 2012 through 2030. The values reported in the tables below
are the averages of the adjusted values of annual car and light truck use by age for calendar years
2012-2030. These averages were used to analyze the impacts of improved fuel efficiency over
the lifetimes of MY 2012-2016 cars and light trucks. In contrast, the estimated fuel savings and
other impacts of improved fuel efficiency reported for individual calendar years used the
adjusted values of car and light truck use by age during those specific calendar years.
As previously stated, the estimates of survival-weighted mileage decline to less than
1,000 miles per year after age 20 for cars and after age 27 for light trucks. Thus they have
relatively little impact on lifetime fuel savings or other benefits from higher fuel economy,
particularly after discounting the benefits that occur in those distant future years to their present
values.
j It was not possible to estimate separate growth rates in average annual use for cars and light trucks, because of the
significant reclassification of light truck models as passenger cars discussed previously.
4-13
-------
Draft Joint Technical Support Document
Table 4-4 Survival Rates and Adjusted Annual Vehicle-Miles Traveled (VMT) by
Age for Passenger Cars
Vehicle Age
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Estimated Survival
Fraction
0.9950
0.9900
0.9831
0.9731
0.9593
0.9413
0.9188
0.8918
0.8604
0.8252
0.7866
0.7170
0.6125
0.5094
0.4142
0.3308
0.2604
0.2028
0.1565
0.1200
0.0916
0.0696
0.0527
0.0399
0.0301
0.0227
Estimated Annual
VMT
16,932
16,603
16,257
15,814
15,414
14,993
14,545
14,105
13,624
13,192
12,668
12,222
11,705
11,191
10,727
10,283
9,878
9,482
9,090
8,691
8,366
8,126
8,003
7,774
7,587
7,424
Adjusted Lifetime Passenger Car VMT
Survival-Weighted
Annual VMT
16,847
16,437
15,983
15,389
14,787
14,113
13,364
12,578
11,722
10,886
9,964
8,763
7,170
5,700
4,443
3,402
2,572
1,923
1,423
1,043
766
566
422
310
228
169
190,971
4-14
-------
Economic Assumptions Used in the Agencies' Analyses
Table 4-5 Survival Rates and Adjusted Annual Vehicle-Miles Traveled (VMT) by Age for
Light Trucks
Vehicle Age
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Estimated Survival
Fraction
0.9950
0.9741
0.9603
0.9420
0.9190
0.8913
0.8590
0.8226
0.7827
0.7401
0.6956
0.6501
0.6042
0.5517
0.5009
0.4522
0.4062
0.3633
0.3236
0.2873
0.2542
0.2244
0.1975
0.1735
0.1522
0.1332
0.1165
0.1017
0.0887
0.0773
0.0673
0.0586
0.0509
0.0443
0.0385
0.0334
Estimated Annual
VMT
18,847
18,408
18,050
17,575
17,142
16,593
16,095
15,493
14,891
14,336
13,689
13,160
12,554
11,945
11,342
10,822
10,383
9,900
9,433
9,033
8,692
8,499
8,246
8,261
8,066
8,066
8,101
8,098
8,096
8,095
8,093
8,092
8,086
8,080
8,064
8,050
Adjusted Lifetime Light Track VMT
Survival-Weighted
Annual VMT
18,752
17,931
17,333
16,556
15,753
14,790
13,826
12,745
11,655
10,610
9,522
8,555
7,585
6,590
5,681
4,894
4,218
3,597
3,053
2,595
2,210
1,907
1,629
1,433
1,228
1,074
944
824
718
626
545
474
412
358
310
269
221,199
4-15
-------
Draft Joint Technical Support Document
4.2.5 Accounting for the Fuel Economy Rebound Effect
The rebound effect refers to the fraction of fuel savings expected to result from an
increase in vehicle fuel economy that is offset by additional vehicle use. The increase in vehicle
use that stems from improved fuel economy occurs because vehicle owners respond to the
resulting reduction in vehicle fuel consumption and operating costs by driving slightly more.
The magnitude of the rebound effect is one of the determinants of the actual fuel savings
that are likely to result from adopting stricter fuel economy or emissions standards, and thus an
important parameter affecting EPA's and NHTSA's evaluation of alternative standards for future
model years. It can be measured directly by estimating the elasticity of vehicle use with respect
to fuel economy itself, or indirectly by the elasticity of vehicle use with respect to fuel cost per
mile driven.k When expressed as a positive percentage, either of these parameters gives the
fraction of fuel savings that would otherwise result from adopting stricter standards, but is offset
by the increase in fuel consumption that results when vehicles with increased fuel economy are
driven more.
The fuel economy rebound effect for light-duty vehicles has been the subject of a large
number of studies since the early 1980s. Although they have reported a wide range of estimates
of its exact magnitude, these studies generally conclude that a significant rebound effect occurs
when vehicle fuel efficiency improves.1 The most common approach to estimating its magnitude
has been to analyze household survey data on vehicle use, fuel consumption, fuel prices (often
obtained from external sources), and other determinants of household travel demand to isolate
the response of vehicle use to higher fuel economy. Other studies have relied on econometric
analysis of annual U.S. data on vehicle use, fuel economy, fuel prices, and other variables to
identify the response of total or average vehicle use to changes in fleet-wide average fuel
economy and its effect of fuel cost per mile driven. Two recent studies analyzed yearly variation
in vehicle ownership and use, fuel prices, and fuel economy among individual states over an
extended time period in order to measure the response of vehicle use to changing fuel economy.
This section surveys these previous studies, summarizes recent work on the rebound
effect, and explains the basis for the 10 percent rebound effect EPA and NHTSA are using in this
proposed rulemaking.7
k Fuel cost per mile is equal to the price of fuel in dollars per gallon divided by fuel economy in miles per gallon, so
this figure declines when a vehicle's fuel economy increases.
1 Some studies estimate that the long-run rebound effect is significantly larger than the immediate response to
increased fuel efficiency. Although their estimates of the adjustment period required for the rebound effect to reach
its long-run magnitude vary, this long-run effect is most appropriate for evaluating the fuel savings and emissions
reductions resulting from stricter standards that would apply to future model years.
m In effect, these studies treat U.S. states as a data "panel" by applying appropriate estimation procedures to data
consisting of each year's average values of these variables for the separate states.
4-16
-------
Economic Assumptions Used in the Agencies' Analyses
4.2.5.1 Summary of Historical Literature on Rebound Effect
It is important to note that a majority of the studies previously conducted on the rebound
effect rely on data from the 1950-1990s. While these older studies provide valuable information
on the potential magnitude of the rebound effect, studies that include more recent information
(e.g., data within the last decade) may provide more reliable estimates of how this proposal will
affect future driving behavior. Therefore, the more recent studies have been described in more
detail below.
Estimates based on aggregate U.S. vehicle travel data published by the U.S. Department
of Transportation, Federal Highway Administration, covering the period from roughly 1950 to
1990, have found long-run rebound effects on the order of 10-30 percent. Some of these studies
are summarized in the following table.
Table 4-6 Estimates of the Rebound Effect Using U.S. Aggregate Time-Series Data
on Vehicle Travel1
AUTHOR
(YEAR)
Mayo & Mathis
(1988)
Gately(1992)
Greene (1992)
Jones (1992)
Schimek(1996)
SHORT-
RUN
22%
9%
Linear 5-
19%
Log-linear
13%
13%
5-7%
LONG-RUN
26%
9%
Linear 5- 19%
Log-linear
13%
30%
21-29%
TIME PERIOD
1958-84
1966-88
1957-89
1957-89
1950-94
1 Source: Sorrell and Dimitropolous (2007) table 4.6.
Table 4-7 Estimates of the Rebound Effect Using U.S. State Level Data1
AUTHOR
(YEAR)
Haughton &
Sarkar(1996)
Small and Van
Dender (2007)
SHORT-RUN
9-16%
4.5%
LONG-RUN
22%
22%
TIME PERIOD
1973-1992
1961-2001
1 Source: Sorrell and Dimitropolous (2007) table 4.7.
While national (Table 4-6) and state level (Table 4-7) data have found relatively
consistent long-run estimates of the rebound effect, household surveys display more variability
(Table 4-8). There are several possible explanations for this larger variability. One explanation
is that some of these studies do not include vehicle age as an explanatory variable, thus leading
to omitted variable bias.8 Another explanation is that it is difficult to differentiate between the
4-17
-------
Draft Joint Technical Support Document
impacts of residential density and fuel prices, since households with higher fuel prices are more
likely to be in urban areas.9
Table 4-8 Estimates of the Rebound Effect Using U.S. Survey Data
AUTHOR
(YEAR)
Goldberg (1966)
Greene (1999)
Pickrell &
Schimek(1999)
Puller &
Greening (1999)
West (2004)
SHORT-
RUN
0%
49%
87%
LONG-RUN
23%
4-34%
TIME PERIOD
CES 1984-90
EIA RTECS
1979-1994
NPTS 1995 Single year
CES 1980-90
Single year, cross-sectional
CES 1997
Single year
An important distinction among studies of the rebound effect is whether they assume that
the effect is constant, or varies over time in response to the absolute levels of fuel costs, personal
income, or household vehicle ownership. Most studies using aggregate annual data for the U.S.
assume a constant rebound effect, although some of these studies test whether the effect can vary
as changes in retail fuel prices or average fuel economy alter fuel cost per mile driven. Many
studies using household survey data estimate significantly different rebound effects for
households owning varying numbers of vehicles, although they arrive at differing conclusions
about whether the rebound effect is larger among households that own more vehicles. Finally,
one recent study using state-level data concludes that the rebound effect varies directly in
response to changes in personal income and the degree of urbanization of U.S. cities, as well as
fuel costs.
In order to provide a more comprehensive overview of previous estimates of the rebound
effect, EPA and NHTSA reviewed 22 studies of the rebound effect conducted from 1983 through
2005. The agencies then performed a detailed analysis of the 66 separate estimates of the long-
run rebound effect reported in these studies, which is summarized in Table 4-9 below.n As the
table indicates, these 66 estimates of the long-run rebound effect range from as low as 7 percent
to as high as 75 percent, with a mean value of 23 percent. Limiting the sample to 50 estimates
reported in the 17 published studies of the rebound effect yields the same range, but a slightly
higher mean estimate (24 percent).
The type of data used and authors' assumption about whether the rebound effect varies
over time have important effects on its estimated magnitude. The 34 estimates derived from
n In some cases, NHTSA derived estimates of the overall rebound effect from more detailed results reported in the
studies. For example, where studies estimated different rebound effects for households owning different numbers
of vehicles but did not report an overall value, we computed a weighted average of the reported values using the
distribution of households among vehicle ownership categories.
4-18
-------
Economic Assumptions Used in the Agencies' Analyses
analysis of U.S. annual time-series data produce a mean estimate of 18 percent for the long-run
rebound effect, while the mean of 23 estimates based on household survey data is considerably
larger (31 percent), and the mean of 9 estimates based on state data (25 percent) is close to that
for the entire sample. The 37 estimates assuming a constant rebound effect produce a mean of
23 percent, identical to the mean of the 29 estimates reported in studies that allowed the rebound
effect to vary in response to fuel prices, vehicle ownership, or household income.
Table 4-9 Summary Statistics for Estimates of the Rebound Effect
Category of Estimates
All Estimates
Published Estimates
U.S. Time-Series Data
Household Survey Data
Pooled U.S. State Data
Constant Rebound Effect (1)
Variable Rebound Effect: (1)
Number of
Studies
22
17
7
13
2
15
10
Number of
Estimates
66
50
34
23
9
37
29
Range
Low
7%
7%
7%
9%
8%
7%
10%
High
75%
75%
45%
75%
58%
75%
45%
Distribution
Median
22%
22%
14%
31%
22%
20%
23%
Mean
23%
24%
18%
31%
25%
23%
23%
Std. Dev.
14%
14%
9%
16%
14%
16%
10%
4.2.5.2 Summary of Recent Studies and Analyses of the Rebound Effect
More recent studies since 2007 indicate that the impacts of fuel prices and fuel economy
have been decreasing over time as incomes increase and real gasoline prices fall. The theoretical
argument for why the rebound effect should vary is that the responsiveness to the fuel cost of
driving will be larger when it is a larger proportion of the total cost of driving. As incomes rise
(or fuel prices fall), the responsiveness to the fuel cost per mile of driving will decrease if people
view the time cost of driving - which is likely to be related to their income levels - as a larger
component of the total cost.
Small and Van Dender combined time series data for each of the 50 States and the
District of Columbia to estimate the rebound effect, allowing the magnitude of the rebound to
vary over time.10 For the time period from 1966-2001, their study found a long-run rebound
effect of about 22 percent, which is consistent with previously published studies. But for the
most recent five year period (1997-2001), the long-run rebound effect decreased to 11 percent.
Furthermore, when the authors updated their estimates with data through 2004, the long-run
rebound effect for the most recent five year period (2000-2004) dropped to 6 percent.11
New research conducted by David Greene in 2008-2009 at EPA's request further supports
the theory that the magnitude of the rebound effect is declining over time.12 Over the entire time
period analyzed (1966-2007), Greene found that fuel economy had no statistically significant
impact on VMT. One potential explanation for Greene's conclusion is that there is limited
variation in fleet-wide fuel economy in the later time period (1982-2007), so that most of the
variation in VMT over that period is attributed to factors other than fuel economy. While Small
and Van Dender did not find a statistically significant coefficient for fuel efficiency, they did
find a statistically significant coefficient for the price of fuel. They then tested the hypothesis
4-19
-------
Draft Joint Technical Support Document
that the elasticity of vehicle travel with respect to the price of fuel was equal to its elasticity with
respect to the rate of fuel consumption (gallons per mile), and found that the data could not reject
that hypothesis. As a consequence, they estimated the rebound effect as the elasticity of vehicle
travel with respect to fuel cost per mile, which constrains these elasticities to be equal in
magnitude. Greene also attempted to re-estimate the Small and Van Dender results using his
national data set. When using Greene's preferred functional form, the projected rebound effect is
approximately 12 percent in 2007, and drops to 10 percent in 2010, 9 percent in 2016 and 8
percent in 2030.
4.2.5.3 NHTSA analysis of the rebound effect
To provide additional insight into the rebound effect for the purposes of this rulemaking,
NHTSA developed several new estimates of its magnitude. These estimates were developed by
estimating and testing several econometric models of the relationship between vehicle miles-
traveled and factors that influence it, including household income, fuel prices, vehicle fuel
efficiency, road supply, the number of vehicles in use, vehicle prices, and other factors.
As the study by Small and Van Dender pointed out, it is important to account for the
effect of fuel prices when attempting to estimate the rebound effect. Failing to control for
changes in fuel prices is likely to bias estimates of the rebound effect. Therefore, changes in fuel
prices are taken into account in Volpe's analysis of the rebound effect. Several different
approaches were used to estimate the fuel economy rebound effect for light duty vehicles,
including single equation OLS regressions, multiple equation systems estimated using two- and
three-stage least squares regression, vector autoregression, and vector error-correction models.
With the except of single-equation regressions estimated using OLS, all of these procedures
attempt to account for the endogenous relationship of fuel efficiency to fuel prices.
The results from each of these approaches are presented in Table 4-10 below. The table
reports the value of the rebound effect calculated over the entire period from 1950 through 2006,
as well as for the final year of that period. In addition, the table presents forecasts of the average
rebound effect between 2010 and 2030, which utilize forecasts of personal income, fuel prices,
and fuel efficiency from EIA's AEO 2009 Reference Case.
The results of NHTSA's analysis are broadly consistent with the findings from previous
research summarized above. The historical average long-run rebound effect is estimated to range
from 16-30 percent, and comparing these estimates to its calculated values for 2006 (which range
from 8-14 percent) gives some an indication that it is declining in magnitude. The forecast
values of the rebound effect shown in the table also suggest that this decline is likely to continue
through 2030, as they range from 4-16 percent.
4-20
-------
Economic Assumptions Used in the Agencies' Analyses
Table 4-10 Summary of NHTSA Estimates of the Rebound Effect
Model
Small- Van Dender
single VMT
equation
Small- Van Dender
three-equation
system
Single-equation
VMT model
Single-equation
VMT model
Single-equation
VMT model
Single-equation
VMT model
Three-equation
system for VMT,
fuel efficiency, and
vehicle stock
Three-equation
system for VMT,
fuel efficiency, and
vehicle stock
Three-equation
system for VMT,
VMT
Measure
annual VMT
per adult
annual VMT
per adult
annual VMT
per adult
annual VMT
per vehicle
annual VMT
per adult
annual VMT
per vehicle
annual VMT
per vehicle
annual VMT
per vehicle
annual VMT
per vehicle
Variables Included in VMT Equation
fuel cost per mile, per Capita income, vehicle stock, road
miles per adult, fraction of population that is adult,
fraction of population living in urban areas, fraction of
population living in urban areas with heavy rail, dummy
variables for fuel rationing, time trend
fuel cost per mile, per Capita income, vehicle stock, road
miles per adult, fraction of population that is adult,
fraction of population living in urban areas, fraction of
population living in urban areas with heavy rail, dummy
variables for fuel rationing, time trend
personal income, road miles per Capita, time trend
fuel cost per mile, personal income, road miles per
Capita, time trend
fuel cost per mile, personal income, road miles per
Capita, dummy variables for fuel rationing, time trend
fuel cost per mile, personal income, vehicles per road
mile, % of fleet manufactured under CAFE standards,
new vehicle prices
fuel cost per mile, personal income, vehicles per capita,
vehicles per road mile, fraction of adult population
licensed to drive, new vehicle prices, % of fleet
manufactured under CAFE standards
fuel cost per mile, personal income, vehicles per capita,
vehicles per road mile, fraction of adult population
licensed to drive, new vehicle prices, % of fleet
manufactured under CAFE standards
fuel cost per mile, personal income, vehicles per capita,
vehicles per road mile, fraction of adult population
Estimation
Technique
OLS
3SLS
OLS
OLS
OLS
IV (for fuel
cost per mile)
2SLS
3SLS
Vector auto-
regression
Rebound Effects:
1950-
2006
33.0%
21.6%
18.4%
17.6%
34.0%
16.3%
29.5%
29.8%
19.9%
2006
15.8%
5.8%
11.7%
15.2%
20.8%
9.2%
13.4%
13.7%
10.8%
2010-
2030*
8.0%
3.4%
9.2%
15.7%
13.6%
7.0%
15.9%
16.2%
4.2.5.4 Basis for Rebound Effect Used by EPA and NHTSA in This Proposed Rule
Based on a combination of historical estimates of the rebound effect and more recent
analyses conducted by EPA and NHTSA, an estimate of 10 percent for the rebound effect was
used for this rulemaking. Although this value represents the lower bound of the historical range
of values for the rebound effect, more recent analysis strongly suggests that the rebound effect is
declining over time. As part of this proposal, we are also providing information on options for
projecting the rebound effect in the future. This projection approach, as described in the Small
Report,13 takes into account changes in income and gasoline prices when estimating the
magnitude of the rebound effect in the future. We will continue to evaluate this methodology
between the proposed and final rules, and we will take into account public comments when
determining whether it is more appropriate to use this projection approach for predicting the
rebound effect.
EPA and NHTSA also seek comment on other alternatives for estimating the rebound
effect. As one illustration, variation in the price per gallon of gasoline directly affects the per-
mile cost of driving, and drivers may respond just as they would to a change in the cost of
driving resulting from a change in fuel economy, by varying the number of miles they drive.
Because vehicles' fuel economy is fixed in the short run, variation in the number of miles driven
in response to changes in fuel prices will be reflected in changes in gasoline consumption.
Under the assumption that drivers respond similarly to changes in the cost of driving whether
they are caused by variation in fuel prices or fuel economy, the short-run price elasticity of
4-21
-------
Draft Joint Technical Support Document
demand for gasoline - which measures the sensitivity of gasoline consumption to changes in its
price per gallon - may provide some indication about the magnitude of the rebound effect itself.
The agencies also invite comment on the extent to which the short run elasticity of demand for
gasoline with respect to its price can provide useful information about the size of the rebound
effect. Specifically, the agencies seek comment on whether it would be appropriate to use the
price elasticity of demand for gasoline, or other alternative approaches, to guide their choice of a
value for the rebound effect.
4.2.6 Benefits from Increased Vehicle Use
The increase in vehicle use from the rebound effect provides additional benefits to their
owners, who may make more frequent trips or travel farther to reach more desirable destinations.
This additional travel provides benefits to drivers and their passengers by improving their access
to social and economic opportunities away from home. As evidenced by their decisions to make
more frequent or longer trips when improved fuel economy reduces their costs for driving, the
benefits from this additional travel exceed the costs drivers and passengers incur in making more
frequent or longer trips.
The analysis estimates the economic benefits from increased rebound-effect driving as
the sum of fuel costs drivers incur plus the consumer surplus they receive from the additional
accessibility it provides.0 Because the increased in travel depend on the extent of improvement
in fuel economy, the value of benefits it provides differs among model years and alternative
CAFE standards. Under even those alternatives that would impose the highest standards,
however, the magnitude of these benefits represents a small fraction of total benefits.
4.2.7 Added Costs from Increased Vehicle Use
Although it provides some benefits to drivers, increased vehicle use associated with the
rebound effect also contributes to increased traffic congestion, motor vehicle accidents, and
highway noise. Depending on how the additional travel is distributed over the day and on where
it takes place, additional vehicle use can contribute to traffic congestion and delays by increasing
traffic volumes on facilities that are already heavily traveled during peak periods. These added
delays impose higher costs on drivers and other vehicle occupants in the form of increased travel
time and operating expenses. Because drivers do not take these added costs into account in
deciding when and where to travel, they must be accounted for separately as a cost of the added
driving associated with the rebound effect.
Increased vehicle use due to the rebound effect may also increase the costs associated
with traffic accidents. Drivers may take account of the potential costs they (and their passengers)
face from the possibility of being involved in an accident when they decide to make additional
trips. However, they probably do not consider all of the potential costs they impose on
occupants of other vehicles and on pedestrians when accidents occur, so any increase in these
0 The consumer surplus provided by added travel is estimated as one-half of the product of the decline in fuel cost
per mile and the resulting increase in the annual number of miles driven.
4-22
-------
Economic Assumptions Used in the Agencies' Analyses
"external" accident costs must be considered as another cost of additional rebound-effect driving.
Like increased delay costs, any increase in these external accident costs caused by added driving
is likely to depend on the traffic conditions under which it takes place, since accidents are more
frequent in heavier traffic (although their severity may be reduced by the slower speeds at which
heavier traffic typically moves).
Finally, added vehicle use from the rebound effect may also increase traffic noise. Noise
generated by vehicles causes inconvenience, irritation, and potentially even discomfort to
occupants of other vehicles, to pedestrians and other bystanders, and to residents or occupants of
surrounding property. Because these effects are unlikely to be taken into account by the drivers
whose vehicles contribute to traffic noise, they represent additional externalities associated with
motor vehicle use. Although there is considerable uncertainty in measuring their value, any
increase in the economic costs of traffic noise resulting from added vehicle use must be included
together with other increased external costs from the rebound effect.
EPA and NHTSA rely on estimates of congestion, accident, and noise costs caused by
automobiles and light trucks developed by the Federal Highway Administration to estimate the
increased external costs caused by added driving due to the rebound effect.14 NHTSA employed
these estimates previously in its analysis accompanying the MY 2011 final rule, and continues to
find them appropriate for this NPRM after reviewing the procedures used by FHWA to develop
them and considering other available estimates of these values. They are intended to measure the
increases in costs from added congestion, property damages and injuries in traffic accidents, and
noise levels caused by automobiles and light trucks that are borne by persons other than their
drivers (or "marginal" external costs).
Updated to 2007 dollars, FHWA's "Middle" estimates for marginal congestion, accident,
and noise costs caused by automobile use amount to 5.2 cents, 2.3 cents, and 0.1 cents per
vehicle-mile (for a total of 7.6 cents per mile), while those for pickup trucks and vans are 4.7
cents, 2.5 cents, and 0.1 cents per vehicle-mile (for a total of 7.3 cents per mile).15'p These costs
are multiplied by the annual increases in automobile and light truck use from the rebound effect
to yield the estimated increases in congestion, accident, and noise externality costs during each
future year.
4.2.8 Petroleum and Energy Security Impacts
4.2.8.1 Impact on U.S. Petroleum Imports
In 2008, U.S. petroleum import expenditures represented 21 percent of total U.S. imports
of all goods and services.16 In 2008, the United States imported 66 percent of the petroleum it
p The Federal Highway Administration's estimates of these costs agree closely with some other recent estimates.
For example, recent published research conducted by Resources for the Future (RFF) estimates marginal congestion
and external accident costs for increased light-duty vehicle use in the U.S. to be 3.5 and 3.0 cents per vehicle-mile in
year-2002 dollars. See Ian W.H. Parry and Kenneth A. Small, "Does Britain or the U.S. Have the Right Gasoline
Tax?" Discussion Paper 02-12, Resources for the Future, 19 and Table 1 (March 2002). Available at
http://www.rff.org/rff/Documents/RFF-DP-02-12.pdf (last accessed July 27, 2009).
4-23
-------
Draft Joint Technical Support Document
consumed, and the transportation sector accounts for 70 percent of total U.S. petroleum
consumption. This compares roughly to 37 percent of petroleum from imports and 55 percent
consumption of petroleum in the transportation sector in 1975.17 It is clear that petroleum
imports have a significant impact on the U.S. economy. Requiring lower-GHG vehicle
technology and improved fuel economy in the U.S. is expected to lower U.S. petroleum imports.
4.2.8.2 Background on U.S. Energy Security
U.S. energy security is broadly defined as protecting the U.S. economy against
circumstances that threaten significant short- and long-term increases in energy costs. Most
discussion of U.S. energy security revolves around the topic of the economic costs of U.S.
dependence on oil imports. The problem is that the U.S. relies on sources of imported oil from
potentially unstable sources. In addition, oil exporters have the ability to raise the price of oil by
exerting monopoly power through the formation of a cartel, the Organization of Petroleum
Exporting Countries (OPEC). Finally, these factors contribute to the vulnerability of the U.S.
economy to episodic oil supply shocks and price spikes. In 2008, U.S. imports of crude oil were
$326 billion (in 2007$, see Figure 4-1).
U.S. Expenditures on Crude Oil
700
1970
1975
1980
1985
1990
Year
1995
2000
2005
Figure 4-1 U.S. Expenditures on Crude Oil from 1970 through 2008q
q For historical data through 2006: EIA Annual Energy Review, various editions.
For data 2006-2008: EIA Annual Energy Outlook (AEO) 2009 (Update Reference (Stimulus) Base Case).
See file "aeostimtab_ll.xls" available at http://www.eia.doe.gov/oiaf/servicerpt/stimulus/aeostim.html
4-24
-------
Economic Assumptions Used in the Agencies' Analyses
One effect of the EPA/NHTSA joint proposal is that it promotes more efficient use of
transportation fuels in the U.S. The result is that it reduces U.S. oil imports, which reduces both
financial and strategic risks associated with a potential disruption in supply or a spike in the cost
of a particular energy source. This reduction in risks is a measure of improved U.S. energy
security. For this proposal, an "oil premium" approach is utilized to identify those energy-
security related impacts which are not reflected in the market price of oil, and which are
expected to change in response to an incremental change in the level of U.S. oil imports.
4.2.8.2.1 Methodology Used to Estimate U.S. Energy Security Benefits
In order to understand the energy security implications of reducing U.S. oil imports, EPA
has worked with Oak Ridge National Laboratory (ORNL), which has developed approaches for
evaluating the social costs and energy security implications of oil use. The energy security
estimates provide below are based upon a methodology developed in a peer-reviewed study
entitled, "The Energy Security Benefits of Reduced Oil Use, 2006-2015," completed in March
1 S
2008. This recent study is included as part of the docket for this rulemaking. This ORNL
study is an update version of the approach to estimating the energy security benefits of U.S. oil
import reductions developed in an ORNL 1997 Report by Leiby, Paul N., Donald W. Jones, T.
Randall Curlee, and Russell Lee, entitled "Oil Imports: An Assessment of Benefits and Costs."
When conducting this recent analysis, ORNL considered the full cost of importing
petroleum into the U.S. The full economic cost is defined to include two components in addition
to the purchase price of petroleum itself. These are: (1) the higher costs for oil imports resulting
from the effect of U.S. import demand on the world oil price and on OPEC market power (i.e.,
the "demand" or "monopsony" costs); and (2) the risk of reductions in U.S. economic output and
disruption of the U.S. economy caused by sudden disruptions in the supply of imported oil to the
U.S. (i.e., macroeconomic disruption/adjustment costs). Maintaining a U.S. military presence to
help secure stable oil supply from potentially vulnerable regions of the world was not included in
this analysis because its attribution to particular missions or activities is difficult (as discussed
further below). Section III.H.S.b of the preamble contains a detailed discussion of how the
monopsony and macroeconomic disruption/adjustment components were treated for this
analysis.
As part of the process for using the ORNL energy security estimates, EPA sponsored an
independent-expert peer review of this ORNL study. A report compiling the peer reviewers'
comments is provided in the docket.19 In addition, EPA has worked with ORNL to address
comments raised in the peer review and to develop estimates of the energy security benefits
associated with a reduction in U.S. oil imports for this proposal. In response to peer reviewer
comments, ORNL modified its model by changing several key parameters involving OPEC
supply behavior, the responsiveness of oil demand and supply to a change in the world oil price,
and the responsiveness of U.S. economic output to a change in the world oil price.
For this proposal, ORNL further updated the energy security premium by incorporating
the most recent oil price forecast and energy market trends in AEO 2009 into its model. In order
for the energy security premium estimated to be used in EPA's VECTOR model, ORNL
developed energy security estimates for a number of different years; e.g., 2015, 2020, 2030, and
4-25
-------
Draft Joint Technical Support Document
2040. NHTSA used just the 2015 estimates to estimate the energy security benefits associated
with this rule.
For 2015, ORNL has estimated that the total energy security premium associated with a
reduction of imported oil is $18.49/barrel. On a dollar per gallon basis, energy security benefits
for 2015 are $0.44/gallon. Based upon alternative sensitivities regarding OPEC supply behavior,
and the responsiveness of oil demand and supply to a change in the world oil price, the energy
security premium ranges from $9.80 to $28.08/barrel. In terms of dollars per gallon, the energy
security premium ranges from $0.23/gallon to $0.67/gallon. Please refer to Table 4-11 for
similar information for years 2015, 2020, 2030 and 2040,r as well as a breakdown of the
components of the energy security premium for each year. The components of the energy
security premium and their values are discussed below.
Table 4-11 Energy Security Premium in 2015, 2020, 2030 and 2040
(2007$/Barrel)
Year
2015
2020
2030
2040
Monopsony
(Range)
$11.79
($4.26 - $21.37)
$12.31
($4.46 - $22.53)
$10.57
($3.84-18.94)
$10.57
($3.84 -$18.94)
Macroeconomic
Disruption/Adjustment Costs
(Range)
$6.70
$3.11 -$10.67
$7.62
($3.77 -$12.46)
$8.12
($3.90 -$13.04)
$8.12
($3.90 -$13.04)
Total Mid-Point
(Range)
$18.49
$9.80 - $28.08
$19.94
($10.58 - $30.47)
$18.69
($10.52 - $27.89)
$18.69
($10.52 - $27.89)
4.2.8.2.1.1 Effect of Oil Use on Long-Run Oil Price, U.S. Import Costs, and
Economic Output
The first component of the full economic costs of importing petroleum into the U.S.
follows from the effect of U.S. import demand on the world oil price over the long-run. Because
the U.S. is a sufficiently large purchaser of foreign oil supplies, its purchases can affect the world
oil price. This monopsony power means that increases in U.S. petroleum demand can cause the
world price of crude oil to rise, and conversely, that reduced U.S. petroleum demand can reduce
the world price of crude oil. Thus, one benefit of decreasing U.S. oil purchases, due to the
increased availability and use of other transportation fuels, is the potential decrease in the crude
oil price paid for all crude oil purchased.
The demand or monopsony effect can be readily illustrated with an example. If the U.S.
imports 10 million barrels per day at a world oil price of $50 per barrel, its total daily bill for oil
r AEO 2009 forecasts energy market trends and values only to 2030. The energy security premia post-2030 were
assumed to be the 2030 estimate.
4-26
-------
Economic Assumptions Used in the Agencies' Analyses
imports is $500 million. If a decrease in U.S. imports to 9 million barrels per day causes the
world oil price to drop to $49 per barrel, the daily U.S. oil import bill drops to $441 million (9
million barrels times $49 per barrel). While the world oil price only declines $1, the resulting
decrease in oil purchase payments of $59 million per day ($500 million minus $441 million) is
equivalent to an incremental benefit of $59 per barrel of oil imports reduced, or $10 more than
the newly-decreased world price of $49 per barrel. This additional $10 per barrel "import cost
premium" represents the incremental external benefits to the U.S. for avoided import costs
beyond the price paid oil purchases. This additional benefit arises only to the extent that
reduction in U.S. oil imports affects the world oil price. ORNL estimates this component of the
energy security benefit in 2015 to be $11.79/barrel, with a range of $4.26 - $21.37/barrel of
imported oil reduced.
4.2.8.2.1.2 Short-Run Disruption Premium from Expected Costs of Sudden Supply
Disruptions
The second component of the oil import premium, the "macroeconomic
disruption/adjustment costs premium," arises from the effect of oil imports on the expected cost
of disruptions. A sudden increase in oil prices triggered by a disruption in world oil supplies has
two main effects: (1) it increases the costs of imports in the short run, further expanding the
transfer of U.S. wealth to foreign producers, and (2) it can lead to macroeconomic contraction,
dislocation and Gross Domestic Product (GDP) losses. ORNL estimates the composite estimate
of these two factors that comprise the macroeconomic disruption/adjustment costs premium to be
$6.70/barrel in 2015, with a range of $3.11- $10.67/barrel of imported oil reduced.
4.2.8.2.1.2.1 Higher Costs of Oil Imports and Wealth Transfer during Shocks
During oil price shocks, the higher price of imported oil causes increased payments for
imports and a transfer of wealth from the U.S to oil exporters. This increased claim on U.S.
economic output is a loss to the U.S. that is separate from and additional to any reduction in
economic output due to the shock. The increased wealth transfer during shocks is counted as a
loss to the degree that the expected price increase is not anticipated and internalized by oil
consumers.
4.2.8.2.1.2.2 Macroeconomic Costs: Potential Output Loss and
Dislocation/Adjustment Costs
Macroeconomic losses during price shocks reflect both aggregate output losses and
"allocative" losses. The former are a reduction in the level of output that the U.S. economy can
produce fully using its available resources; and the latter stem from temporary dislocation and
underutilization of available resources due to the shock, such as labor unemployment and idle
plant capacity. The aggregate output effect, a reduction in "potential" economic output, will last
so long as the price is elevated. It depends on the extent and duration of any disruption in the
world supply of oil, since these factors determine the magnitude of the resulting increase in
prices for petroleum products, as well as whether and how rapidly these prices return to their pre-
disruption levels.
4-27
-------
Draft Joint Technical Support Document
In addition to the aggregate contraction, there are "allocative" or "adjustment" costs
associated with dislocated energy markets. Because supply disruptions and resulting price
increases occur suddenly, empirical evidence shows they also impose additional costs on
businesses and households which must adjust their use of petroleum and other productive factors
more rapidly than if the same price increase had occurred gradually. Dislocational effects
include the unemployment of workers and other resources during the time needed for their
intersectoral or interregional reallocation, and pauses in capital investment due to uncertainty.
These adjustments temporarily reduce the level of economic output that can be achieved even
below the "potential" output level that would ultimately be reached once the economy's
adaptation to higher petroleum prices was complete. The additional costs imposed on businesses
and households for making these adjustments reflect their limited ability to adjust prices, output
levels, and their use of energy, labor and other inputs quickly and smoothly in response to rapid
changes in prices for petroleum products.
Since future disruptions in foreign oil supplies are an uncertain prospect, each of the
disruption cost components must be weighted by the probability that the supply of petroleum to
the U.S. will actually be disrupted. Thus, the "expected value" of these costs - the product of the
probability that a supply disruption will occur and the sum of costs from reduced economic
output and the economy's abrupt adjustment to sharply higher petroleum prices - is the relevant
measure of their magnitude. Further, when assessing the energy security value of a policy to
reduce oil use, it is only the change in the expected costs of disruption that results from the
policy that is relevant. The expected costs of disruption may change from lowering the normal
(i.e., pre-disruption) level of domestic petroleum use and imports, from any induced alteration in
the likelihood or size of disruption, or from altering the short-run flexibility (e.g., elasticity) of
petroleum use.
In summary, the steps needed to calculate the disruption or security premium are: 1)
determine the likelihood of an oil supply disruption in the future; 2) assess the likely impacts of a
potential oil supply disruption on the world oil price; 3) assess the impact of the oil price shock
on the U.S. economy (in terms of import costs and macroeconomic losses); and 4) determine
how these costs change with oil imports. The value of price spike costs avoided by reducing oil
imports becomes the oil security portion of the premium.
4.2.8.2.2 Costs of Existing U.S. Energy Security Policies
The last often-identified component of the full economic costs of U.S. oil imports is the
costs to the U.S. taxpayers of existing U.S. energy security policies. The two primary examples
are maintaining a military presence to help secure a stable oil supply from potentially vulnerable
regions of the world and maintaining the Strategic Petroleum Reserve (SPR). The SPR is the
largest stockpile of government-owned emergency crude oil in the world. Established in the
aftermath of the 1973-74 oil embargo, the SPR provides the U.S. a response option should a
disruption in commercial oil supplies threaten the U.S. economy. It also allows the U.S. to meet
part of its International Energy Agency obligation to maintain emergency oil stocks, and it
provides a national defense fuel reserve.
U.S. military costs are excluded from the analysis performed by ORNL because their
attribution to particular missions or activities is difficult. Most military forces serve a broad
4-28
-------
Economic Assumptions Used in the Agencies' Analyses
range of security and foreign policy objectives. Attempts to attribute some share of U.S. military
costs to oil imports are further challenged by the need to estimate how those costs might vary
with incremental variations in U.S. oil imports. Similarly, while the costs for building and
maintaining the SPR are more clearly related to U.S. oil use and imports, historically these costs
have not varied in response to changes in U.S. oil import levels. Thus, while SPR is factored
into the ORNL analysis, the cost of maintaining the SPR is excluded.
4.2.8.2.3 Modifications to Analysis Based Upon Peer Reviewer Comments
The EPA commissioned ORNL to conduct a number of sensitivity analyses to address the
comments of the peer reviewers. Based upon the peer reviewer comments, key parameters that
influence the "oil import" premium were assessed. Since not all the comments were in
agreement with each other, several ranges of different parameters were developed for the
analyses. These sensitivities used the most recent price forecasts and energy market trends
available at the time the peer review was being conducted and completed, the AEO 2007
Reference Case. Thus, the results presented below are suggestive of how the energy security
premium is influenced by alternative assumptions of key parameters that influence world oil
markets. A summary of the results of the analyses are shown in Table 4-12.
Three key parameters were varied in order to assess their impacts on the oil import
premium: (1) the response of OPEC supply, (2) the combined response of non-U.S., non-OPEC
demand and supply and (3) the GDP response to a change in the world oil as a result of reduced
U.S. oil imports. The cases used updated supply/demand elasticities for non-U.S./non-OPEC
region after considering more recent estimates than those used in 1997 study. As a result, the
total market responsiveness is greater than previous ORNL estimates. Only small changes to the
world oil price are anticipated from a substantial reduction in U.S. demand, on average, about
$0.70/barrel for every million-barrels-per-day reduction in demand.
In the ORNL framework, OPEC-behavior is treated parametrically, with a wide range of
possible responses represented by a range of supply elasticities. Case One in Table 4-12 below
refers to the AEO 2007 estimates of energy market trends and uses the elasticity parameters from
the original 1997 ORNL study. In Case Two, the OPEC supply elasticities range for 0.25 to 6
with a mean elasticity of 1.7. Case Three alters the distribution of the OPEC supply elasticities
so that the mean elasticity is 2.2 instead of 1.7. With the more elastic OPEC oil supply in Case
Three, the oil premium is lower. Alternatively, a candidate rule for OPEC strategic response
behavior, adapted from a lead article on what behavior maximizes OPEC's long run net revenue
in a robust way,20 would have OPEC responding to preserve its worldwide oil market share.
This is presented as Case Seven. Application of this rule instead of the range of OPEC supply
responses used leads to an estimate of the oil import premium that is between Case Two and
Case Three.
The second key parameter that was varied based upon peer reviewer comments was non-
OPEC, non-U.S. demand and supply responsiveness to a change in the U.S. oil import demand
and, hence, the world oil price. In Case Four, the mean non-U.S./non-OPEC demand and supply
elasticities are taken to each be 0.3 in absolute value terms. When combined together, the net
elasticity of import demand from this region the non-U.S./non-OPEC region is roughly
1.6. Case Five takes the Case Four assumptions of a more elastic OPEC supply behavior and
4-29
-------
Draft Joint Technical Support Document
combines those assumptions with the 1.6 net elasticity of import demand for the non-U.S./non-
OPEC region. Case Six looks at the consequences of a yet higher net elasticity of import
demand 2.28 for the non-U.S./non-OPEC region. The impact on the oil import premium is
relatively modest.
Cases Eight and Nine consider a reduced GDP elasticity, the parameter which
summarizes the sensitivity of GDP to oil price shocks. Several reviewers suggested a lower
estimate for this parameter. In response to their comments, a couple of cases were examined
where the GDP elasticity was lowered to 0.032 in comparison to the original ORNL estimate of
0.0495. As anticipated, this change lowered the oil import premium modestly. For example,
compared with Case Four where OPEC supply is more elastic, lowering the GDP elasticity with
respect to the world oil price reduced the oil import premium by roughly $0.40/barrel. This is
because the GDP-dislocation component is only about one-quarter of the total premium, and
there are offsetting changes in other components. The last case examined, Case Nine, looks at
the consequences for the oil import premium with a reduced elasticity of GDP if OPEC attempts
to maintain its share of the world oil market.
Clearly there is an unavoidable degree of uncertainty about the magnitude of marginal
economic costs from the U.S. importation of petroleum, and the size of the oil import premium.
ORNL sought to reflect this with probabilistic risk analysis over key input factors, guided by the
available literature and the best judgment of oil market experts. Cases shown in Table 4-12
explore some reasonable variations in the ranges of input assumptions and the mean oil premium
estimates vary in a fairly moderate range between roughly $11 and $15/barrel of imported oil.
On balance, Case Eight suggested a reasonable and cautious assessment of the premium value to
ORNL, and is ORNL's recommended case. This is based on a review of important driving
factors, the numerical evaluations and simulations over major uncertainties, and taking into
consideration the many comments and suggestion from the reviewers, the EPA and other
Agencies. This recommended case, and the premium range resulting from 90% of the simulated
outcomes, encompasses a wide array of perspectives and potential market outcomes in response
to a reduction of U.S. imports.
As mentioned previously, this recommended case relied on the most recent available
projections of the U.S. and world oil market for the next ten years based upon the AEO 2007
Reference Case. OPEC-behavior was treated parametrically, with a wide range of possible
responses represented by a wide range of supply elasticities, from small to quite large. This
recommended case recognized that the OPEC response is the most uncertain single element of
this analysis. It could vary between inelastic defense of output levels, or market share, or could
be highly elastic in defense of price, probably at the expense of longer run cartel power and
discounted net profits. The balance between possible elastic and inelastic OPEC response was
essentially even over a fairly wide range of elasticities. ORNL concluded that this is the best
way to estimate OPEC behavior until greater progress can be made in synthesizing what insights
are available from the evolving strategic game-theoretic and empirical research on OPEC
behavior, and advancing that research. An alternative would have been to use OPEC strategic
response behavior to maximize long-run net revenue, which may well correspond to market-
share preservation behavior (e.g., Case Seven), and a somewhat higher premium value.
4-30
-------
Economic Assumptions Used in the Agencies' Analyses
Finally, ORNL's recommended case used a GDP elasticity range, the parameter which
summarizes the sensitivity of GDP to oil price shocks, which is reduced compared to earlier
estimates, and compared to the full range of historically-based estimates. This helped address
the concerns of those who either question the conclusions of past empirical estimates or expect
that the impacts of oil shocks may well be declining.
4-31
-------
Draft Joint Technical Support Document
Table 4-12 Summary Results - Oil Import Premium Under Various Cases ($2007/BBL)
Component
Monopsony Component
Disruption Import Costs
Disruption Dislocation Costs
Economic Disruption/Adjustment
Costs
Total Mid- Point
Total Premium, in $/Gallon
Price Reduction ($/MMBD)
Statistic
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Range
Mean
Mean
1) AEO2007
Base Outlook,
1997 Study
Elasticities
$5.57
($3. 60 -$8.1 9)
$2.36
($0.51 - $4.66)
$3.83
($1 .06 - $6.69)
$6.19
($2.94 -$10.01)
$1 1 .75
($8. 04 -$15. 96)
$0.28
$0.52
2) AEO2007
Base Outlook,
Wider Range of
OPEC Supply
Elasticities
$10.26
($3.10 -$21 .22)
$2.33
($0.58 - $4.57)
$3.70
($1 .04 - $6.57)
$6.03
($2.94 - $9.75)
$16.29
($8.10 -$27 .42)
$0.39
$1.04
3) Case 2 with
Revised Wider
Range of OPEC
Supply
Elasticities
$8.16
($3.1 2 -$19.30)
$2.34
($0.66 -$4.61)
$3.76
($1 .06 - $6.74)
$6.10
($2.91 - $9.85)
$14.27
($7.90 - $25.49)
$0.34
$0.81
4) Case 2 with
Updated Non-
OPEC
Supply/Demand
Elasticities
$7.77
($2.94 -$13.75)
$1.92
($0.39 - $3.75)
$3.42
($0.92 - $6.23)
$5.34
($2.53 - $8.69)
$13.11
($7.26 -$19.59)
$0.31
$0.75
5) Case 4 plus
Revised Wider
Range of OPEC
Supply
Elasticities
$6.52
($2.90 -$13.06)
$1.94
($0.38 - $3.84)
$3.45
($0.94 -$6.10)
$5.39
($2.60 - $8.62)
$1 1 .90
($6.96 -$18.62)
$0.28
$0.61
6) Case 4
Variant with
Wider, Higher
Range of Non-
U.S./Non-OPEC
Supply/Demand
Elasticities
$6.44
($2.69 - $1 1 .33)
$1.93
($0.42 - $3.70)
$3.47
($0.92 -$6.1 3)
$5.40
($2.58 - $8.92)
$1 1 .86
($6.79 -$17.23)
$0.28
$0.60
7) Case 6 with
Applied Strategic
OPEC Behavioral
Rule: Maintain
Market Share
$9.36
($6.69 -$12.42)
$1.89
($0.44 - $3.66)
$3.41
($0.87 - $6.26)
$5.30
($2.54 - $8.74)
$14.65
($11. 03 -$18.61)
$0.35
$0.90
8) Case 4 with
Reduced GDP
Elasticity
$7.86
($2.94 -$13.89)
$2.20
($0.63 - $4.07)
$2.67
($0.90 - $4.84)
$4.87
($2.23 - $7.85)
$12.71
($7.07 -$19.02)
$0.30
$0.75
9) Case 7 with
Reduced GDP
Elasticity
$9.42
($6.72 -$12.54)
$2.16
($0.62 - $3.96)
$2.64
($0.87 - $4.87)
$4.81
($2.25 - $7.77)
$14.23
($10.76 -$18.25)
$0.34
$0.90
Cases
1) Based on AEO2007. Updated oil market outlook from AEO1994 Base Case to AEO2007 Base Case. Among other things, this means average crude price rises from $20.33 to $48.34. All elasticities
match 1997 values. Non-U.S. elasticity of import demand = -0.
2) AEO2007 Base Outlook, with wider range of OPEC supply elasticities, 0.25 to 6.0 and a mean elasticity of 1.76.
3) Revise Case 2, with OPEC behavior distributed over elasticities 0 to 6, so that 25% of response is inelastic (< 1.0), mode elasticity is 2.0 (mean elasticity is 2.2 rather than 1.76)
4) Updated Case 2 supply/demand elasticities for non-OPEC region with more recent estimates. Elasticity of non-U.S. demand -0.2 to -0.4, with mean and mode -0.3, non-U.S. Supply = 0.2 to 0.4, with
mean and mode 0.3, implying (mode) net elasticity of impo
5) Revise Case 4, with OPEC behavior distributed over elasticities 0 to 6, so that 25% of response is inelastic (< 1.0), mode elasticity is 2.0 (mean elasticity is 2.2 rather than 1.76). Net elasticity of import
demand is -1.6 for the non-U.S./non-OPEC r
6) Alternative to Case 4 with expanded (and higher) range of non-U.S. supply/demand elasticities. Elasticity of non-U.S. demand = -0.3 to -0.7, triangular distribution with mode -0.4 mean -0.467, elasticity of
non-U.S. supply = 0.2 to 0.6, mode 0.3 and m
7) Applied Strategic OPEC Behavioral Rule to Case 6: Maintain Market Share (Gately 2004 paper best strategy). This rule implies that OPEC Supply elasticity matches that of all non-OPEC supply. As a
result non-U.S. elasticity of import demand ranges from
8) Variant on version Case 4, considered reduced GDP elasticity for future disruptions (range -0.01 to -0.054; midcase value -0.032; mean value is -0.032, reduced from mean value of -0.0495). OPEC-
behavior treated parametrically.
9) Revise Case 7 (which applied Strategic OPEC Behavioral Rule to Case 6: Maintain Market Share (Gately 2004 paper best strategy)) with reduced GDP elasticity for future disruptions (range -0.01 to -
0.054; midcase value -0.032; mean value is -0.032, reduc
4-32
-------
Economic Assumptions Used in the Agencies' Analyses
4.2.8.2.4 Energy Security Premium Values for This Proposal
After the peer review process was completed, EPA worked with ORNL to further
update the analysis for this proposed vehicle program. Using the same methodology as the
peer reviewed model, but updating the analysis using AEO 2009 world oil price values,
ORNL calculated energy security premium estimates for the years 2015, 2020, and 2030.
Since the AEO 2009 ends its forecast in the year 2030, EPA assumed that the post-2030
energy security premium estimates were the same as the 2030 estimate. The results of this
analysis are shown above in Table 4-1 lin Chapter 4.2.8.
4.2.8.3 The Impact of Fuel Savings on U.S. Petroleum Imports
Based on a detailed analysis of differences in fuel consumption, petroleum imports,
and imports of refined petroleum products among the Reference Case, High Economic
Growth, and Low Economic Growth Scenarios presented in AEO 2009, NHTSA and EPA
estimate that approximately 50 percent of the reduction in fuel consumption resulting from
adopting improved fuel efficiency standards and GHG standards is likely to be reflected in
reduced U.S. imports of refined fuel, while the remaining 50 percent would be expected to be
reflected in reduced domestic fuel refining.8 Of this latter figure, 90 percent is anticipated to
reduce U.S. imports of crude petroleum for use as a refinery feedstock, while the remaining
10 percent is expected to reduce U.S. domestic production of crude petroleum.1 Thus on
balance, each gallon of fuel saved as a consequence of improved fuel efficiency standards and
GHG standards is anticipated to reduce total U.S. imports of crude petroleum or refined fuel
by 0.95 gallons." Section 5.3 of the RIA contains a discussion regarding caveats for the fuel
savings estimated due to implementation of this rule.
4.2.9 Air Pollutant Emissions
Car and light truck use, fuel refining, and fuel distribution and retailing also generate
emissions of certain criteria air pollutants, including carbon monoxide (CO), hydrocarbon
compounds (usually referred to as "volatile organic compounds," or VOC), nitrogen oxides
(NOX), fine particulate matter (PMi.s), and sulfur dioxide (802). Emissions of most of these
pollutants are associated with the number of vehicle miles driven, rather than with fuel
consumption. While reductions in fuel refining and distribution that result from lower fuel
consumption will reduce emissions of criteria pollutants, additional vehicle use associated
with the rebound effect from higher fuel economy will increase emissions of most of these
pollutants. Sulfur dioxide is an exception, as described below.
s Differences between forecast annual U.S. imports of crude petroleum and refined products among these three
scenarios range from 24-89% of differences in projected annual gasoline and diesel fuel consumption in the U.S.
These differences average 49% over the forecast period spanned by AEO 2009.
' Differences between forecast annual U.S. imports of crude petroleum among these three scenarios range from
67-97% of differences in total U.S. refining of crude petroleum, and average 85% over the forecast period
spanned by AEO 2009.
u This figure is calculated as 0.50 + 0.50*0.9 = 0.50 + 0.45 = 0.95.
4-33
-------
Draft Joint Technical Support Document
Thus the net effect of stricter fuel efficiency and GHG standards on total emissions of
each criteria pollutant depends on the relative magnitudes of reduced emissions in fuel
refining and distribution, and increases in emissions from vehicle use. Because the
relationship between emission rates (emissions per gallon refined of fuel or mile driven) in
fuel refining and vehicle use is different for each criteria pollutant, the net effect of fuel
savings from increased fuel efficiency and GHG standards on total emissions of each
pollutant is likely to differ.
4.2.9.1 Reductions in Emissions of Criteria Air Pollutants
EPA and NHTSA estimate the increases in emissions of each criteria air pollutant
from additional vehicle use by multiplying the increase in total miles driven by cars and light
trucks of each model year and age by their estimated emission rates per vehicle-mile of each
pollutant. These emission rates differ between cars and light trucks as well as between
gasoline and diesel vehicles, and both their values for new vehicles and the rates at which they
increase with age and accumulated mileage can vary among model years. With the exception
of SOi, the agencies calculated the increase in emissions of these criteria pollutants from
added car and light truck use by multiplying the estimated increases in their vehicles' use
during each year over their expected lifetimes by per-mile emission rates appropriate to each
vehicle type, fuel used, model year, and age as of that future year.
These emission rates were estimated by U.S. EPA using its recently-developed Motor
Vehicle Emission Simulator (Draft MOVES 2009). The MOVES model assumes that the per-
mile rates at which these pollutants are emitted are determined by EPA regulations and the
effectiveness of catalytic after-treatment of engine exhaust emissions, and are thus unaffected
by changes in car and light truck fuel economy. As a consequence, the effects of required
increases in fuel economy on emissions of these pollutants from car and light truck use are
determined entirely by the increases in driving that result from the fuel economy rebound
effect.
Emission factors in the MOVES database are expressed in the form of grams per
vehicle-hour of operation. To convert these emission factors to grams per mile for use in EPA
and NHTSA's calculations, MOVES was run for the year 2050, and was programmed to
report aggregate emissions from vehicle starting, operation, storage, and refueling. EPA
analysts selected the year 2050 in order to generate emission factors that were representative
of lifetime average emission rates for vehicles meeting the agency's Tier 2 emission
standard/ Separate estimates were developed for each vehicle type and model year, as well as
for each state and month, in order to reflect the effects of regional and temporal variation in
temperature and other relevant variables on emissions.
v Because all light-duty emission rates in Draft MOVES 2009 are assumed to be invariant after MY
2010, a calendar-year 2050 run produced a full set of emission rates that reflect anticipated deterioration
in the effectiveness of vehicles' emission control systems with increasing age and accumulated mileage
for post-MY 2010 vehicles.
4-34
-------
Economic Assumptions Used in the Agencies' Analyses
The MOVES emissions estimates were then summed to the model year level and
divided by average distance traveled in order to produce per-mile emission factors for each
pollutant. The resulting emission rates represent average values across the nation, and
incorporate typical temperature variations over an entire calendar year. These national
average rates also reflect county-specific differences in fuel composition, as well as in the
presence and type of vehicle inspection and maintenance programs.w
Emission rates for the criteria pollutant SOi were calculated by EPA and NHTSA
using average fuel sulfur content estimates supplied by EPA, together with the assumption
that the entire sulfur content of fuel is emitted in the form of SOi- These calculations assumed
that national average gasoline and diesel sulfur levels would remain at current levels/ There,
unlike many other criteria pollutants, sulfur dioxide emissions from vehicle use decrease
proportionally to the decrease in fuel consumption.
Emissions of criteria air pollutants also occur during each phase of fuel production and
distribution, including crude oil extraction and transportation, fuel refining, and fuel storage
and transportation. The reduction in emissions during each of these phases depends on the
extent to which fuel savings result in lower imports of refined fuel, or in reduced domestic
fuel refining. To a lesser extent, they also depend on whether reductions in domestic gasoline
refining are reflected in reduced imports of crude oil or in reduced domestic extraction of
petroleum.
Both EPA's and NHTSA's analyses assume that reductions in imports of refined fuel
would reduce criteria pollutant emissions during fuel storage and distribution only.
Reductions in domestic fuel refining using imported crude oil as a feedstock are assumed to
reduce emissions during fuel refining, storage, and distribution, because each of these
activities would be reduced. Finally, reduced domestic fuel refining using domestically-
produced crude oil is assumed to reduce emissions during all four phases of fuel production
and distribution/
EPA estimated the reductions in criteria pollutant emissions from producing and
distributing fuel that would occur with alternative fuel efficiency and GHG standards using
emission rates obtained from Argonne National Laboratories' Greenhouse Gases and
Regulated Emissions in Transportation (GREET) model.21 The GREET model provides
w The national mix of fuel types includes county-level market shares of conventional and reformulated gasoline,
as well as county-level variation in sulfur content, ethanol fractions, and other fuel properties.
Inspection/maintenance programs at the county level account for detailed program design elements such as test
type, inspection frequency, and program coverage by vehicle type and age.
x These are 30 and 15 parts per million (ppm, measured on a mass basis) for gasoline and diesel respectively,
which produces emission rates of 0.17 grams of SO2 per gallon of gasoline and 0.10 grams per gallon of diesel.
y In effect, this assumes that the distances crude oil travels to U.S. refineries are approximately the same
regardless of whether it travels from domestic oilfields or import terminals, and that the distances that gasoline
travels from refineries to retail stations are approximately the same as those from import terminals to gasoline
stations.
4-35
-------
Draft Joint Technical Support Document
separate estimates of air pollutant emissions that occur in four phases of fuel production and
distribution: crude oil extraction, crude oil transportation and storage, fuel refining, and fuel
distribution and storage.2 EPA modified the GREET model to change certain assumptions
about emissions during crude petroleum extraction and transportation, as well as to update its
emission rates to reflect adopted and pending EPA emission standards. EPA also
incorporated emission factors for the air toxics estimated in this analysis: benzene, 1,3-
butadiene, acetaldehyde, acrolein, and formaldehyde. The resulting emission factors are
shown in Table 4-13.
z Emissions that occur during vehicle refueling at retail gasoline stations (primarily evaporative emissions of
volatile organic compounds, or VOCs) are already accounted for in the "tailpipe" emission factors used to
estimate the emissions generated by increased light truck use. GREET estimates emissions in each phase of
gasoline production and distribution in mass per unit of gasoline energy content; these factors are then converted
to mass per gallon of gasoline using the average energy content of gasoline.
4-36
-------
Economic Assumptions Used in the Agencies' Analyses
Table 4-13 Emissions by Stage of Fuel Production and Distribution (grams/million Btu)
Pollutant
CO
VOC
NOx
SOx
PM2.5
Fuel Type
Conventional Gasoline
Reformulated Gasoline
Low Sulfur Diesel
Conventional Gasoline
Reformulated Gasoline
Low Sulfur Diesel
Conventional Gasoline
Reformulated Gasoline
Low Sulfur Diesel
Conventional Gasoline
Reformulated Gasoline
Low Sulfur Diesel
Conventional Gasoline
Reformulated Gasoline
Low Sulfur Diesel
Petroleum
Extraction &
Transportation1
4.908
4.908
4.908
3.035
3.035
3.035
14.91
14.91
14.91
3.926
3.926
3.926
0.645
0.645
0.645
Refinery
Energy
Use
Upstream
Emissions
0.928
0.908
0.800
0.602
0.627
0.552
3.328
3.288
2.895
4.398
4.422
3.893
1.442
1.487
1.309
Petroleum
Refining
On-Site
5.596
5.662
5.103
2.560
2.584
2.511
14.442
14.575
12.972
9.678
9.922
9.187
1.789
1.838
1.635
Petroleum
Refining2
6.525
6.571
5.903
3.162
3.211
3.063
17.771
17.863
15.866
14.076
14.344
13.080
3.231
3.325
2.943
Fuel
Transport,
Storage,
Distribution3
0.748
0.768
0.780
42.91
42.92
1.261
3.691
3.786
3.570
0.886
0.909
0.840
0.155
0.159
0.133
Air Toxics
1 ,3-Butadiene
Acetaldehyde
Acrolein
Benzene
Formaldehyde
Conventional Gasoline
Reformulated Gasoline
Low Sulfur Diesel
Conventional Gasoline
Reformulated Gasoline
Low Sulfur Diesel
Conventional Gasoline
Reformulated Gasoline
Low Sulfur Diesel
Conventional Gasoline
Reformulated Gasoline
Low Sulfur Diesel
Conventional Gasoline
Reformulated Gasoline
Low Sulfur Diesel
0.0017
0.0017
0.0017
0.0002
0.0002
0.0002
0.0001
0.0001
0.0001
0.0313
0.0313
0.0313
0.0050
0.0050
0.0050
0.0003
0.0003
0.0003
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0062
0.0064
0.0058
0.0010
0.0010
0.0009
0.0014
0.0014
0.0014
0.0002
0.0002
0.0002
0.0001
0.0001
0.0001
0.0264
0.0264
0.0264
0.0042
0.0042
0.0042
0.0017
0.0018
0.0017
0.0002
0.0002
0.0002
0.0001
0.0001
0.0001
0.0326
0.0328
0.0322
0.0052
0.0052
0.0051
0.0001
0.0001
0.0001
0.0046
0.0047
0.0044
0.0006
0.0006
0.0006
0.0787
0.0788
0.0015
0.0326
0.0335
0.0316
The petroleum extraction and transport emission factors are only applied to domestic crude oil.
2 Refinery emissions factors are applied to domestic refineries, whether refining domestic or imported crude.
3 Fuel transport, storage, and distribution emission factors represent domestic emissions and are applied to all
finished fuel, whether refined domestically or internationally.
The agency converted these emission rates from the mass per fuel energy content basis
on which GREET reports them to mass per gallon of fuel supplied using the estimates of fuel
energy content reported by GREET. The resulting emission rates were applied to both EPA's
and NHTSA's estimates of fuel consumption under alternative fuel efficiency standards to
develop estimates of total emissions of each criteria pollutant during fuel production and
distribution. The assumptions about the effects of changes in fuel consumption on domestic
4-37
-------
Draft Joint Technical Support Document
and imported sources of fuel supply discussed above were then employed to calculate the
effects of reductions in fuel use from alternative fuel efficiency and GHG standards on
changes in domestic emissions of each criteria pollutant.
Finally, EPA and NHTSA calculated the net changes in domestic emissions of each
criteria pollutant by summing the increases in its emissions projected to result from increased
vehicle use, and the reductions in emissions anticipated to result from lower domestic fuel
refining and distribution.aa As indicated previously, the effect of adopting improved fuel
efficiency and GHG standards on total emissions of each criteria pollutant depends on the
relative magnitudes of the resulting reduction in emissions from fuel refining and distribution,
and the increase in emissions from additional vehicle use. Although these net changes vary
significantly among individual criteria pollutants, we project that on balance, adopting
improved fuel efficiency and GHG standards would reduce emissions of all criteria air
pollutants except carbon monoxide (CO).
4.2.9.2 Estimated Values of Reducing Criteria Pollutant Emissions
In this analysis, EPA and NHTSA estimate the economic value of the human health
benefits associated with reducing exposure to PMi.5. Due to analytical limitations, the
estimated values do not include comparable benefits related to reductions in other criteria
pollutants (such as ozone, NO2 or SCb) or toxic air pollutants, nor do they monetize all of the
potential health and welfare effects associated with PM2.5 or the other criteria pollutants. As a
result these values underestimate the benefits associated with reductions of these pollutants.
This analysis uses a "benefit-per-ton" method to estimate selected PM2.5-related health
benefits. These PMi.s-related benefit-per-ton estimates provide the total monetized human
health benefits (the sum of premature mortality and premature morbidity) of reducing one ton
of directly emitted PM2.5, or one ton of a pollutant that contributes to secondarily-formed
PM2.5 (such as NOx, SOx, and VOCs) from a specified source. Ideally, the human health
benefits would be estimated based on changes in ambient PM2.5 concentrations and population
exposure, as determined by full-scale air quality and exposure modeling. However,
conducting such detailed modeling was not possible within the timeframe for this proposal.
The dollar-per-ton estimates for reductions in PM2.5 emissions used in this analysis are
provided in Table 4-14.
aa All emissions from increased vehicle use are assumed to occur within the U.S., since Fuel efficiency standards
would apply only to vehicles produced for sale in the U.S.
4-38
-------
Economic Assumptions Used in the Agencies' Analyses
Table 4-14 Benefits-per-ton Values (2007$) Derived Using the ACS Cohort Study for
PM-related Premature Mortality (Pope et al., 2002)a and a 3% Discount Rateb
Yearc
2015
2020
2030
2040
All Sources"
SOx VOC
$28,000 $1,200
$31,000 $1,300
$36,000 $1,500
$43,000 $1,800
Stationary (Non-EGU)
Sources6
NOx
$4,700
$5,100
$6,100
$7,200
Direct PM2. 5
$220,000
$240,000
$280,000
$330,000
Mobile Sources
NOx Direct PM2. 5
$4,900 $270,000
$5,300 $290,000
$6,400 $350,000
$7,600 $420,000
a The benefit-per-ton estimates presented in this table are based on an estimate of premature mortality derived
from the ACS study (Pope et al., 2002). If the benefit-per-ton estimates were based on the Six Cities study
(Laden et al., 2006), the values would be approximately 145% (nearly two-and-a-half times) larger.
b The benefit-per-ton estimates presented in this table assume a 3% discount rate in the valuation of premature
mortality to account for a twenty-year segmented cessation lag. If a 7% discount rate had been used, the values
would be approximately 9% lower.
0 Benefit-per-ton values were estimated for the years 2015, 2020, and 2030. For 2040, EPA and NHTSA
extrapolated exponentially based on the growth between 2020 and 2030.
d Note that the benefit-per-ton value for SOx is based on the value for Stationary (Non-EGU) sources; no SOx
value was estimated for mobile sources. The benefit-per-ton value for VOCs was estimated across all sources.
e Non-EGU denotes stationary sources of emissions other than electric generating units.
As Table 4-14 indicates, EPA projects that the per-ton values for reducing emissions
of criteria pollutants from both vehicle use and stationary sources such as fuel refineries and
storage facilities will increase over time. These projected increases reflect rising income
levels, which are assumed to increase affected individuals' willingness to pay for reduced
exposure to health threats from air pollution, as well as population growth into the future,
expanding the population exposed to air pollution in both urban and rural areas.
For certain PMi.s-related pollutants (such as direct PM2.5 and NOx), EPA estimates
different per-ton values for reducing mobile source emissions than for reductions in
emissions of the same pollutant from stationary sources such as fuel refineries and storage
facilities. These reflect differences in the typical geographic distributions of emissions of
each pollutant, their contributions to ambient levels of PMi.5, and resulting changes in
population exposure. EPA and NHTSA apply these separate values to its estimates of
changes in emissions from vehicle use and from fuel production and distribution to determine
the net change in total economic damages from emissions of those pollutants.
The benefit per-ton technique has been used in previous analyses, including EPA's
recent Ozone National Ambient Air Quality Standards (NAAQS) RIA (U.S. EPA, 2008a),22
Portland Cement National Emissions Standards for Hazardous Air Pollutants (NESHAP) RIA
(U.S. EPA, 2009a),23 and NO2 NAAQS (U.S. EPA, 2009b).24 Table 4-15 shows the
quantified and unquantified PMi.5 -related co-benefits captured in these benefit-per-ton
estimates.
4-39
-------
Draft Joint Technical Support Document
Table 4-15 Human Health and Welfare Effects of PM2.5
Pollutant /
Effect
Quantified and Monetized
in Primary Estimates
Unquantified Effects
Changes in:
PM2.5
Adult premature mortality
Bronchitis: chronic and acute
Hospital admissions: respiratory and
cardiovascular
Emergency room visits for asthma
Nonfatal heart attacks (myocardial
infarction)
Lower and upper respiratory illness
Minor restricted-activity days
Work loss days
Asthma exacerbations (asthmatic
population)
Infant mortality
Subchronic bronchitis cases
Low birth weight
Pulmonary function
Chronic respiratory diseases other
than chronic bronchitis
Non-asthma respiratory emergency
room visits
Visibility
Household soiling
Consistent with the NOi NAAQS,bb the benefits estimates utilize concentration-
response functions as reported in the epidemiology literature. Readers interested in reviewing
the complete methodology for creating the benefit-per-ton estimates used in this analysis can
consult the Technical Support Document (TSD)25 accompanying the recent final ozone
NAAQS RIA (U.S. EPA, 2008a). Readers can also refer to Fann et al. (2009)26 for a detailed
description of the benefit-per-ton methodology.cc
As described above, national per-ton estimates were developed for selected
pollutant/source category combinations. The per-ton values calculated therefore apply only to
tons reduced from those specific pollutant/source combinations (e.g., NOi emitted from
mobile sources; direct PM emitted from stationary sources). Our estimate of total PM2.5
benefits is therefore based on the total direct PM2.5 and PMi.s-related precursor emissions
(NOx, SOx, and VOCs) controlled from each source and multiplied by the respective per-ton
values of reducing emissions from that source.
The benefit-per-ton coefficients in this analysis were derived using modified versions
of the health impact functions used in the PM NAAQS Regulatory Impact Analysis.
bb Although we summarize the main issues in this chapter, we encourage interested readers to see
benefits chapter of the NO2 NAAQS for a more detailed description of recent changes to the PM benefits
presentation and preference for the no-threshold model.
cc The values included in this report are different from those presented in the article cited above. Benefits
methods change to reflect new information and evaluation of the science. Since publication of the June 2009
article, EPA has made two significant changes to its benefits methods: (1) We no longer assume that a threshold
exists in PM-related models of health impacts; and (2) We have revised the Value of a Statistical Life to equal
$6.3 million (year 2000$), up from an estimate of $5.5 million (year 2000$) used in the June 2009 report. Please
refer to the following website for updates to the dollar-per-ton estimates:
http://www.epa.gov/air/benmap/bpt.html
4-40
-------
Economic Assumptions Used in the Agencies' Analyses
Specifically, this analysis uses the benefit-per-ton estimates first applied in the Portland
Cement NESHAP RIA (U.S. EPA, 2009a), which incorporated concentration-response
functions directly from the epidemiology studies, without any adjustment for an assumed
threshold. Removing the threshold assumption is a key difference between the method used
in this analysis to estimate PM co-benefits and the methods used in analyses prior to EPA's
Portland Cement NESHAP. As a consequence, the benefit-per-ton estimates used in this
analysis include incremental benefits of reductions in PM2.5 concentrations down to their
lowest modeled levels.
Reductions in PM-related mortality provide the majority of the monetized value in
each benefit-per-ton estimate. Typically, the premature mortality-related effect coefficients
that underlie the benefits-per-ton estimates are drawn from epidemiology studies that examine
two large population cohorts: the American Cancer Society cohort (Pope et al., 2002)27 and
the Harvard Six Cities cohort (Laden et al., 2006).28 The concentration-response (C-R)
function developed from the extended analysis of American Cancer Society (ACS) cohort, as
reported in Pope et al. (2002), has previously been used by EPA to generate its primary
benefits estimate. The extended analysis of the Harvard Six Cities cohort, as reported by
Laden et al (2006), was published after the completion of the Staff Paper for the 2006 PM2.5
NAAQS and has been used as an alternative estimate in the PM2.5 NAAQS RIA and PMi.5 co-
benefits estimates in analyses completed since the PM2.5 NAAQS.
These studies provide logical choices for anchor points when presenting PM-related
benefits because, while both studies are well designed and peer-reviewed, there are strengths
and weaknesses inherent in each. Although this argues for using both studies to generate
benefits estimates, due to the analytical limitations associated with this analysis, EPA and
NHTSA have chosen to use the benefit-per-ton value derived from the ACS study. The
agencies note that benefits would be approximately 145 percent (or nearly two-and-a-half
times) larger if the per-ton benefit values based on the Harvard Six Cities were used instead.
As is the nature of benefits analyses, assumptions and methods evolve over time to
reflect the most current interpretation of the scientific and economic literature. For a period
of time (2004-2008), EPA's Office of Air and Radiation (OAR) valued mortality risk
reductions using a value of statistical life (VSL) estimate derived from a limited analysis of
some of the available studies. OAR arrived at a VSL using a range of $1 million to $10
million (2000$) consistent with two meta-analyses of the wage-risk literature.
The $1 million value represented the lower end of the interquartile range from the
Mrozek and Taylor (2002)29 meta-analysis of 33 studies. The $10 million value represented
the upper end of the interquartile range from the Viscusi and Aldy (2003)30 meta-analysis of
43 studies. The mean estimate of $5.5 million (2000$) was also consistent with the mean
VSL of $5.4 million estimated in the Kochi et al. (2006)31 meta-analysis. However, the
Agency neither changed its official guidance on the use of VSL in rule-makings nor subjected
the interim estimate to a scientific peer-review process through the Science Advisory Board
(SAB) or other peer-review group.
Until updated guidance is available, EPA determined that a single, peer-reviewed
estimate applied consistently best reflects the SAB-EEAC advice it has received. Therefore,
4-41
-------
Draft Joint Technical Support Document
EPA has decided to apply the VSL that was vetted and endorsed by the SAB in the Guidelines
for Preparing Economic Analyses (U.S. EPA, 2000)32 while they continue efforts to update
their guidance on this issue.dd This approach calculates a mean value across VSL estimates
derived from 26 labor market and contingent valuation studies published between 1974 and
1991. The mean VSL across these studies is $6.3 million (2000$). The dollar-per-ton
estimates used in this analysis are based on this revised VSL.
The benefit-per-ton estimates are subject to a number of assumptions and
uncertainties.
a. They do not reflect local variability in population density, meteorology, exposure,
baseline health incidence rates, or other local factors that might lead to an
overestimate or underestimate of the actual benefits of controlling fine particulates in
specific locations. EPA will conduct full-scale air quality modeling for the final
rulemaking in an effort to capture this variability. Please refer to Chapter 7.2 of the
DRIA for a description of EPA's modeling plans and to Chapter 7.3 for the description
of the agency's plans for quantification and monetization of health impacts for the
FRM.
b. This analysis assumes that all fine particles, regardless of their chemical composition,
are equally potent in causing premature mortality. This is an important assumption,
because PMi.5 produced via transported precursors emitted from stationary sources
may differ significantly from direct PM2.5 released from diesel engines and other
industrial sources. At the present time, however, no clear scientific grounds exist for
supporting differential effects estimates by particle type.
c. This analysis assumes that the health impact function for fine particles is linear within
the range of ambient concentrations under consideration. Thus, the estimates include
health benefits from reducing fine particles in areas with varied initial concentrations
of PM2.5, including both regions that are in attainment with fine particle standard and
those that do not meet the standard, down to the lowest modeled concentrations.
d. There are several health benefits categories that EPA and NHTSA were unable to
quantify due to limitations associated with using benefits-per-ton estimates, several of
which could be substantial. Because the NOx and VOC emission reductions
associated with this proposal are also precursors to ozone, reductions in NOx and VOC
would also reduce ozone formation and the health effects associated with ozone
exposure. Unfortunately, benefits-per-ton estimates for ozone do not exist due to
issues associated with the complexity of the atmospheric air chemistry and
nonlinearities associated with ozone formation. The PM-related benefits-per-ton
estimates also do not include any human welfare or ecological benefits. Please refer to
Chapter 7.3 of the DRIA for a description of the quantification and monetization of
dd In the (draft) update of the Economic Guidelines (U.S. EPA, 2008c), EPA retained the VSL endorsed by the
SAB with the understanding that further updates to the mortality risk valuation guidance would be forthcoming
in the near future. Therefore, this report does not represent final agency policy. The draft update of the
Economic Guidelines is available on the Internet at .
4-42
-------
Economic Assumptions Used in the Agencies' Analyses
health impact for the FRM and a description of the unqualified co-pollutant benefits
associated with this rulemaking.
As mentioned above, emissions changes and benefits-per-ton estimates alone are not a
good indication of local or regional air quality and health impacts, as the localized impacts
associated with the proposed rulemaking may vary significantly. Additionally, the
atmospheric chemistry related to ambient concentrations of PM2.5, ozone and air toxics is very
complex. Full-scale photochemical modeling is therefore necessary to provide the needed
spatial and temporal detail to more completely and accurately estimate the changes in ambient
levels of these pollutants and their associated health and welfare impacts. As discussed
above, timing and resource constraints precluded EPA and NHTSA from conducting a full-
scale photochemical air quality modeling analysis in time to include in the NPRM. For the
final rule, however, a national-scale air quality modeling analysis will be performed to
analyze the impacts of the standards on PM2.5, ozone, and selected air toxics.
4.2.10 Reductions in Emissions of Greenhouse Gases
Emissions of carbon dioxide and other greenhouse gases (GHGs) occur throughout the
process of producing and distributing transportation fuels, as well as from fuel combustion
itself. By reducing the volume of fuel consumed by passenger cars and light trucks, higher
fuel efficiency standards will thus reduce GHG emissions generated by fuel use, as well as
throughout the fuel supply cycle. Lowering these emissions is likely to slow the projected
pace and reduce the ultimate extent of future changes in the global climate, thus reducing
future economic damages that changes in the global climate are otherwise expected to cause.
Further, by reducing the probability that climate changes with potentially catastrophic
economic or environmental impacts will occur, lowering GHG emissions may also result in
economic benefits that exceed the resulting reduction in the expected future economic costs
caused by gradual changes in the earth's climatic systems. Quantifying and monetizing
benefits from reducing GHG emissions is thus an important step in estimating the total
economic benefits likely to result from establishing improved fuel efficiency and GHG
standards.
4.2.10.1 Estimating Reductions in GHG Emissions
The agencies estimated emissions of carbon dioxide (CO2) from passenger car and
light truck use by multiplying the number of gallons of each type of fuel (gasoline and diesel)
they are projected to consume with each alternative CAFE standard in effect by the quantity
or mass of CO2 emissions released per gallon of fuel consumed. This calculation assumes
that the entire carbon content of each fuel is ultimately converted to COi emissions during the
combustion process. The weighted average CO2 content of gasoline is estimated to be
approximately 8,900 grams per gallon, while that of diesel fuel is estimated to be
approximately 10,200 grams per gallon. For details, please see EPA's DRIA and NHTSA's
PRIA.
Although carbon dioxide emissions account for nearly 95 percent of total GHG
emissions that result from fuel combustion during vehicle use, emissions of other GHGs are
potentially significant as well because of their higher "potency" as GHGs than that of COi
4-43
-------
Draft Joint Technical Support Document
itself. EPA and NHTSA estimated the increases in emissions of methane (CELO and nitrous
oxide (NiO) from additional vehicle use by multiplying the increase in total miles driven by
cars and light trucks of each model year and age by emission rates per vehicle-mile for these
GHGs. These emission rates, which differ between cars and light trucks as well as between
gasoline and diesel vehicles, were estimated by EPA using its recently-developed Motor
Vehicle Emission Simulator (Draft MOVES 2009), as subsequently updated for use in this
analysis.
The MOVES model assumes that the per-mile rates at which cars and light trucks emit
these GHGs are determined by the efficiency of fuel combustion during engine operation and
chemical reactions that occur during catalytic after-treatment of engine exhaust, and are thus
independent of vehicles' fuel consumption rates. Thus MOVES emission factors for these
GHGs, which are expressed per mile of vehicle travel, are assumed to be unaffected by
changes in fuel economy.
Increases in emissions of non-CO2 GHGs are converted to equivalent increases in COi
emissions using estimates of the Global Warming Potential (GWP) of methane and nitrous
oxide. These GWPs account for the higher radiative forcing capacity of methane and nitrous
oxide when they are released into the earth's atmosphere, measured relative to that of CO2
itself. The COi equivalents of increases in emissions of these gases are then added to the
increases in emissions of CO2 itself to determine the effect of the total increase in COi-
equivalent GHG emissions from vehicle use.
4.2.10.2 Economic Benefits from Reducing GHG Emissions
EPA and NHTSA assigned a dollar value to reductions in COi emissions using the
marginal dollar value (i.e., cost) of climate-related damages resulting from carbon emissions,
also referred to as "social cost of carbon" (SCC). The SCC is intended to measure the
monetary value society places on impacts resulting from increased GHGs, such as property
damage from sea level rise, forced migration due to dry land loss, and mortality changes
associated with vector-borne diseases. Published estimates of the SCC vary widely as a result
of uncertainties about future economic growth, climate sensitivity to GHG emissions,
procedures used to model the economic impacts of climate change, and the choice of
discount rates. EPA and NHTSA's coordinated proposals present a set of interim SCC values
reflecting a federal interagency group's interpretation of the relevant climate economics
literature. Sections III.H.6 and IV.C.3.1 of the preamble provide more detail about SCC.
4.2.11 The Value of Increased Driving Range
Improving the fuel economy of passenger cars and light-duty trucks may also increase
their driving range before they require refueling. By reducing the frequency with which
drivers typically refuel their vehicles and extending the upper limit of the range they can
travel before requiring refueling, improving fuel economy thus provides some additional
benefits to their owners. Alternatively, if manufacturers respond to improved fuel economy
by reducing the size of fuel tanks to maintain a constant driving range, the resulting cost
saving will presumably be reflected in lower vehicle sales prices. If manufacturers respond
4-44
-------
Economic Assumptions Used in the Agencies' Analyses
by doing so, this presumably reflects their judgment that the value to economic benefits to
vehicle buyers from lower purchase prices exceeds that from extended refueling range.
No direct estimates of the value of extended vehicle range are readily available, so the
agencies' analyses calculate the reduction in the annual number of required refueling cycles
that results from improved fuel economy, and applies DOT-recommended values of travel
time savings to convert the resulting time savings to their economic value.33 As a coarse
illustration of how the value of extended refueling range is estimated, a typical small light
truck model has an average fuel tank size of approximately 20 gallons.ee Based on a
California Air Resources Board Study, the average fuel purchase is approximately 55% of
tank volume.34 Therefore, increasing this model's actual on-road fuel economy from 24 to 25
mpg would extend its driving range from 216 miles (= 9 gallons x 24 mpg) to 225 miles (= 9
gallons x 25 mpg). Assuming that this vehicle is driven 12,000 miles/year, this reduces the
number of times it needs to be refueled each year from 55.5 (= 12,000 miles per year 7216
miles per refueling) to 53.3 (= 12,000 miles per year / 225 miles per refueling), or by 2.2
refueling s per year.
Weighted by the nationwide mix of urban (about 2/3) and rural (about 1/3) driving and
average vehicle occupancy for all driving trips (1.6 persons), the DOT-recommended value of
travel time per vehicle-hour is $24.00 (in 2006 dollars).ff Assuming that locating a station and
filling up requires five minutes, the annual value of time saved as a result of less frequent
refueling amounts to $4.40 (calculated as 5/60 x 2.2 x $24.00). This calculation is repeated
for each future calendar year that light trucks of each model year affected by the alternative
fuel economy standards considered in this rule would remain in service. Like fuel savings
and other benefits, however, the value of this benefit declines over a model year's lifetime,
because a smaller number of vehicles originally produced during that model year remain in
service each year, and those remaining in service are driven fewer miles. The detailed values
used in this calculation by each agency are available in EPA's DRIA and NHTSA's PRIA
respectively.
The agencies' estimate of benefits from less frequent refueling is subject to several
sources of uncertainty.
First, this analysis assumes that manufacturers will not adjust fuel tank capacities
downward (from the current average of 19.3 gallons) when they improve the fuel economy of
their vehicle models, so that the entire increase in fuel economy will be reflected in increased
ee Based on the Volpe Model Market Data file for Model Year 2011, average tank volumes for cars and trucks
are 16.6 gallons and 23.0 gallons, respectively. This produces a production weighted average of 19.3 gallons.
ff The hourly wage rate during 2006 is estimated to be $24.00. Personal travel (94.4% of urban travel) is valued
at 50 percent of the hourly wage rate. Business travel (5.6% or urban travel) is valued at 100 percent of the
hourly wage rate. For intercity travel, personal travel (87%) is valued at 70 percent of the wage rate, while
business travel (13%) is valued at 100 percent of the wage rate. The resulting values of travel time are $12.67
for urban travel and $17.66 for intercity travel, and must be multiplied by vehicle occupancy (1.6) to obtain the
estimate value of time per vehicle hour.
4-45
-------
Draft Joint Technical Support Document
driving range. Should manufacturers choose to downsize fuel tanks, and all other factors have
been estimated with no error, the current estimates of refueling benefits would be overstated.
On the other hand, vehicle space, utility and value could increase and vehicle weight will
decrease, improving fuel economy and CC>2 emissions (all other things being equal). In the
context of the rule, this will decrease the cost of compliance with the proposed standards.
Second, the agencies' analysis assumes that fuel purchases average 55% of fuel tank
capacity. However, as shown in the California Air Resource Board (CARB) report, refueling
patterns vary. Moreover, the 55% estimate implies that drivers, on average, are either
refueling when nearly a half tank of gas remains in their vehicles, or that they are habitually
not filling their tanks. Since many drivers only refuel when their tanks are very low, and
since many drivers habitually fill up their tanks, this in turn implies that many drivers in the
CARB study are refueling when their tanks are still well above 50% full. Certainly instances
of this type of behavior occur, but the CARB study implies that it is the norm. Behavior that
maximizes the number of fill-ups implies a very low value of time, but it is also possible that
the results of the CARB study are not representative of typical behavior across the country.
While based on field data, this estimate may thus overestimate the impact of refueling
benefits.
Third, the agencies' estimate of refueling benefits assumes that refueling stops involve
the same number of vehicle occupants as the overall average for all vehicle trips (1.6 persons).
To the extent that drivers refuel while doing other errands or in advance of picking up
passengers, this figure may overestimate the typical vehicle occupancy during refueling, and
thus the total savings in refueling time. Similarly, the hourly value used to estimate the
economic value of savings in refueling time reflects the typical mix of personal and business
travel purposes, and drivers are likely to assign different values to their time when traveling
for these different purposes. To the extent that drivers seek to refuel when traveling for
purposes that typically use less valuable time, the hourly value used in the agencies' analysis
may overstate the benefits from saving refueling time.
Finally, the agencies assume that both finding and using a refueling station takes, on
average, five minutes. There are few, if any, data sources on average refueling time, and this
estimate is subject to significant uncertainty.
For these reasons, the agencies' estimate of savings in refueling time is uncertain. As
noted in section II of the preamble, the agencies seek comment or data on of each of the
assumptions they use to estimate this benefit, including alternative empirical estimates of the
parameters used in its calculations.
4.2.12 Discounting Future Benefits and Costs
Discounting future fuel savings and other benefits is intended to account for the
reduction in their value to society when they are deferred until some future date, rather than
received immediately. The discount rate expresses the percent decline in the value of these
benefits - as viewed from current perspective - for each year they are deferred into the future.
In evaluating the benefits from alternative increases in fuel economy and GHG standards for
4-46
-------
Economic Assumptions Used in the Agencies' Analyses
MY 2012-2016 passenger car sand light tracks, EPA and NHTSA have employed a discount
rate of 3 percent per year.
The primary reason that the agencies have selected 3 percent as the appropriate rate
for discounting future benefits from increased fuel economy and GHG standards is that most
or all of vehicle manufacturers' costs for complying with improved fuel economy and GHG
standards are likely to be reflected in higher sales prices for their new vehicle models. By
increasing sales prices for new cars and light trucks, CAFE regulation will thus primarily
affect vehicle purchases and other private consumption decisions.
Both economic theory and OMB guidance on discounting indicate that the future
benefits and costs of regulations that mainly affect private consumption should be discounted
at the consumption rate of time preference. 5 OMB guidance indicates that savers appear to
discount future consumption at an average real (that is, adjusted to remove the effect of
inflation) rate of about 3 percent when they face little risk about its likely level, which makes
it a reasonable estimate of the consumption rate of time preference.36 Thus EPA and NHTSA
have employed the 3 percent rate to discount projected future benefits and costs resulting
from improved fuel economy and GHG standards for MY 2012-2016 passenger cars and light
trucks.
Because there is some uncertainty about the extent to which vehicle manufacturers
will be able to recover their costs for complying with improved fuel economy and GHG
standards by increasing vehicle sales prices, however, the agencies have also tested the
sensitivity of these benefit and cost estimates to the use of a higher percent discount rate.
OMB guidance indicates that the real economy-wide opportunity cost of capital is the
appropriate discount rate to apply to future benefits and costs when the primary effect of a
regulation is ".. .to displace or alter the use of capital in the private sector," and estimates that
this rate currently averages about 7 percent.37 Thus the agencies have also tested the
sensitivity of its benefit and cost estimates for alternative MY 2012-2016 fuel economy and
GHG standards to the use of a 7 percent real discount rate.
4-47
-------
Draft Joint Technical Support Document
Reference
1 EPA, Fuel Economy Labeling of Motor Vehicles: Revisions To Improve Calculation of Fuel Economy
Estimates; Final Rule, 40 CFR Parts 86 and 600, Federal Register, December 27, 2006, pp. 77872-77969,
http://www.epa.gov/fedrgstr/EPA-AIR/2006/December/Day-27/a9749.pdf.
2 EPA, Final Technical Support Document: Fuel Economy Labeling of Motor Vehicle Revisions to Improve
Calculation of Fuel Economy Estimates, Office of Transportation and Air Quality EPA420-R-06-017 December
2006, Chapter II, http://www.epa.gov/fueleconomy/420r06017.pdf.
3 Energy Information Administration, Annual Energy Outlook 2009, Revised Updated Reference Case (April
2009), Table 12. Available at http://www.eia.doe.gov/oiaf/servicerpt/stimulus/excel/aeostimtab 12.xls (last
accessed July 26, 2009). EIA's Updated Reference Case reflects the effects of the American Reinvestment and
Recovery Act of 2009, as well as the most recent revisions to the U.S. and global economic outlook.
4 http://www.eia.doe.gov/oiaf/aeo/forecast.html (last accessed July 26, 2009), and Table 16,
http://www.eia.doe.gov/oiaf/aeo/pdf/tablel6.pdf (last accessed July 26, 2009).
5 Lu, S., NHTSA, Regulatory Analysis and Evaluation Division, "Vehicle Survivability and Travel Mileage
Schedules," DOT HS 809 952, 8-11 (January 2006). Available at http://www-nrd.nhtsa.dot.gov/pdf/nrd-
30/NCSA/Rpts/2006/809952.pdf (last accessed April 20, 2008). These updated survival rates suggest that the
expected lifetimes of recent-model passenger cars and light trucks are 13.8 and 14.5 years.
6 FHWA, Highway Statistics, Summary to 1995, Table vm201at
http://www.fhwa.dot.gov/ohim/summarv95/vm201a.xlw , and annual editions 1996-2005, Table VM-1 at
http://www.fhwa.dot.gov/policv/ohpi/hss/hsspubs.htm (last accessed April 20, 2008).
7 Sorrell, S. and J. Dimitropoulos, 2007. "UKERC Review of Evidence for the Rebound Effect, Technical Report
2: Econometric Studies", UKERC/WP/TPA/2007/010, UK Energy Research Centre, London, October and
Greening, L.A., D.L. Greene and C. Difiglio, 2000. "Energy Efficiency and Consumption - The Rebound Effect
- A Survey", Energy Policy, vol. 28, pp. 389-401.
8 Greening et al. 2000.
9 Pickrell, D. and P. Schimek, 1999. "Growth in Motor Vehicle Ownership and Use: Evidence from the
Nationwide Personal Transportation Survey", Journal of Transportation and Statistics, vol. 2, no. 1, pp. 1-17.
10 Small, K. and K. Van Dender, 2007a. "Fuel Efficiency and Motor Vehicle Travel: The Declining Rebound
Effect", The Energy Journal, vol. 28, no. 1, pp. 25-51.
11 Small, K. and K. Van Dender, 2007b. "Long Run Trends in Transport Demand, Fuel Price Elasticities and
Implications of the Oil Outlook for Transport Policy," OECD/ITF Joint Transport Research Centre Discussion
Papers 2007/16, OECD, International Transport Forum.
12 Greene Report to EPA, March 24, 2009. Note, this report has been submitted for peer review. Completion of
the peer review process is expected prior to the final rule.
13 Small Report to EPA, June 12, 2009.
14 FHWA. 1997 Federal Highway Cost Allocation Study, http://www.fhwa.dot.gov/policy/hcas/final/index.htm
(last accessed July 29, 2009).
15 Federal Highway Administration, 1997 Federal Highway Cost Allocation Study,
http://www.fhwa.dot.gov/policy/hcas/final/index.htm, Tables V-22, V-23, and V-24 (last accessed July 27,
2009).
16 U.S. Bureau of Economic Analysis, U.S. International Transactions Accounts Data, as shown on June 24,
2009.
17 U.S. Department of Energy, Annual Energy Review 2008, Report No. DOE/EIA-0384(2008), Tables 5.1 and
5.13c, June 26, 2009.
4-48
-------
Economic Assumptions Used in the Agencies' Analyses
8 Leiby, Paul N. "Estimating the Energy Security Benefits of Reduced U.S. Oil Imports," Oak Ridge National
Laboratory, ORNL/TM-2007/028, Final Report, 2008.
19 Peer Review Report Summary: Estimating the Energy Security Benefits of Reduced U.S. Oil Imports, ICE,
Inc., September 2007.
20 Gately, Dermot 2004. "OPEC's Incentives for Faster Output Growth," The Energy Journal, 25(2):75-96,
"What Oil Export Levels Should We Expect From OPEC?" The Energy Journal, 28(2): 151-173, 2007
21 Argonne National Laboratories, The Greenhouse Gas and Regulated Emissions from Transportation (GREET)
Model, Version 1.8, June 2007, available at http://www.transportation.anl.gov/software/GREET/index.html
(last accessed April 20, 2008).
22 U.S. Environmental Protection Agency (U.S. EPA). 2008a. Regulatory Impact Analysis, 2008 National
Ambient Air Quality Standards for Ground-level Ozone, Chapter 6. Office of Air Quality Planning and
Standards, Research Triangle Park, NC. March. Available at .
23 U.S. Environmental Protection Agency (U.S. EPA). 2009a. Regulatory Impact Analysis: National Emission
Standards for Hazardous Air Pollutants from the Portland Cement Manufacturing Industry. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. April. Available on the Internet at
.
24 U.S. Environmental Protection Agency (U.S. EPA). 2009b. Proposed NO2 NAAQS Regulatory Impact
Analysis (RIA). Office of Air Quality Planning and Standards, Research Triangle Park, NC. April. Available
on the Internet at http://www.epa.gov/ttn/ecas/regdata/RIAs/proposedno2ria.pdf.
25 U.S. Environmental Protection Agency (U.S. EPA). 2008b. Technical Support Document: Calculating Benefit
Per-Ton estimates, Ozone NAAQS Docket #EPA-HQ-OAR-2007-0225-0284. Office of Air Quality Planning
and Standards, Research Triangle Park, NC. March. Available on the Internet at .
26 Fann, N. et al. (2009). The influence of location, source, and emission type in estimates of the human health
benefits of reducing a ton of air pollution. Air Qua! Atmos Health. Published online: 09 June, 2009.
27 Pope, C.A., III, R.T. Burnett, M.J. Thun, E.E. Calle, D. Krewski, K. Ito, and G.D. Thurston. 2002. "Lung
Cancer, Cardiopulmonary Mortality, and Long-term Exposure to Fine Particulate Air Pollution." Journal of the
American Medical Association 287:1132-1141.
28 Laden, F., J. Schwartz, F.E. Speizer, and D.W. Dockery. 2006. "Reduction in Fine Particulate Air Pollution
and Mortality." American Journal of Respiratory and Critical Care Medicine 173:667-672. Estimating the
Public Health Benefits of Proposed Air Pollution Regulations. Washington, DC: The National Academies Press.
29 Mrozek, J.R., and L.O. Taylor. 2002. "What Determines the Value of Life? A Meta-Analysis." Journal of
Policy Analysis and Management 21(2):253-270.
30 Viscusi, V.K., and J.E. Aldy. 2003. "The Value of a Statistical Life: A Critical Review of Market Estimates
throughout the World." Journal of Risk and Uncertainty 27(l):5-76.
31 Kochi, I., B. Hubbell, and R. Kramer. 2006. An Empirical Bayes Approach to Combining Estimates of the
Value of Statistical Life for Environmental Policy Analysis. Environmental and Resource Economics. 34: 385-
406.
32 U.S. Environmental Protection Agency (U.S. EPA). 2000. Guidelines for Preparing Economic Analyses.
EPA 240-R-00-003. National Center for Environmental Economics, Office of Policy Economics and
Innovation. Washington, DC. September. Available on the Internet at
.
33 See http://ostpxweb.dot.gov/policv/Data/VOT97guid.pdf and
http://ostpxweb.dot.gov/poricv/Data/VOTrevisionl_2-l l-03.pdf
4-49
-------
Draft Joint Technical Support Document
34 California Environmental Protection Agency, Air Resources Board. Draft Assessment of the Real-World
Impacts of Commingling California Phase 3 Reformulated Gasoline. August 2003
35 Id.
36 Office of Management and Budget, Circular A-4, "Regulatory Analysis," September 17, 2003, 33. Available
at http://www.whitehouse.gov/omb/circulars/a004/a-4.pdf (last accessed July 24, 2009).
37 Id.
4-50
------- |