Automotive Features: Mass Impact and

Deployment Characterization
by
Stephen M. Zoepf
B.S. Electrical Engineering, Massachusetts Institute of Technology,
2001
Submitted to the Engineering Systems Division
in partial fulfillment of the requirements for the degree of
Master of Science in Technology and Policy
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
June 2011

c Massachusetts Institute of Technology 2011. All rights reserved.

Author . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Engineering Systems Division
May 13, 2011

Certified by . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Professor John B. Heywood
Professor of Mechanical Engineering
Sun Jae Professor, Emeritus
Thesis Supervisor

Accepted by . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Professor Dava J. Newman
Professor of Aeronautics and Astronautics and Engineering Systems
Director, Technology and Policy Program
2
 Automotive Features: Mass Impact and Deployment
Characterization
by
Stephen M. Zoepf

Submitted to the Engineering Systems Division

on May 13, 2011, in partial fulfillment of the
requirements for the degree of
Master of Science in Technology and Policy

Abstract
Passenger car use is a major driver of greenhouse gas (GHG) emissions and fossil fuel
consumption in the United States. Vehicles continue to incorporate increasing levels
of technology, these advances do not translate directly into improved fuel economy.
Vehicle weight, interior volume and performance have all grown substantially in the
past 30 years, as has feature content.
This thesis shows that safety features, emissions controls, and optional equipment
account for a total mass that mirrors growth in vehicle mass during this time pe-
riod. Chief among these are optional features designed to improve the comfort and
convenience of passenger cars.
This thesis also examines historical deployment rates of vehicle features. Safety
features and emissions controls achieve faster deployment rates than other optional
features. While these features are those most governed by regulation, it is not clear
that regulations push technology deployment rates higher. Automotive product de-
velopment is complex and features require significant time to overcome deployment
constraints. This lag time, from first production use to most rapid deployment across
the vehicle fleet, is found to be exponentially decreasing for all feature types and has
dropped to approximately a decade.
These analyses provide two countering assessments. New vehicles will continue to
grow heavier due to the continued incorporation of new features, but technology that
may improve overall efficiency can be brought to market ever faster.

Thesis Supervisor: Professor John B. Heywood

Title: Professor of Mechanical Engineering
Sun Jae Professor, Emeritus

3
4
Acknowledgments
This thesis would not have been possible without the help and support of many people
from the MIT community and beyond.
First and foremost I would like to offer my sincere thanks to Prof. John Heywood
for his guidance and support these past two years as my advisor. I am humbled by
his immense knowledge, talent and kindness.
I would like to thank my fellow students from the Sloan Automotive Laboratory:
Don MacKenzie, Fernando de Sisternes, Lynette Cheah, Valerie Karplus, David Keith,
Irene Berry, Eriko Nishimura and Mike Khusid. My work has built upon much of
their own and has been improved greatly by their feedback. Many thanks also to Prof.
Randy Kirchain, Rich Roth and Elisa Alonso from the Materials Systems Laboratory
for their contributions.
I would also like to thank the staff of the U.S. Department of Transportation and
the Volpe Center for their information and feedback.
The faculty and staff of the Engineering Systems Division and the Technology and
Policy Program have been a great help at every step. It has been a pleasure to learn
from their vast experience these past two years.
Finally, I would like to thank my wife, Irma, for her unwavering support and
encouragement. She is a pillar of strength and I am forever indebted for the sacrifices
she has made in accompanying me to graduate school.
This work has been sponsored by Ford Motor Company and the MIT Portugal
Program.

5
6
Contents

1 Introduction 15
1.1 Vehicle Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.2 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.3 Research Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.4 Methodology and Comments . . . . . . . . . . . . . . . . . . . . . . . 21
1.4.1 Data Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
1.4.2 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

2 Vehicle Features 25
2.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2 Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.3 Safety Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.4 Emissions Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.5 Optional Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.6 Secondary Mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

3 Measuring Technological Sophistication 39

3.1 BMEP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.2 PSFI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.3 Adjusted PSFI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.4 Conclusions and Tradeoffs . . . . . . . . . . . . . . . . . . . . . . . . 43

7
4 Deployment Rates of Vehicle Technology 47
4.1 Common Models of Diffusion/Deployment . . . . . . . . . . . . . . . 49
4.2 Applying Regression . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.3 Prior Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.4 Important Characteristics and Secondary Regression . . . . . . . . . 51
4.4.1 Maximum Take Rate . . . . . . . . . . . . . . . . . . . . . . . 51
4.4.2 Maximum Growth Rate . . . . . . . . . . . . . . . . . . . . . 54
4.4.3 Developmental Lag Time . . . . . . . . . . . . . . . . . . . . . 57
4.5 Alternative Methodology: Z-Curves and Logistic Approximation . . . 59
4.6 Differences Among Manufacturers . . . . . . . . . . . . . . . . . . . . 62
4.6.1 Manufactuer-Specific Action and Fleet Impact . . . . . . . . . 62
4.6.2 Strategies Within Manufacturers . . . . . . . . . . . . . . . . 65
4.7 Supply Side Constraints on Deployment . . . . . . . . . . . . . . . . 66
4.8 Impact of Regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
4.8.1 Frontal Air Bags . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.8.2 Anti-lock Brakes . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.8.3 Multiple valves and Variable Valve Timing . . . . . . . . . . . 72
4.9 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

5 Findings, Applications and Conclusions 75

5.1 Vehicle Mass Findings . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.2 Feature Deployment Findings . . . . . . . . . . . . . . . . . . . . . . 76
5.3 Applications for Policy and Vehicle Fleet Models . . . . . . . . . . . . 77
5.4 Other Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
5.4.1 Alternative Fuels and Parallel Innovation . . . . . . . . . . . . 80
5.4.2 Disruptive Innovation . . . . . . . . . . . . . . . . . . . . . . . 81
5.5 Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

A Tables 85

B Mass Decompounding Studies 89

8
C Safety Features: Take Rates and Regression 91

D Powertrain Features: Take Rates and Regression 93

E Comfort and Convenience Features: Take Rates and Regression 95

F Other: Features with Limited Data 99

G Regression Data Points 103

9
10
List of Figures

1-1 Average performance, size and fuel economy of new U.S. cars, 1977-
2008, indexed to 1977 averages. Graphic from MacKenzie, 2009 using
data from U.S. EPA, 2008 . . . . . . . . . . . . . . . . . . . . . . . . 18
1-2 Growth in mass of vehicles from four major classes from [Glennan, 2007] 20

2-1 The average mass of passenger cars in the United States has climbed
since the mid-1980s. Source: [EPA, 2010] . . . . . . . . . . . . . . . . 26
2-2 Average mass added by safety regulations from 1968-2001 as reported
by U.S. DOT in report HS 809 834. . . . . . . . . . . . . . . . . . . . 28
2-3 Estimated mass of emissions control equipment by year. . . . . . . . . 31
2-4 Mass and take rate of ten features with highest fleetwide impact. Lines
of isometric fleet impact at levels of 20, 10, 5 and 3 kg/car are also
shown. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2-5 Estimated mass impact of optional features by year. Dashed lines show
projected future impact using fits of existing take rate data, while solid
line is a linear fit of historic data. . . . . . . . . . . . . . . . . . . . . 35
2-6 Mass of passenger cars 1975-2010 and weight attributed to Safety,
Emissions and Comfort/Convenience features (Secondary mass included). 36

3-1 Trends in BMEP for U.S. engines from 2000 - 2008 [Heywood and Welling, 2009] 41
3-2 Performance Size Fuel Economy Index, 1975 - 2005 from [An and DeCicco, 2007].
This metric of vehicle sophistication has grown with remarkable linear-
ity since 1975. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3-3 Traditional PSFI vs. Adjusted PSFI growth from 1977 - 2010. . . . . 44

11
4-1 Functional form of the diffusion of innovations as envisioned by [Rogers, 2003] 48
4-2 Functional form of the Bass diffusion model as envisioned by [Sterman, 2000] 49
4-3 Important characteristics of a logistic function. . . . . . . . . . . . . . 52
4-4 Technology supersession in fuel injection and engine management as
envisioned by [Amey, 1995] . . . . . . . . . . . . . . . . . . . . . . . . 53
4-5 Histrogram of maximum feature growth rates. . . . . . . . . . . . . . 55
4-6 Histogram of maximum feature growth rate by category. . . . . . . . 56
4-7 Historical phase-in time of all features. . . . . . . . . . . . . . . . . . 58
4-8 Historical decrease in phase-in time of features by category. . . . . . . 59
4-9 Product variety and model changes by region, adapter from [Clark and Fujimoto, 1991]. 60
4-10 Stages of deployment of new vehicle technology. Image from [Bandivadekar, 2008],
adapter from [Schafer et al., 2006] . . . . . . . . . . . . . . . . . . . . 61
4-11 Differences among automotive manufacturers in phase-in of fuel injection. 63
4-12 Differences among automotive manufacturers in phase-in of multivalve
engines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4-13 Differences among automotive manufacturers in phase-in of variable
valve timing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4-14 Differences in maximum growth rates between regulated and non-regulated
features. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4-15 Maximum growth rate vs. number of passenger cars equipped at time
regulation is announced. . . . . . . . . . . . . . . . . . . . . . . . . . 70
4-16 Adoption of ABS and timing of FMVSS 126 NPRM. . . . . . . . . . 71

12
List of Acronyms and Common Terms
ABS Anti-lock Brakes

BAU Business As Usual, a term used to describe a modeling scenario where present
conditions are presumed to carry forward unchanged.

BEV Battery Electric Vehicle

BMEP Brake Mean Effective Pressure

CAFE Corporate Average Fuel Economy, the sales-weighted average fuel economy
of the vehicles produced by a manufacturer in a given year.

CARB California Air Resources Board

Criteria Pollutant Six common air pollutants regulated by the U.S. EPA: Ozone,
Particulate Matter, Carbon Monoxide, Nitrogen Oxides, Sulfur Dioxide and
Lead.

DOT U.S. Department of Transportation

EIA Energy Information Administration

EISA Energy Independence and Security Act

EPA U.S. Environmental Protection Agency

EPCA Energy Policy Conservation Act

ERFC Emphasis on Reduction of Fuel Consumption

ESC Electronic Stability Control

Feature Equipment on a vehicle not strictly necessary for basic vehicle functionality
but required by law to improve safety or emissions, or optionally added by
automakers to improve the comfort, convenience or consumer appeal of a vehicle.

FI Fuel Injection

13
FMVSS Federal Motor Vehicle Safety Standard

GHG Greenhouse Gas

ICE Internal Combustion Engine

LDV, Light Duty Vehicle Passenger cars and light trucks with a gross vehicle
weight under 8,500 lbs

MPG Miles per gallon

MSRP Manufacturer’s Suggested Retail Price

NHTSA U.S. National Highway Traffic Safety Administration

NPRM Notice of Proposed Rulemaking

OEM Original Equipment Manufacturer

PHEV Plug-in Hybrid Electric Vehicle

PSFI Performance-size-fuel economy index

SI Spark-Ignition, generally referring to gasoline-powered engines

SUV Sport Utility Vehicle

Take Rate The percentage of new vehicles or passenger cars equipped with a specific
feature in a given year.

VMT, VKT Vehicle Miles Traveled, Vehicle Kilometers Traveled: An expression

for the average distance traveled by each vehicle in a given fleet.

VVT Variable Valve Timing

14
Chapter 1

Introduction

Chapter 1
Chapter 1 outlines the importance of transportation in petroleum and
GHG contexts, discusses the importance of vehicle features in the
evolution of the passenger car, and outlines the themes to be discussed
in the remainder of the document.

For nearly a century the automobile has been the dominant form of personal
transportation in the United States. The passenger car has made inexpensive, reliable
transportation accessible to virtually every citizen and has profoundly affected the
development of modern society.
At the same time, transportation has also become a dominant source of greenhouse
gas (GHG) emissions and fossil fuel consumption in the United States. As shown
in Table 1.1, the Transportation sector represents a growing driver of energy and
fossil fuel use. The Transportation sector is now responsible for 71% of petroleum
consumption and 34% of GHG emissions. Since 1999 the Transportation sector has
exceeded the Industrial sector as the largest emitter of greenhouse gases.[EIA, 2010]
The Transportation sector derives 94% of its energy from Petroleum. 45% of
this petroleum is used to produce motor gasoline, the vast majority of which is used
to fuel Light Duty Vehicles (LDVs). The GHG emissions and fossil fuel consump-
tion driven by the Transportation sector have been modeled by numerous groups.

15
Table 1.1: Data from EIA showing historical energy consumption and GHG emissions
by sector.

The fundamental factors driving such consumptions are shown below, adapter from
[MIT, 2008].

GHGemissions = LP K ∗ V KT ∗ F I

Where,
GHGemissions = Greenhouse Gas Emissions (tons/year)
LP K = Liters per Kilometer (L/100km)
V KT = Vehicle Kilometers Traveled (VKT in km/year)
F I = GHG intensity of Fuel (GHG tons/liter of fuel)

16
 While there are interactions among these factors, LPK is the only factor here
reliant on the technical sophistication of vehicles themselves while FI and LPK are
primarily a product of infrastructure and usage.

1.1 Vehicle Evolution

During its development over the past century the passenger car has evolved from a
rudimentary, self-propelled ”horseless carriage” to a sophisticated device capable of
traveling hundreds of miles at high speeds. All new vehicles sold in the United States
incorporate seat belts and at least two airbags; virtually all new vehicles include air
conditioning and automatic transmissions to increase driver comfort and many also
include navigation systems, moonroofs, and a bevy of other electronic aids.
Gains in performance, size and fuel economy are relatively straightforward to
analyze, as these factors have been carefully tracked since 1975 in the EPA Fuel
Economy trends report. As shown in Figure 1-1 below fuel economy grew rapidly
from approximately 1977-1987 and has since remained relatively constant. Vehicle
performance has continually increased since the mid-1980s and vehicle size, measured
using interior volume, has also grown modestly over the same time period.
However, vehicle capability has improved in other ways as well. Vehicles are
safer, emit fewer criteria pollutants per mile traveled, and include a broad array of
equipment designed to improve dynamic capabilities provide the driver with a high
level of comfort and information.
[Cheah, 2010] highlights the importance of vehicle features by comparing two ve-
hicles: a 2008 Audi A4 and a 2006 Mazda3 Touring, as shown in Table 1.2. The two
vehicles are nearly identical in performance, interior space, exterior dimensions and
highway fuel economy. However, the Audi has a significantly longer list of standard
equipment, performs better in several crash tests, is 50% more expensive and weighs
345 kg more than the Mazda. Cheah attributes the mass increase primarily to the
higher feature content of the Audi. However, this is only a single data point: a closer
examination is needed to address the fleet-wide impact of features on the mass of

17
Figure 1-1: Average performance, size and fuel economy of new U.S. cars, 1977-2008,
indexed to 1977 averages. Graphic from MacKenzie, 2009 using data from U.S. EPA,
2008

vehicles.
Figure 1-2 graphically illustrates four case studies of growth in the mass of vehicles
of various classes.

1.2 Motivation
Existing work by [An and DeCicco, 2007, MacKenzie, 2009, Knittel, 2009] has in-
vestigated the tradeoffs among performance, size and fuel economy and consumer
willingness to pay for said attributes. However, as shown by Cheah in the previous
section, vehicles similar in these common metrics may still be substantially differ-
ent in their ability to deliver functionality to the purchaser in the form of additional
features. The need to quantify this tradeoff dimension has, to date, remained unfilled.
Performing an analysis of vehicle features pays dividends in other ways. While
features such as fuel injection, multiple valves per cylinder and variable valve timing

18
Table 1.2: Audi A4 vs. Mazda3: A comparison of vehicles with similar primary
attributes but very different features. Adapted from [Cheah, 2010].
Year, Model 2008 Audi A4 2.0T Quattro 2006 Mazda3 s Touring

Body style 4-dr sedan (seats 5) 4-dr sedan (seats 5)

Wheelbase 104.3 in 103.9 in
length x width x height 180.6 x 69.8 x 56.2 in3 178.7 x 69.1 x 57.7 in3
Overall volume (l*w*h) 11.61 m3 11.68 m3
Horsepower 200 hp 160 hp
Power-to-weight ratio 0.057 hp/lb 0.058 hp/lb
0-60mph (est.) 9.58 sec 9.59 sec
Engine displacement 2.0L I4 2.3L I4
Front head, leg room 37.9 in, 41.3 in 39.1 in, 41.9 in
Luggage capacity 13.4 f t3 11.4f t3
City/hwy (adj.) fuel econ- 22/31 MPG 26/32 MPG
omy
Features 16“ wheels, AWD, power seat ad- 17“ wheels, FWD, manual
justment, sunroof, heated mirrors, seat adjustment, sunroof op-
dual zone climate controls, 10 speak- tional
ers + subwoofer, traction and stabil-
ity control
NHTSA crash test ratings 4, 4, 5, 4, 4 stars 4, 4, 3, 3, 4 stars
driver, passenger, side im-
pact front, rear, rollover
MSRP $28,900 $17,600
Curb weight 1,595 kg (3,516 lb) 1,251 kg (2,758 lb)

(VVT) are tracked by the U.S. EPA, the deployment rates of these features are
not strictly the product of consumer demand and automaker ability to deploy such
technology. To varying degrees each of these technologies has been driven by criteria
pollutant regulations. Because of this, viewing the deployment of such features in
the broader context of safety regulation-driven and optional features serves to better
characterize deployment of technology in the automotive fleet.

1.3 Research Objectives

This thesis serves to address the following primary research questions described below.

Mass Impact What has been the impact of safety and emissions requirements and

19
Figure 1-2: Growth in mass of vehicles from four major classes from [Glennan, 2007]

other optional features on the mass of automobiles in the United States?

Trends in Features What is the trend in deployment of features in each of these

areas?

Technical Sophistication How can information about vehicle features be used to

refine existing measures of technical sophistication?

Deployment Scenarios What do past trends in feature deployment rates tell us

about possible future deployment of vehicle features?

20
1.4 Methodology and Comments

This thesis is composed of two distinct analyses. Chapter 2 combines estimated

deployment data of vehicle features of three categories (safety equipment, emissions
controls, and optional features) and estimated mass of each feature to calculate the
total mass added to an average passenger car as a result of this equipment.
Chapter 4 analyses the characteristics of the deployment of vehicle features for
which time series deployment data is available and identifies key parameters in the
historical deployment of approximately 35 unique vehicle features at fleet-wide and
manufacturer levels.

1.4.1 Data Sources

Estimates of mass impact of legally mandated safety features are drawn directly
from analysis performed by NHTSA in DOT HS 809 834.[DOT, 2004a]. This study
addresses both the deployment rate of specific features and their weight contribution
year-over-year. Learning effects are addressed: vehicles from a range of dates are
torn down and the mass contribution from each NHTSA regulation is individually
addressed.
Specific equipment required for emissions controls from 1970 - 1990 has been
interpreted from a timeline of EPA regulations [EPA, 2011]. The mass contributions
from each of the regulations during this period are drawn from vehicle teardown data
discussed below.
Take rates of some optional features is drawn from Wards Factory Installed databases.
[Ward’s, 2010b, Ward’s, 2010a]. Of the 74 optional features analyzed, 34 of these fea-
tures had historical take rate data available. Where option take rates are unavailable,
yearly take rates have been estimated based on feature introduction dates and avail-
ability on top-selling 2010 vehicles as discussed in Chapter 2. Of the forty features
estimated using this linear method, half resulted in estimated take rates of less than
10% and with a minimal mass impact.
The mass of each vehicle feature is approximated by summing the masses of indi-

21
vidual vehicle components attributed to the application of a specific feature. Confi-
dential component-level teardown data, including mass, is provided by an unnamed
OEM for four individual passenger cars (two compact and two full-sized). For twenty-
five of these features, mass was calculated as an average value of the applicable equip-
ment from these four vehicles. Where mass data for individual features is unavailable,
an estimated value has been used. Of the features where estimated mass was used,
twenty-four were on features with estimated take rates of less than 10% with minimal
overall impact.

1.4.2 Limitations

Over the course of their development many features incorporate improvements that
result in their implementation with a reduced mass (i.e. electric power steering vs.
hydraulic power steering). [DOT, 2004a] incorporates a learning effect the mass of
specific safety features by tearing down two vehicles a specific number of years apart
and assumes a linear reduction in mass over the intervening years. All other features
are assumed to have a constant mass over time. This assumption is the result of
limitations on data availability. Vehicle teardown data for every year is not available,
so component mass values from 2010 vehicles have been used to determine feature
weights.
Since the teardown data used for the purposes of this analysis is at a component
level, only components that can be directly attributed to a specific feature are in-
cluded. For example, weight values used for power seats include only the switches,
motors and harnesses identified separately in the data set. Gearboxes or other struc-
tural changes to the seats in order to accommodate power assistance are not included.
Because not all components of every system are composed of individually identifiable
parts, the mass values for optional features calculated by this analysis could rea-
sonably be called a lower bound of the possible range of masses of a given optional
feature. This is distinct from the concept of secondary mass, discussed in Chapter
2, which refers to mass resulting from reinforcing other systems to maintain vehicle
performance.

22
 While light trucks represent an increasingly large share of new vehicle sales (40%
in 2009), the use and makeup of the light truck fleet has changed dramatically.
Light trucks, originally envisioned primarily as work vehicles, now include SUVs and
crossover vehicles that have replaced passenger cars in private use. Since it would not
be practical to isolate demographic groups among light truck purchasers, the analysis
performed in this document focuses strictly on passenger cars and does not include
information on light trucks.

23
24
Chapter 2

Vehicle Features

Chapter 2
Chapter 2 dissects the increasing presence of features into those im-
proving safety, reducing emissions, and those offering luxury amenities.
This chapter further quantifies the resultant mass increase in the pas-
senger car fleet and proposes estimates of how fleet fuel consumption
might change in the absence of these features.

2.1 Definitions
For the purposes of this thesis, a feature is defined as any component that is not
strictly necessary for the basic functionality of the vehicle but that is either required
by legal mandate or advertised as providing a benefit to the consumer.
In this section features are categorized as Safety, Emissions, or Optional. Safety
and Emissions equipment are strictly required components as defined by respective
governing agencies. Components with driver and passenger safety implications that
are not currently mandated, such as rollover airbags and park distance control are
classified as optional equipment for the purposes of mass impact, but are classified
as safety features in Chapter 4. Table 2.1 identifies how features have been classified
for the purposes of this analysis.1
1
In this analysis, both curtain airbags and anti-lock brakes are identified as optional features.

25
 Table 2.1: Examples of feature categorization.
Optional Features Safety Emissions
Automatic Transmission Power Brakes Catalytic Converters
Power Steering Frontal Airbags Evaporative Emissions Controls
Int. Wiper / Washers Child Seat Anchors EGR valves
Air conditioning Side Door Beams Fuel injection

2.2 Context
Following a decline in the late seventies and early eighties, the average mass of pas-
senger cars in the United States has climbed steadily as shown in Figure 2-1. While
some of this mass increase is attributable to a modest growth in interior volume over
the same time period, a large portion of this growth in mass is the result of features
that improve the safety, emissions, and comfort/convenience of modern vehicles.

Figure 2-1: The average mass of passenger cars in the United States has climbed since
the mid-1980s. Source: [EPA, 2010]

This chapter proposes a framework for estimating the mass increase in modern
However, both of these features will soon become required safety features. Anti-lock brakes, a
component of most Electronic Stability Control systems, are being phased in from 2009-2012 in
compliance with FMVSS 126. Proposed legislation requiring curtain airbags under FMVSS 208 was
issued during the writing of this document and is yet to be finalized.

26
vehicles that is the result of comfort/convenience features and also identifies estimates
of mass attributable to regulations governing safety and emissions. Secondary mass
estimates (defined in Section 2.6) are applied and the results of these estimates are
summed to produce overall mass impact of features.
These calculations serve two primary purposes. First, using these estimates of
feature mass we can estimate what the mass of current production vehicles might be
without the presence of these additional features. Second, since the feature content of
vehicles continues to grow, we can extrapolate trends in vehicle equipment to predict
the mass of features in future vehicles.

2.3 Safety Features

Modern vehicles incorporate a vast array of technologies to reduce the likelihood of in-
juries and fatalities in the event of a crash. Advanced materials strengthen critical ar-
eas around vehicle occupants, seat belts reduce the likelihood of ejection, and airbags
control occupant deceleration rates. New ways to protect occupants are constantly
under development, and the National Highway Transportation Safety Administration
(NHTSA) continually reviews available technologies. When it deems appropriate,
NHTSA issues a ruling mandating that all vehicle incorporate new technologies.
In 2004 NHTSA compiled a report [DOT, 2004a] that quantified the mass added
to passenger cars and light trucks as a result of the cumulative impact of safety
regulations from 1968-2001. According to this report, 2001 model year passenger
cars were, on average, 57 kg (125.44 lbs) heavier as a result of these regulations.
While a subsequent report has not been released, a regression of this data suggests
that weight attributable to NHTSA regulations has risen to 62 kg (136.4 lbs) in 2010
as shown in Figure 2-2 below and, if current trends are maintained, would exceed 100
kg (220 lbs) by 2050.
This analysis incorporates only features that are explicitly mandated by NHTSA
requirements but were not in widespread use prior to regulation. For example, rear
window defoggers, although now required, were already in widespread use at the time

27
Figure 2-2: Average mass added by safety regulations from 1968-2001 as reported by
U.S. DOT in report HS 809 834.

the regulation governing them was issued and are not included here.

Other features that have an impact on safety are classified as optional features
for the purposes of this analysis. For example, while NHTSA does not specifi-
cally track driveway incidents as they occur on private property, [NHTSA, 2011]
and [Safe Kids USA, 2011] discuss the large number of children involved in such inci-
dents each year. Ultrasonic parking assistance (Park Distance Control) and rearview
cameras have the potential to reduce the number of such accidents. Since NHTSA
regulations are frequently updated, features such as this may become required equip-
ment in the future. In the meantime, however, they are classified as optional.

28
2.4 Emissions Controls
Since the first Clean Air Act in 1970, the Environmental Protection Agency has in-
troduced a series of regulations governing automotive emissions of criteria pollutants.
However, the EPA has not published a report that performs a tear-down analysis of
the cumulative impact of its legislation comparable to the work published in DOT
report HS 809 834.
To simulate the results of a comprehensive teardown report this thesis identifies
individual vehicle components attributable to EPA regulations in the Bill of Ma-
terials (BOM) of four typical passenger cars and then allocates the average mass
to specific regulatory milestones from 1970 - 1990 published by the EPA on its
webpage.[EPA, 2011] The allocation of these specific technologies to the milestones
was confirmed as plausbile in an interview with an EPA staffperson in 2010. A sum-
mary of this analysis is shown in Table 2.2.

Table 2.2: Emissions control legislation with associated components and mass.
Date Regulation Components Introduced Estimated
Mass (kg)
1970 Clean Air Act PCV (Positive Crankcase 1.75
Ventilation) valves and
early evaporative canisters.
1972 NOx Standards EGR (Exhaust Gas Recir- 0.42
culation) Valves
1975 First generation catalytic 4.17
converters
1981 Amended Clean Air Act 3-way catalytic converters 4.17
1990 Clean Air Act Enhanced evaporative emis- 3.03
sions controls and electronic
engine management
1994-1997 Clean Air Act ”Tier I” various
1994 Clean Air Act Amendments On board diagnostics
1994-2003 CARB LEV I various
1999-2003 Transitional NLEV Pro- various
gram
2004-2009 Clean Air Act ”Tier II” various
2004-2010 CARB LEV II various

As shown in Table 2.2 the EPA is not the only body regulating automotive emis-

29
sions in the United States. Since 1990 the California Air Resource Board (CARB)
has introduced a series of regulations under the Low Emissions Vehicle (LEV) pro-
gram. Under the Clean Air Act, other states are permitted to adopt either Federal
standards or those set by CARB.
Since regulations may now only apply to vehicles sold in certain states, identifying
the vehicle mass attributable to exclusively to emissions controls has become far
more complex. To further complicate matters, the technologies used to meet these
regulations (such as variable valve timing, multiple valves per cylinder, and direct
injection) offer other benefits such as improved drivability and performance. From
the EPA Tier 2 FRM Regulatory Impact Assessment[EPA, 1999]:

In addition to gains in breathing, the multiple-valve (typically 4-valve)

design allows the spark plug to be positioned closer to the center of the
combustion chamber (as discussed above) which decreases the distance
the flame must travel inside the chamber. In addition, the two streams of
incoming gas can be used to achieve greater mixing of air and fuel, further
increasing combustion efficiency which lowers engine-out HC emissions.

and

Variable valve timing can allow for increased swirl and intake charge
velocity, especially during low load operating conditions where sufficient
swirl and turbulence tend to be lacking. By providing a strong swirl
formation in the combustion chamber, the air-fuel mixture can mix suffi-
ciently, resulting in a faster, more complete combustion, even under lean
air-fuel conditions, thereby reducing emissions.

.
As a result it is not appropriate to ascribe the mass increase from these technical
changes exclusively to emissions regulations.
Since estimating the additional mass of emissions equipment beyond 1990 is quite
complex, for the purposes of this analysis a linear regression has been performed on

30
the mass increases identified from 1970 - 1990 and extrapolated. The resultant mass
increase values have been discussed with EPA staff members who agreed the values
are plausible.
Figure 2-3 shows the calculated mass of emissions equipment from 1970-1990 and
the projected trend based on a linear regression of values during this time period. By
1990, approximately 13.6 kg of emissions equipment had been added to the average
passenger car. Regression suggests that by 2010 this value has nearly doubled to 24.6
kg per passenger car.

Figure 2-3: Estimated mass of emissions control equipment by year.

It is questionable whether a linear extrapolation of past trends is a reasonable

assumption, and it could be argued that the shift from criteria pollutant regulations to
GHG emissions regulations could lead to mass reductions through material selection
and advanced design in an effort to meet GHG reduction targets. However, it is also
possible that GHG reduction requirements will necessity the deployment of BEVs
with large battery packs and lead to an increasing mass requirement attributable to

31
such regulations. In the absence of an alternative, a linear extrapolation has been
used.

2.5 Optional Features

Most new cars sold in the U.S. are equipped with a large number of features designed
to improve the comfort of passengers or provide additional information to the driver.
Air conditioning, automatic transmissions, radios and other features are present in
virtually every passenger car sold today, and dozens of other features are equipped at
lower rates. A full list of features and their availability on top-selling 2010 passenger
cars is shown in Appendix A.
Since not all cars are equipped with all features, the average mass added to each
new vehicle on average by n features is determined according to the following equation:

n
X
M assf eatures = (massi ) ∗ (takeratei )
i=1

The take rate (the percentage of new cars in a given year equipped with a feature)
of many features is tracked by Ward’s Automotive in its Factory Installed equipment
databases. [Ward’s, 2010b, Ward’s, 2010a] The data provided by Ward’s Automotive
are sales-weighted take rates
For features that are not tracked by Ward’s data, yearly equipment rate is based on
the availability of features on top-selling vehicles in early 2010 and the introduction
date of the feature as shown in the formula below. This formula assumes strictly
linear growth rather than the S-curve shaped growth shown to be typical, but since
only two data points are used (date of introduction and take rate in 2010) it is not
reasonable to generate an S-curve.

T akerate2010 = .85 ∗ Vstandard + .25 ∗ Voptional + .01 ∗ Vunavailable

Where Vstandard is defined as the number of top ten selling vehicles that include said

32
feature standard, Voptional is the number of top ten selling vehicles that offer said
feature as an option, and Vunavailable is the number of top ten selling vehicles that do
not offer the feature at all.

Table 2.3: An example of vehicle feature availability in 2010. S = feature is standard

in all trim levels; O = feature is optional or available on some trim levels; N = feature
is not offered on this vehicle.

Chevrolet Malibu
Chevrolet Impala
Toyota Corolla
Toyota Camry

Honda Accord

Nissan Altima

Toyota Prius
Honda Civic

Ford Fusion

Ford Focus
Feature
Center Console / Armrest S S S O S S S S N S

Since many features are included standard and certain trim levels, a feature is
said to be standard only if it is included on all trim levels. Using the example taken
from Table 2.3, Vstandard = 8, Voptional = 1 and Vunavailable = 1 for the feature Center
Console/Armrest in 2010. As a result, the take rate for 2010 would be determined
to be (8 ∗ .85) + (1 ∗ .25) + (1 ∗ .01) = .68 + .025 + .001 = 71%. The take rates for
these features are assumed to grow linearly from the date of their first production,
street-legal application.
Figure 2-4 shows the distribution in mass and take rates for the ten features
with the highest impact in 2010. Relatively only two features (air conditioning and
automatic transmission) have an impact of more than 5kg per car in 2010, but a large
number of features with varying masses and take rates produce a fleetwide impact of
2-3kg per passenger car.
The cumulative additional mass from comfort and convenience features from 1975 -
2010 is shown in Figure 2-5. This analysis suggests that in 2010, the average passenger
car was equipped with 136.1 kg in features. Project growth in the take rates of these
features suggest that the mass impact would grow to 232.2 kg in 2050. This value
includes only project growth in features already known and does not attempt to
predict the impact of new features not yet available.

33
Figure 2-4: Mass and take rate of ten features with highest fleetwide impact. Lines
of isometric fleet impact at levels of 20, 10, 5 and 3 kg/car are also shown.

2.6 Secondary Mass

Secondary mass is defined as the mass of additional components and structure nec-
essary to support a primary component or system and maintain vehicle performance.
For example, the primary mass of a sunroof system is contained in the sunroof cas-
sette, drains, wiring harness and electronic modules necessary to operate it. Sec-
ondary mass for a sunroof system would be additional structural reinforcement, in-
sulation and marginal engine and brake component growth necessary to offset the
increased vehicle mass. Secondary mass is usually considered in the context of pos-
sible mass reductions but the reverse effect–secondary mass accumulation–is what is
considered here.
A literature review by the Materials Systems Laboratory (MSL) at MIT of re-
cent studies of secondary mass is shown in Appendix B. The Mass Decompounding
Coefficient cited in the table refers to the percentage of primary mass attributable

34
Figure 2-5: Estimated mass impact of optional features by year. Dashed lines show
projected future impact using fits of existing take rate data, while solid line is a linear
fit of historic data.

to secondary mass. For example, a mass decompounding coefficient of 50% would

indicate that the study found for each kilogram of primary mass, 0.5 kg of secondary
mass was identified.
As reported in the literature the coefficients cited in the table vary widely from
23-129%. A strict average of these studies results in a mass decompounding coefficient
of 79.6%. Thus for the purposes of this analysis a secondary mass coefficient of 80%
has been used.
Recent studies, such as those of the Materials Systems Lab at MIT, have narrowed
focus to identify whether power train components are held constant or not. Where
power train components are allowed to change, the mass decompounding coefficient
is generally higher. However, the ability to incorporate power train changes into a ve-
hicle design is typically a function of timescale: year-to-year power train components
usually remain constant but during a vehicle refresh or facelift such changes would

35
be possible. Since this thesis identifies changes over decades on a fleet-average basis,
such changes are abstracted out.

2.7 Conclusions
Figure 2-6 highlights the impact that vehicle features have had since 1975, incorpo-
rating a secondary mass value of 80% applied to each feature category.

Figure 2-6: Mass of passenger cars 1975-2010 and weight attributed to Safety, Emis-
sions and Comfort/Convenience features (Secondary mass included).

Required safety and emissions equipment are the source of approximately 62 kg

(3.9%) and 24.6 kg (1.5%), respectively to a 2010 vehicle. On average, optional equip-
ment added to new 2010 vehicles added more than 136 kg (8.6%) to 2010 vehicles. By
comparison, 1975 vehicles on average incorporated 31.2 kg (1.7%) of safety equipment,
6.35 kg (0.3%) emissions equipment, and 71 kg (3.9%) of optional features.

36
 According to this analysis, basic sales-weighted passenger car mass (average pas-
senger car mass less mass attributed to features) was 1183 kg (2603 lbs) in 1982. This
base mass actually continued to drop until a low of 1154 kg was reached in 1987. Ba-
sic vehicle mass has remained relatively constant, reaching a value of 1190 kg in 2010.
Virtually all of the 200 kg in growth in vehicle mass since the mid-1980s is attributable
to increasing adoption of vehicle safety, emissions, and comfort/convenience features.

37
38
Chapter 3

Measuring Technological
Sophistication

Chapter 3
Chapter 3 reviews existing methods used to estimate vehicle techno-
logical sophistication. This chapter further uses the estimates of mass
devoted to features to inform these estimates.

This chapter discusses a variety of measures of technical sophistication and their

relatively consistent growth over time. Brake Mean Effective Pressure (BMEP) is
a measure of technical efficiency, while the Performance-Size-Fuel Economy Index
(PSFI), developed by An & DeCicco (2007), is a measure of the ability to deliver
measurable amenities to consumers.
Neither is a perfect measure: technical efficiency is a relatively foreign concept to
most car buyers and improvements in thermal efficiency can be used to either generate
greater horsepower or improve fuel economy. On the other hand, PSFI is unable to
capture all attributes that are important consumers and therefore underestimates
actual growth.
This chapter proposes incorporating vehicle features, using mass values identified
in Chapter 2, into the PSFI metric. The resulting metric, adjusted PSFI, thereby
more closely reflects customer-perceivable amenities.

39
3.1 BMEP
A common measure of internal combustion engine efficiency is Brake Mean Effective
Pressure (BMEP). As discussed in [Chon and Heywood, 2000] and [Heywood and Welling, 2009],
BMEP is a measure of the brake work per cycle an engine produces divided by the
engine’s displacement:

4π ∗ Tmax
BM EPmax =
Vd

or can be expressed as a function of volumetric efficiency (ηV ), fuel conversion

efficiency (ηf,i ) and mechanical efficiency (ηm ) multiplied by the fuel to air mass ratio
F
( ), intake/atmospheric air density (ρa,i ) and fuel heating value (%HV ):
A

F
BM EP = ηV ηf,i ηm ρa,i %HV
A

In Figure 3-1 Heywood and Welling show that the BMEP of engines increased
at a relatively constant rate of approximately 1.5% per year from 2000 - 2008, with
slight variations between naturally aspirated and forced-induction engines. These
measurements are not sales-weighted, but are derived from the engines available each
year.
These results suggest that engines are becoming more efficient at a relatively
constant rate. However, since BMEP is strictly a measure of engine efficiency it is
silent on the issue of how the entire vehicle system might have developed due to
advances in manufacturing technology, materials and design.

3.2 PSFI
The BMEP growth rate of approximately 1.5% per year shown by Heywood and
Welling correlates well with the growth rate in ton*mpg, a unit used by the EPA to de-
scribe vehicle efficiency. An and DeCicco (2007), using data from [Heavenrich, 2006],
calculate that ton*mpg increased an average of 1.6% per year from 1975 - 2005.

40
Figure 3-1: Trends in BMEP for U.S. engines from 2000 - 2008
[Heywood and Welling, 2009]

However, recognizing that ton*mpg does not incorporate growth in amenities such
as size and performance, An and DeDicco (2007) introduce the concept of the Perfor-
mance Size Fuel Economy Index (PFSI) as a measure of whole-vehicle sophistication
based on readily measurable amenities.1 For passenger cars PSFI is defined as:

hp
P SF I = P ∗ S ∗ F = ∗ F T 3 ∗ MP G
lb

In words An and DeCicco describe this measure as ”the ratio of moving a spatial
carrying capacity a unit distance with a given performance capability per unit of fuel
consumed.” As shown in Figure 3-2, An and DeCicco, incorporating growth in all
1
An and DeCicco use the term amenities to describe desirable characteristics tracked by the
EPA. This term should not be confused with the term feature used in this document to describe the
addition of physical equipment to perform specific functions.

41
primary vehicle attributes tracked by the EPA, find that PSFI has grown at a rate
of 5.3% for passenger cars from 1977 - 2005.

Figure 3-2: Performance Size Fuel Economy Index, 1975 - 2005 from
[An and DeCicco, 2007]. This metric of vehicle sophistication has grown with re-
markable linearity since 1975.

3.3 Adjusted PSFI

By incorporating three commonly measured attributes, PSFI is clearly a step closer to

accurately representing a measure of the technical sophistication of modern vehicles.
However, a critical omission of PSFI is that it is unable to capture developments in
feature content that result in improved vehicle safety, reduced emissions and com-
fort/convenience.

42
 As a quick illustrative example, take the values used for PSFI in 2005: 3463 lbs,
182 hp, 111 cu ft and 29.5 mpg. Assume that several years later car companies have,
on average, added 346 lbs (10% of weight) of equipment to improve safety, emissions,
and comfort/convenience of the vehicle fleet. Further assume that powertrains have
improved sufficiently to increase horsepower 10% without any resulting decrease in
fuel economy. The resulting fleet of vehicles–with an average weight of 3809 lbs and
200 horsepower but with identical interior volume and fuel economy–would clearly be
more sophisticated than the vehicles that preceded them but the calculated average
PSFI would be exactly the same.
To ameliorate such a shortcoming of PSFI, this paper proposes an Adjusted PSFI
metric. Adjusted PSFI simply adds a fourth term to the PSFI that represents a ratio
of vehicle mass with safety, emissions and comfort convenience features over the same
vehicle without these features. The resultant P SF IAdjusted metric is as follows:

M assw/f eatures hp lb
P SF IAdjusted = P ∗ S ∗ F ∗ = ∗ F T 3 ∗ MP G ∗
M assw/oF eatures lb lb

Since this fourth term is unitless, the P SF IAdjusted metric has the same units as
the standard PSFI metric. However, as shown in Figure 3-3, using the new metric
we can see that technical sophistication has grown at a faster rate than that shown
by PSFI.

3.4 Conclusions and Tradeoffs

These analyses indicate that the average technological sophistication of vehicles, mea-
sured by thermal efficiency of available engines, has grown consistently by a rate of
approximately 1.5% per year since the 1980s. When determined by ability to deliver
customer-perceivable amenities the growth is shown to be approximately 5.3% per
year since 1977.
Both of these indexes are measures of average vehicle performance–BMEP refer-
ring to non-sales weighted availability and PSFI referring to sales weighted delivery

43
 Figure 3-3: Traditional PSFI vs. Adjusted PSFI growth from 1977 - 2010.

of measurable vehicle amenities. However, both of these metrics are silent on the
question of how different vehicle amenities can be traded off against one another.
[Cheah, 2010] uses a combination of techniques summarized in Table 3.1 to deter-
mine a ”10-7” rule: that each ten percent in weight reduction results in a approxi-
mately a 7% reduction in fuel consumption for gasoline-powered midsize cars.
Alternatively [Knittel, 2009] uses a Cobb-Douglas model to determine that a ten
percent decrease in weight would correlate to a 4.26% increase in fuel economy, or
that a ten percent decrease in horsepower would result in a 2.57% increase in fuel
economy.
Cheah, Knittel and MacKenzie have focused on tradeoffs among the conventional
amenities size, performance and fuel economy. Literature to-date has not explored to
what extent safety, emissions, and comfort/convenience features could be traded off
against conventional attributes. A tradeoff analysis of features should extent beyond

44
Table 3.1: Fuel consumption (FC)-curb weight relationship for a current conventional
gasoline midsize car, from [Cheah, 2010].
Approach FC reduction per 10% FC reduction per 100 kg
mass reduction mass reduction
Literature Review 5.6-8.2% 0.36-0.58 L/100 km
Empirical data (MY2006- 5.6% 0.36 L/100 km
2008)
Engineering Simulation 6.9% 0.39 L/100 km
(ADVISOR)

the exploration of vehicle mass as features also impose parasitic losses that reduce
powertrain efficiency.

45
46
Chapter 4

Deployment Rates of Vehicle

Technology

Chapter 4
Chapter 4 reviews literature on the technology adoption and, using
data on deployment of features, identifies typical rates of deployment
of specific features in the automotive fleet.

Each of the metrics discussed in the previous chapter: BMEP, ton*mpg, PSFI and
P SF IAdjusted suggest that the growth in vehicle capability is strictly an evolutionary
process resulting from marginal gains in performance in multiple areas. However,
such marginal improvement in vehicle functionality can be better characterized as
the product of the independent development of hundreds of individual technologies.
The deployment of technology is aptly characterized by a logistic curve (also S-
Curve). One of the earliest uses of this functional form was used by Everett Rogers in
describing the process of Diffusion of Innovations. An image of this process is shown
1
in 4-1.
Per [Rogers, 2003] consumer adoption has depended primarily on five key factors:
1
In this chapter the terms diffusion and deployment are used interchangeably. These terms both
describe the process by which innovations are brought to market, although the traditional term
diffusion characterizes the process as consumer-driven, while deployment has connotations of being
producer-driven.

47
Figure 4-1: Functional form of the diffusion of innovations as envisioned by
[Rogers, 2003]

Relative Advantage People will purchase an invention if they believe it will en-
hance their utility in some fashion.

Complexity The degree to which an innovation is perceived as relatively difficult to

understand and use.

Compatibility The degree to which an innovation is perceived as consistent with

the existing values, past experiences, and needs of potential adopters.

Trialability The degree to which an innovation can be experimented with on a

limited basis.

Observability The degree to which the results of an innovation are visible to others.

48
4.1 Common Models of Diffusion/Deployment

A wealth of literature has been published on the diffusion of innovations in various

markets. One of the most commonly-cited of these, the Bass diffusion model, describes
process of adoption as primarily the function of two effects: advertising and word-of-
mouth.[Bass, 1969] Figure 4-2 shows a representation of this process using a System
Dynamics model. The counterbalancing feedback loops of market saturation and
word of mouth communication create the familiar S-shaped curve seen in diffusion
models.

Figure 4-2: Functional form of the Bass diffusion model as envisioned by

[Sterman, 2000]

4.2 Applying Regression

This chapter provides characterization of historical deployment rates of vehicle fea-

tures intended to inform plausible deployment rates for future automotive technology.
For the purposes of this analysis, a least-squares regression with a logistic (S-

49
Curve) form has been applied to feature take rates gathered from the EPA Fuel
Economy Trends report, Ward’s Factory Installed data, and DOT report HS 809 834.
The functional form of the regression used in these analysis is as follows:

Limit
T akeRate(t) =
1 + αe−βt

Where,
Limit = Maximum Take of Features
t = time in years
α = regression parameter approximating lag
β = regression parameter approximating steepness

A complete list of features with charts of take rate data and applied regressions
is provided in Appendices C - F.

4.3 Prior Work

[DeCicco, 2010] applies regression with a logistic form to feature data available from
EPA for front-wheel drive, fuel injection, multivalve engines and VVT. The analy-
sis proposes a logistic function and discusses both the steepness parameter of the
adoption curve and also the number of years since the ”first significant use” although
it is unclear exactly what criteria have been used to establish this date. DeCicco
also proposes a logistic function within the range of other powertrain technologies
as a plausible deployment scenario for hybrid electric vehicles (HEVs) although the
author notes that HEVs will compete with other technologies for incorporation into
future vehicle fleets.
While [Kramer and Haigh, 2009] do not use the term ”logistic function,” the au-
thors use descriptive language to identify phases of growth in the power generation
sector that are remarkably similar to those seen in the automotive industry. Kramer
and Haigh also discuss the importance of the developmental phase of new technol-

50
ogy, citing that it takes ”time and industrial capacity”–not just capital investment–to
deploy new technology.
[Nakicenovic, 1986] discusses the logistic form of the diffusion of technology in a
variety of fields and identifies several examples of diffusion of automotive features.
Nakicenovic also discusses differences among different types of vehicle features, a
concept continued here with the differentation among safety, powertrain, and com-
fort/convenience features. Nakicenovic cites examples of the time to reach 50% pen-
etration of a new technology, a parameter referred to later in this chapter as ”devel-
opmental lag time.”
The work performed in this chapter updates the work performed by Nakicenovic
with nearly 25 years of new data on a broader array of technologies. The larger
quantity of data available also allows secondary regressions of fit characteristics.

4.4 Important Characteristics and Secondary Re-

gression

In the form used here logistic regressions incorporate three primary characteristics:
maximum take rate (or application), maximum growth rate (slope of the curve at its
inflection point) and lag time, or delay in reaching the period of maximum growth as
shown in Figure 4-3.

4.4.1 Maximum Take Rate

Unlike stand-alone products where a potential market needs to be identified, the sales
of a particular feature of a passenger car cannot exceed sales of the car itself. As a
result, the potential market is defined as the percentage of the new car fleet equipped
with a given feature.
When examining historical maximum take rates, one soon encounters a problem:
for most features where data is available, take rates have either already approached
100% or are projected to reach this saturation point. The reason for this is simple:

51
 Figure 4-3: Important characteristics of a logistic function.

data is not closely tracked for technologies that are unsuccessful! At first blush this
would seem to lead to a problem of selection bias, but a closer look is appropriate.

Examples of actual failed attempts to bring features to market in the automotive

industry are not found in this data set.2 When take rates for a given feature do fall it
is generally the result of a technology supersession: eight-track players are replaced
by cassette decks and CD players; traction control is replaced by stability control.

The question, then, seems to be one of boundary definition. Broadly defined, fuel
injection has now saturated 100% of the new passenger car market. However, the
growth in the use of fuel injection was in fact characterized by the the successive up-
take of mechanical fuel injection, throttle-body fuel injection, and port fuel injection
as shown in Figure 4-4. Since the development of this figure Gasoline Direct Injection

2
One potential exception to this conclusion are diesels in the US market. After reaching a peak
of 6% of new passenger cars in 1981, diesels have fallen to less than a percent of recent new car sales.
However, the argument can also be made that diesels were simply replaced by more preferable SI
technology, thereby falling into the pattern of technology supersession discussed here.

52
(GDI) has seen large gains in use and may soon surpass other existing types of fuel
injection.

Figure 4-4: Technology supersession in fuel injection and engine management as

envisioned by [Amey, 1995]

Technology supersession is not always entirely contained within the automotive

domain. Two recent examples–mobile phones and navigation systems–also show ev-
idence of technology supersession. In-car phones originally reached a high of 9% of
new passenger cars sold in 2000, but in the past decade the paradigm of in-car elec-
tronics has been replaced by a model in which owners are presumed to buy a separate
device. Take rates of in-car navigation systems are still increasing, but an increasing
number of buyers use a stand alone portable navigation system or cell phone-based
navigation application. Time will tell whether a market will continue to exist for
factory-installed navigation systems with large screens and vehicle integration.
With this concept in mind, three common factors arise as indicators that a tech-

53
nology will not reach 100% take rate:

Limited Appeal (Example: Rear sear entertainment) Some features simply will
not appeal to all buyers. Rear Seat DVD players, for example, are generally
considered desirable by buyers with small children but are of very limited use
to a large fraction of buyers.

Significant Tradeoffs (Examples: Automatic Transmission, Front-Wheel Drive)

Some features or technologies inherently involve tradeoffs in other attributes.
Automatic transmissions, for example, remove the feel of control that some buy-
ers want. Front-wheel drive, while offering a benefit to fuel economy, requires
the sacrifice of rear-wheel drive handling dynamics sought after by some buyers.

Competing Technology / Paradigm (Example: On-board Navigation) Some fea-

tures and technologies compete with others for market share. On board nav-
igation, for example, competes with mobile phones and portable navigation
devices.

4.4.2 Maximum Growth Rate

The maximum rate at which the take rate of a technology grows is dependent on a
variety of factors: consumer demand, producers’ ability to bring the technology to
market on its fleet and, in some cases, the influence of regulation.
DeCicco (2010), while using slightly different regressions than those in this anal-
ysis, points to the maximum growth rate (UM ax ) of 17%/year for fuel injection in
LDVs, 11%/year for front wheel drive in passenger cars. While DeCicco does not
perform an explicit regression on other technologies, the proposed value of 7.6%/year
a plausible adoption rate of HEVs is similar to those he shows for multivalve engines
and VVT.3
3
Non-linear least squares regression requires the use of starting values to begin optimization.
Different starting values may result in the convergence on a variety of solutions, which may account
for different regression values in the literature

54
 However, as discussed in Chapter 2, each of these technologies has, to a varying
extent, been driven by regulation. Figure 4-5 examines a histogram of the maximum
growth rate of all features for which data is available. Figure 4-6 performs the same
analysis but divides technologies into the functional categories of safety, powertrain,
and comfort/convenience.

Figure 4-5: Histrogram of maximum feature growth rates.

Peak annual growth rates range from 1% - 23.9%. Safety features see the fastest
deployment among these categories, ranging from 4.5% per year to 23.9%. This max-
imum growth rate seems to confirm conventional wisdom, adopted in [NHTSA, 2010]
and others, that an average five-year product development cycle is appropriate for
modeling the automotive industry– even technologies with a clear life-saving benefit
cannot be deployed much faster than 20% of the new vehicle fleet per year.
This histogram also seems to confirm the sentiment expressed by DeCicco, that
the examples of technology growth in powertrain development commonly cited (FI,
FWD, VVT, and Multivalve engines) exhibit ”very rapid rates of change.” In the

55
 Figure 4-6: Histogram of maximum feature growth rate by category.

broader context of automotive features, yearly growth rates of less than 5% per year
are more common.
The fastest growing of the comfort/convenience features, Satellite Radio, merits
additional consideration. At 11% per year, Satellite radio is currently exhibiting very
rapid growth in application for a feature where regulatory impact is not a consider-
ation. Satellite radio has been the subject of a very aggressive marketing campaign
in recent years. Since satellite radio is a subscription-based service, it represents a
stream of residual revenue from the vehicle purchaser, typical shared between the
satellite radio provider and vehicle manufacturer. As a result, OEMs have an incen-
tive to push satellite radio to customers in the hopes that many of them will continue
to subscribe. However, since the Wards data used in these regressions shows only the
purchase of an option it is unknown how many of these customers actually activate
the satellite radio service or are even aware of its installation.4
4
Dealers may also receive a bonus for successfully selling options such as satellite radio. This
compensation structure results in a push from every stakeholder to install satellite radio in a vehicle.

56
4.4.3 Developmental Lag Time

When a new technology or feature is brought to market in the automotive industry

it is typically done in limited quantities and often on high-end flagship products.
Fuel injection, for example, was first brought to market by Mercedes in 1954-55 on
its 300SL race homologation vehicle. It was not until 1985, three decades later, that
maximum growth in the mainstream automotive market was achieved.
This period of development and maturation seems to be not only the product of
limited consumer demand, but also of an automotive manufacturing infrastructure
building competence and confidence in a new concept.
The analysis shown here is a secondary regression comparing the inflection point
from primary regressions to the date of the first production, street-going vehicle to
use a technology. Since many features referenced here were developed before the
earliest Wards data sets used, start dates have been identified using a variety of
sources including Wikipedia, patent databases, and company advertising. As such,
small discrepancies may exist in some start dates. A complete table of the values
used for the charts shown below is available in Appendix G.
Figure 4-7 shows a regression of all features simultaneously, while Figure 4-8 shows
independent secondary regressions by feature category. This analysis shows a dra-
matic, exponential decline in the developmental lag time of features deployed over
the past century.
There are a variety of explanations for such a change in the automotive industry.
It is theoretically possible that the marked decrease in developmental lag time of
features is the product of more stringent consumer expectations resulting from more
exposure to new products and features through new media, and a higher level of
communication between consumers leading to greater Word of Mouth interaction
between adopters and potential adopters.
However, improvements in supply side capabilities have likely played a strong role
as well. [Clark and Fujimoto, 1991] and [Ellison et al., 1995] highlight that while U.S.

Many manufacturers also provide the service free of charge for a period of time, a way of increasing
the trialability of the feature for potential customers.

57
 Figure 4-7: Historical phase-in time of all features.

and European automakers were at a significant disadvantage to Japanese producers

in the 1970s and 1980s (Figure 4-9). However, by the mid-1990s they had reduced
overall product lead time by nearly a year. The resultant increase in product changes
allows a manufacturer to incorporate new features into the product mix more readily,
while increasing competition pushes manufacturers to differentiate products by the
incorporation of features with consumer appeal.

The structure of the automotive industry itself has also changed significantly over
this same time period. Automakers are no longer vertically-integrated giants capable
of receiving raw materials in one end of a plant and shipping cars out the other–
the industry is now highly stratified and automakers are dependent on a pyramid
of suppliers for components. [Ellison et al., 1995] highlight the increased role that
suppliers play in the product development process. For U.S. suppliers, Ellison et
al. (1995) find that supplier content in the product development process more than
doubled from 15% to 33% from the 1980s to the 1990s. While Japanese supplier

58
 Figure 4-8: Historical decrease in phase-in time of features by category.

content decreased slightly over the same period, European supplier content increased.
Increasing reliance on suppliers suggests that intellectual property is distributed more
quickly as suppliers are free to market a new technology to a variety of manufacturers
as customers.
However, despite these factors that dramatically different competitive landscape
and decline, current developmental lag time still for new vehicle features remains on
the order of approximately a decade, although this represents a dramatic decline since
the 1950s.

4.5 Alternative Methodology: Z-Curves and Lo-

gistic Approximation
[Schafer et al., 2006] and [Bandivadekar, 2008] describes deployment of new automo-
tive technology in three primary phases. The authors then project potential time

59
Figure 4-9: Product variety and model changes by region, adapter from
[Clark and Fujimoto, 1991].

periods for these phases for a variety of technologies (shown in Figure 4-10). This
section approximates the three phases used by Bandivadekar and Schafer et al. by
extrapolating parameters identified by regression with a logistic form performed in
the previous sections.

Table 4.1: Stages of deployment using parameters extracted from logistic regressions.
Vehicle Technology
Implementation Variable Anti-lock Keyless Satellite
Stage Valve Timing brakes Entry Radio
Market Competitive 17 15 9 4
Vehicle
Penetration across 15 25 12 8
new vehicle produc-
tion
Major fleet penetra- 10+ 10+ 10+ 10+
tion
Total time required 42+ 50+ 31+ 22+

In this analysis the periods of time for each phase of technology deployment are
defined in the following manner. The ”Penetration across new vehicle production”
phase is defined as the maximum fit take rate from logistic regression divided by the
maximum growth rate from the same regression. Conceptually this can be envisioned

60
Figure 4-10: Stages of deployment of new vehicle technology. Image from
[Bandivadekar, 2008], adapter from [Schafer et al., 2006]

as the line of maximum slope being extended until it intersects the x-axis and the
maximum take rate (generally 100%).

The ”Market Competitive Vehicle” phase is defined here as time between the first
production application and the x-axis intercept of maximum growth, or: DateM axGrowth −
DateIntroduction − (GrowthRateM ax /2 ∗ T akeRateM ax ).

The ”Major Fleet Penetration” phase is not addressed here as it is generally the
product of fleet turnover rates. The ”10+ years” value is used in all cases.

Table 4.1 shows the results of applying such parameters to four examples of tech-
nologies with varying function, complexity and timing. This analysis confirms that
for a variety of historical examples the time periods Schafer identifies are plausible.

Z curve fit parameters for all relevant time series feature data is plotted in Ap-
pendix G. The last column, ratio, represents a ratio of the time spent in the devel-
opment phase over the time spent in the deployment phase. Values greater than 1
indicate that more time is spent in development than deployment, while values less
than one indicate that deployment takes longer than development.

61
Table 4.2: Z-curve: Phases of deployment and actual percentages of new vehicles sold
with each technology for the four examples used in the previous table.

4.6 Differences Among Manufacturers

The analyses performed in the previous sections show trends in the new passenger
car fleet as a whole. However, the automotive market is composed of a range of
independent companies from large, full-line producers to boutique operations or those
that compete only in a few segments. A fleet-wide analysis does not show differences
between such varied producers.

4.6.1 Manufactuer-Specific Action and Fleet Impact

Figures 4-11, 4-12, and 4-13 show sales-weighted deployment of three key power-
train technologies in the U.S. passenger car fleet overlaid with the start and finish
of implementation by ten major automotive manufacturers based on data drawn pri-
marily from the EPA fuel economy trends database. The manufacturers include three
U.S., three European, and four Japanese auto manufacturers.

62
 The manufacturer-specific bars start with the release of the first model includ-
ing the technology and end with the phase-out of the last model not including the
technology. The bars thereby represent the transitional period for each manufacturer
with respect to the technology. In each case the bars have been ordered by the first
use of the technology, with the manufacturers at the lower end of the chart beginning
earliest.
Several key factors surface upon examination of technology deployment on a
manufacturer-by-manufacturer level.5

Figure 4-11: Differences among automotive manufacturers in phase-in of fuel injec-

tion.

Individual Producers Are Faster Than the Market While fleet-wide deployment
may take decades, individual manufacturers are capable of deploying technol-
ogy much faster. Mazda deployed Fuel Injection, VVT and Multivalve engines
across its entire product portfolio in three, four, and eight years respectively.
This is a special case as Mazda has traditionally been an OEM with a limited
5
Small distinctions arise in this analysis. Chrysler began experimenting with fuel injection around
the same time as GM and in fact sold a small number of vehicles equipped with the Bendix ”elec-
trojector” system. However, because the system was installed post-production and was purportedly
installed on fewer than one hundred vehicles it is not counted as a start date here.

63
Figure 4-12: Differences among automotive manufacturers in phase-in of multivalve
engines.

product portfolio, but other manufacturers have also shown the ability to move
quickly when necessary.

First mover advantage? An automotive producer that is first to market with a

new technology is not necessarily the first to roll out the technology to its entire
fleet. GM, second only to Mercedes in launching fuel injection, was among the
last to discontinue use of carburetors while Mazda, despite a late start, was
able to completely deploy fuel injection across its vehicle portfolio in only three
years. These small-scale deployments are, however, needed to build experience
with a new technology necessary for a larger rollout.

Fleet Impact One or two manufacturers experimenting with a new technology will
not make a meaningful impact in fleet-wide numbers. Fuel injection had been
in production on street-going vehicles for more than two decades before total
fleet penetration exceeded 10%. Similarly, Variable Valve Timing only exceeded
10% fleet-wide deployment when six major manufacturers had begun to use the
technology on at least one vehicle.

64
Figure 4-13: Differences among automotive manufacturers in phase-in of variable
valve timing.

4.6.2 Strategies Within Manufacturers

Examining one step closer (at the individual model level) we can see that companies
sometimes find it advantageous to buy new technology from a competitor to gain
experience quickly. Ford’s first multivalve engine was found in the Taurus SHO, a
high-performance sedan with an engine built by Yamaha. Chrysler’s first multivalve
engine was in variants of the Dodge Colt, a product of the Diamond Star Motors
(DSM) joint venture with Mitsubishi.
A trend that appears nearly universal is that automakers generally launch new
technology on high-end luxury or performance products first: Toyota’s first fuel in-
jected product was the Celica Supra; Honda’s first car with VVT was the Acura NSX;
Nissan’s first car with VVT was the Infiniti Q45 and Toyota’s first car with Direct
Injection was the Lexus IS F. Many of these technologies descend from racing vehicles
and this seems to be the literal embodiment of the old adage ”race on Sunday, sell on
Monday.” A number of examples of this are discussed in the following paragraphs.
Through 1978, Toyota’s entire fleet was carbureted. In 1979, Toyota equipped the
Celica Supra, a performance car, with fuel injection. In 1979 Toyota added FI to the

65
Cressida, its luxury car. These two models continued until 1983, when the Starlet,
the Camry and certain Celica models with FI. Toyota continued to migrate its vehicle
portfolio over to FI through the subsequent years; the majority of Corolla and Tercel
models migrated to FI in 1990, although a few variants of the Tercel and Corolla were
not changed until 1991, at which point Toyota’s fleet was entirely fuel injected. The
total time of transition from carburetion to electronic fuel injection was eleven years.
While the data in the Trends Report is not sales weighted, the transition began with
more expensive, relatively low volume vehicles and finished with the inexpensive cars
in Toyota’s portfolio.
BMW’s first cars with four valves per cylinder were released in 1988. These
models are listed in the Trends Report database as the 3, 6, and 7 Series, although
the horsepower values in the database indicate that these were limited production
variants using the Motorsport-developed high-output engines. Only these models
were produced with four valves per cylinder for three years. BMW rolled out multi-
valve engines across much of its product portfolio very quickly from 1991-1994 and by
1995 only BMW’s V12 engine used two valves, but production of this engine continued
through 2001. From 2002 - On all BMW’s vehicles have used multivalve engines. The
total time of transition from two-valve to multi-valve engines was thirteen years,
although the majority of the change in the portfolio took place during four years.
The transition began with low-volume, high-performance vehicles.

4.7 Supply Side Constraints on Deployment

[Bandivadekar, 2008] calls attention to the importance of supply side constraints in
determining the speed with which new technology can be brought in to the new
automotive fleet. The four factors listed by Bandivadekar are paraphrased below.

Development lead times and availability across product platforms Time needed
for development and integration of components in a vehicle platform.

Capital investment required Time and capital required to retool a production

66
 facility.

Supply of critical systems/components Possible shortages in production of com-

ponents for new tecnhnologies.

Capacity utilization Need to balance shifting product demand with maintaining

production volumes.

Table 4.3: Supply side constraints by product development role.

Concept De- Integration, Production Market Market Sup-
velopment Testing Adoption port

(Researchers, (Automotive (OEM, Sup- (Consumers, (OEMs, Sup-

Suppliers) OEM) pliers) Regulators) pliers, Dealers)
IP Boundaries Engineering Retooling Attribute Warranty, Re-
Expertise Capital Desirability call costs
Design Capital Platform Redesign Price Repairability
Demands schedule
Component Durability Material Regulatory Dealer Training
Costs Testing Costs Compliance
Safety Testing Capacity
Constraints

Table 4.3 expands upon the constraints identified by Bandivadekar and classifies
them by the role and responsible party with which the task is most closely associated.6
The roles, as described here, reflect the automotive product development environment
as it exists today; thirty years ago these constraints were more wholly within the
domain of the automotive OEM.

4.8 Impact of Regulation

As discussed in Chapter 2, the impact of regulations on automotive technologies and
features is often unclear. Adoption of applicable technology is often well under way
before a regulation requiring it is made law, and many features offer marketable ben-
efits other than mere regulatory compliance. A list of such features where regulation
6
While this table is situated in a discussion of supply constraints, factors within the fourth column
are market-driven and not strictly within the domain of the supply side.

67
has impacted deployment with relevant legislation and dates is shown in Table 4.4,
and a chart of maximum growth rates for regulated vs. unregulated features is shown
in Figure 4-14.

Figure 4-14: Differences in maximum growth rates between regulated and non-
regulated features.

Where possible the ”Date Issued” is in Table 4.4 is the year of governing legislation
or the Notice of Proposed Rulemaking for the relevant standard, identified by the date
added to the Federal Register, with additional clarification from [DOT, 2004a] and
[DOT, 2004b]. This date is generally the first public notice of an agency’s intent to
issue a rule.
On its face this date is a trigger for an automaker to begin preparing for compliance
but these dates are imprecise for two reasons: first, automakers may anticipate that
a rule will be delayed or abandoned during the Notice and Comment process and
postpone compliance activity until a final rule. Alternatively, automakers may become
aware of pending regulations during informal communication with regulatory agencies
and begin compliance activity in advance of an official NPRM.
If deployment of a feature is high prior to regulation, a governmental standard
will only force a feature to be applied to laggard vehicles in the fleet at a minor cost.

68
Table 4.4: Regulations applicable to features analyzed in this chapter. Standards
with an asterisk are flexible standards that could be achieved via this feature or other
means; regulations without an asterisk are strict requirements.
Feature Applicable Regulation Date Issued Effective Date New Passenger
cars equipped
at Issue Date
Intermittent FMVSS 104 1966 1968 unknown
Wiper/Washers
Rear window defogger FMVSS 103 1966 1968 unknown
Front Wheel Drive CAFE* 1975 1978 - Present 7%
Fuel Injection Clean Air Act Amend- 1970 - 1990 1975 - 1994 5%(1970)
ments*
CAFE* 1975 1978 - Present
Multivalve Clean Air Act Tier II* 1999 2004 - 2009 65%
VVT Clean Air Act Tier II* 1999 2004 - 2009 17%
Dual Master Cylinders FMVSS 105 1966 1968 ˜69%
Front Disc Brakes FMVSS 105a* 1970 1976 41%
Power Brakes FMVSS 105* 1975 1976 ˜74%
Anti-Lock Brakes FMVSS 126* 2005 2009 - 2012 63%
Driver Front Airbag FMVSS 208I* 1984 (final) 1986 - 1995 <2%
Dual Front Airbags FMVSS 208 1993 (final) 1996 - 1997 14%
Side Impact Beams FMVSS 214 1968 1973 <17%
Curtain Airbags FMVSS 208 2011 2013-2018 >91%
Traction / Stability FMVSS 126 2005 2009 - 2012 14%
Control
Alarm / Immobilizer Motor Vehicle Theft 1984 1984 <9%
Prevention Act*

Alternatively, if deployment of a feature is low prior to regulation, a governmental

standard requires the entire automotive industry to adapt its fleet in compliance at
a potentially high cost. As shown in Table 4.4, deployment rates prior to regulation
have ranged from less than two percent to more than 90% of the passenger car fleet.

Figure 4-15 plots the maximum yearly growth rate of deployment on new passenger
cars vs. the percent of vehicles equipped with a feature at the time the relevant
regulation was announced. No clear trend linking maximum growth rate to the level
of prior deployment can be shown. If regulatory agencies were engaging in aggressive
forcing of technology, one would expect that when a regulation is announced early in
the development process (when take rates are low) that maximum growth rates would
be higher than when a regulation occurs after most vehicles are already equipped
with a feature. Figure 4-15 shows that this is not necessarily the case, and that the
interaction between regulation and deployment is more complex.

69
Figure 4-15: Maximum growth rate vs. number of passenger cars equipped at time
regulation is announced.

A few brief examples of how legislation has interacted with deployment of vehicle
features are discussed in the following sections.

4.8.1 Frontal Air Bags

Driver frontal airbags, which were installed on less than 2% of passenger cars in 1984,
became a method of compliance with FMVSS 208 standards announced in 1984 (final
rule) requiring passive safety devices. However, FMVSS 208 also initially permitted
automatic seat belts to achieve compliance and it was not until 1993 that a final
rule requiring both passenger and driver side airbags on 100% of passenger cars was
issued.[DOT, 2004a] Impact of these regulations is shown in Table 4.5.

70
 Table 4.5: Adoption of airbags and timing of FMVSS 208 requirements

4.8.2 Anti-lock Brakes

Figure 4-16: Adoption of ABS and timing of FMVSS 126 NPRM.

In the case of Anti-lock Brakes(ABS), regulatory impact is clear despite widespread

adoption before regulation. Anti-lock brakes, first made available in the early 1970s,
were equipped on approximately 63% of the passenger car fleet in 2005. Figure 4-16
indicates that adoption of ABS was leveling off and predicted to stabilize at approx-
imately 65%.

71
 In 2005 NHTSA announced FMVSS 126, which required all vehicles to be equipped
with Electronic Stability Control (ESC) starting in 2012 (phase-in beginning in 2009).
While it is theoretically possible to implement ESC without the use of ABS, the
unusual double sigmoid characteristic of Figure 4-16 suggests that automakers are
treating FMVSS 126 as a de facto requirement for ABS. As a result, application of
ABS has again begun to increase and will reach 100% by 2012. Since this regulation
apparently alters the saturation point of ABS and not particularly the growth rate,
this effect is not captured by Figure 4-15.

4.8.3 Multiple valves and Variable Valve Timing

According to [EPA, 1999], multiple valves per cylinder and variable valve timing are
both technologies that were important in reducing emissions to the levels required by
EPA Tier II requirements. These requirements were announced in 1999 and phased
in from 2004 - 2009.

Table 4.6: Adoption of multiple valves per cylinder and variable valve timing in
comparison to the timing of Tier II requirements

By 1999, 65% of passenger cars were already equipped with multiple valves per
cylinder, but only 17% used variable valve timing. Some manufacturers, including
Mazda and Chrysler, had not marketed any vehicles with variable valve timing by
1999.
Some evidence of the accelerative effect of Tier II requirements on the implemen-
tation of variable valve timing can be seen during the phase in period: from 2004

72
- 2009 variable valve timing grew nearly ten percent per year with an actual peak
growth rate from 2008 - 2009 of 17%, far in excess of the 7% per year maximum
growth rate predicted by the fit logistic curve.

4.9 Conclusions

The objective of this chapter has been to characterize historical deployment rates
of technologies and use these characterizations to make useful predictions about the
plausible future deployment rates.
All features require significant developmental time before they can be deployed
in the vehicle fleet at significant rates. This developmental period has been reduced
drastically in the past sixty years, from more than 50 years of developmental time
prior to 1950 to approximately ten years for recently developed features.
Maximum annual growth in feature penetration of 6% per year or less is far more
common than growth rates of 10 - 24% per year. Features that improve vehicle
safety will generally be adopted in new cars faster than either powertrain or com-
fort/convenience features, which generally see maximum growth rates of under 4%
annually.
Future automotive features could be reasonably expected to follow a similar pat-
tern: small-scale deployment for approximately five years leading to exponential
growth and an inflection point ten or more years after first application.
This analysis also highlights the extent to which regulation has played a part in
technology deployment in the automotive industry. The majority of features with
peak growth rates in excess of 6% per year have been impacted by some form of
regulatory intervention– either emissions standards, as is the case for most powertrain
innovations, or NHTSA safety standards in the case of airbag and brake systems.
However, it is unclear whether regulation actually causes deployment to occur
more rapidly, or whether technologies for which regulation is enacted are also those
where market and supply-side factors lead to rapid growth in adoption rates. Side
and curtain airbags represent counterexamples to the idea of regulatory forcing, with

73
maximum growth rates of 9% and 12.5%, respectively, prior to a regulatory require-
ment.
This analysis does not explore performance metrics such as the Insurance Insti-
tute for Highway Safety (IIHS) or NHTSA New Car Assessment Program (NCAP)
crash tests. Vehicle manufacturers may add additional safety equipment to improve
performance in such tests that is not strictly required by regulation.

74
Chapter 5

Findings, Applications and

Conclusions

Chapter 5
Chapter 5 highlights conclusions that can be drawn from analysis of
vehicle mass and technology deployment trends and ways that these
findings could be applied to existing models. This chapter also notes
limitations to this analysis and potential extensions for further work.

This thesis has focused on two key themes: the role features have played in the
development of passenger cars, and characterization of the deployment of new features
and technology in the passenger car fleet. The conclusions from each of these areas
are discussed below.

5.1 Vehicle Mass Findings

The average mass of U.S. passenger cars dropped from 1845 kg in 1975 to a low
of 1378 kg in 1987. Since then, mass has climbed steadily, reaching 1591 kg in
2010. However, absent developments in safety, emissions and comfort/convenience
the average passenger car would have been approximately 1735 kg in 1975, a low of
1254 kg in 1987 and 1368 kg in 2010, a growth rate of just under 5 kg per year without

75
vehicle features.
Despite the incorporating of increasing levels of technology, vehicle mass continues
to grow, and this growth strongly parallels increased applications of vehicle features.
The analysis of optional and required features indicates that both the absolute mass
contribution and the percentage of passenger car mass devoted to features are increas-
ing, from a total of 109 kg (5.9%) in 1975 to approximately 223 kg (14%) in 2010
on a sales-weighted average basis. Optional features that enhance the comfort and
utility of passenger cars are the largest mass contributor at 136 kg in vehicle mass.
Extrapolation of current trends indicates that features would contribute a total of
386 kg to the average vehicle mass in 2050, suggesting that if other vehicle attributes
are unchanged from 2010, average vehicle mass could reach 1977 kg.

5.2 Feature Deployment Findings

The deployment of technology in the automotive industry is not simply the result
of consumer demands but also the product regulatory influence and infrastructure
constraints that prevent the rapid diffusion of innovation possible in other consumer
industries. The timescales for the development of a new feature, from first application
to maximum growth, are measured in decades. However, this developmental time
has decreased exponentially and modern features achieve maximum growth rates in
approximately ten years.
Features that enhance the safety of passenger cars exhibit more rapid maximum
growth, on average, than either powertrain or optional comfort / convenience features,
but growth in excess of 15% per year is rare and has only been possible for features
with a life-saving benefit. For powertrain features, maximum growth of 6 - 14% is
more typical. The majority comfort / convenience features do not exceed growth
rates of 6% annually, and in only one case have exceeded 8% per year.
The impact of regulations is difficult to ascertain. While features affected by
a regulatory requirements do generally have higher growth rates than unregulated
features, an expected correlation between early regulation and higher growth rates is

76
absent. Specific examples indicate that regulations affect maximum growth rate and
maximum take rate in differing ways.
Individual manufacturers may deploy technologies much more quickly than indus-
try average time scales, but penetration of a feature or technology in the new car fleet
is typically not significant until most major automakers are have begun deploying the
technology. Perhaps contrary to intuition, manufacturers that are first to market with
a specific feature may not be those that are able to apply fully to their fleet first.
High performance vehicles will continue to play a significant role as platforms
for small-scale deployments of technology that lead the mass market. The generally
higher purchase price of these vehicles allows application of technology that could not
be incorporated in lower-priced vehicles.

5.3 Applications for Policy and Vehicle Fleet Mod-

els

An immediate application of these feature and deployment guidelines is to inform

vehicle fleet models used to predict future GHG emissions and fuel consumption.
Many fleet models operate with a presumption of emissions and fuel consumption
reduction due to the increased application of technology, and use ”no change” or
Business As Usual (BAU) scenarios to describe a worst-case where current vehicle
characteristics and focus on performance continue unabated. Such models should
anticipate the continued growth in mass allocated to vehicle features. As a result,
either vehicle mass will continue to grow or increasingly aggressive weight-saving
techniques will be needed just to maintain current vehicle mass.
Other models, such as the Volpe Model [NHTSA, 2010] used to support CAFE
regulations, attempt to identify plausible deployment rates of technology to help set
fuel economy regulations at an economically optimal level. In this model NHTSA has
incorporated an assumption that vehicles are, on average, redesigned every five years.
In Chapter 4, few features exceed growth rates of 20% per year. This indicates that

77
NHTSA’s assumption of a five-year product redesign cycle is a reasonable one.
The Volpe Model uses a combination of ”phase-in caps” and manufacturer redesign
cycles to estimate reasonable limits on technology deployment and as a proxy for
supply-side constraints discussed in Chapter 4. Examples of these phase-in caps are
shown in Table 5.1. Due to the other constraints incorporated in the Volpe model,
actual application of these technologies applied by the model are lower, as shown in
Table 5.2.

Table 5.1: A sample of specific technology phase-in caps from current CAFE legisla-
tion, referenced from Table V-12 in [NHTSA, 2010]
Technology 2012 2013 2014 2015 2016
Engine Friction Reduction 100% 100% 100% 100% 100%
Stoichiometric Gasoline Direct Injection 85% 85% 85% 85% 85%
Combustion Restart 0% 0% 85% 85% 85%
Turbocharging and Downsizing 85% 85% 85% 85% 85%
Dual Clutch or Automated Manual Trans- 85% 100% 100% 100% 100%
mission
Electric Power Steering 85% 85% 85% 100% 100%
Power Split Hybrid 3% 6% 9% 12% 15%

Table 5.2: A sample of specific technology application levels output from the
Volpe model to support 2016 CAFE regulation, referenced from Table V-48 in
[NHTSA, 2010]
Technology 2012 2013 2014 2015 2016
Engine Friction Reduction 28% 50% 55% 58% 66%
Stoichiometric Gasoline Direct Injection 18% 26% 31% 35% 38%
Combustion Restart 0% 0% 2% 5% 11%
Turbocharging and Downsizing 13% 20% 21% 25% 27%
Dual Clutch or Automated Manual Trans- 24% 38% 48% 61% 69%
mission
Electric Power Steering 40% 54% 57% 59% 72%
Power Split Hybrid 5% 5% 6% 5% 5%

Phase-in caps in the 2016 CAFE rule are immediately set to 85 - 100% for a
majority of powertrain technologies. Hybrids, diesels, and a few advanced combustion
technologies are initially set lower. However, when constrained by other model factors
these rates fall. Most of the examples shown in Table 5.2 fall within the ranges of
maximum historical growth rates identified in Chapter 4 for powertrain features. Dual

78
Clutch Transmissions and Electric Power Steering, for example, exhibit a maximum
modeled growth rate of 14% per year. This rapid application of technology is on par
with the most rapid rates of non-safety technology seen in historical data analyzed in
this thesis.

For other features, technology readiness is an important consideration. The ten-

year developmental lag times identified in this analysis have been defined as the period
from the first production application of a technology, and experimental laboratory
development time has not been counted. While most of the powertrain technologies
contained in the 2016 FRIA are already in production, NHTSA incorporates a few
technologies (i.e. combustion restart) that have not yet been incorporated into a
production vehicle, and additional pre-production development time needed for such
technologies would be in additional to the development times discussed here.

Technical readiness is a clear concept when applied to a specific technology, i.e.

Gasoline Direct Injection. However, some technologies identified in 2016 CAFE doc-
umentation are not clearly defined. ”Engine Friction Reduction,” Aero Drag Reduc-
tion,” and ”Low Friction Lubrication” are not achieved through a single technology
or feature and the baseline friction or drag values will be different for each manufac-
turer. As a result, it is difficult to speculate about whether the underlying design
changes and technologies necessary to achieve these efficiency improvements are, in
fact, ready.

Previous CAFE regulations have generally set fuel economy targets only 5-6 years
in the future. However, in September, 2010 the U.S. DOT and EPA issued a Notice of
Intent to establish fuel economy and GHG emissions regulations from 2017 - 2025, and
an NPRM for these standards is expected in 2011. Such a rule will more than double
the horizon of regulation from the current six years. Accurate predictions of the pro-
duction readiness of powertrain technology will be critical to identifying optimal levels
of regulation. Such a task will test the limits of NHTSA’s technology adoption-based
modeling approach, and the very real possibility exists that a presently unknown
technology could be developed, commercialized and produced prior to end of the
regulated period.

79
5.4 Other Considerations

The deployment scenarios described in this thesis could reasonably be used to describe
a broad range of future vehicle features and technologies. However, they are examples
of relatively straightforward, standalone systems developed by automakers and their
suppliers. Automotive developments that will not necessarily follow the patterns
exemplified here are described below.

5.4.1 Alternative Fuels and Parallel Innovation

One of the most critical questions regarding future automotive technology is whether
such deployment rates could be used to describe future powertrain technologies. For
features of a similar level of complexity (advanced injection or induction systems, for
instance) these adoption rates are plausible. However, many powertrain technologies
viewed as central to reducing fuel consumption and GHG emissions are based on
alternative fuels.

Most manufacturers are now developing a range of alternative-fuel vehicles, and

virtually all require some degree of parallel infrastructure development: BEVs require
the development of a public charging infrastructure and major grid improvements; fuel
cell vehicles require a hydrogen refueling infrastructure, and even diesel passenger cars
in large numbers would require changes to the refining and refueling infrastructure.

The need for infrastructure development in parallel with automotive technology

development could significantly extend the deployment rates describe here for two
primary reasons. The parallel deployment of technology by at least two giant indus-
tries would require close alignment of incentives not typically seen. Additionally, the
need for infrastructure development implies that consumers will initially have lim-
ited access to such infrastructure, negatively impacting two of Rogers’ key adoption
factors: compatibility and trialability.

80
5.4.2 Disruptive Innovation

The concept of disruptive innovation, introduced by Bower and Christensen (1995),

illustrates the idea that a new paradigm or product will result in a simpler product
that will either define a new market segment or appeal to low-end customers and
gradually overtake existing producers until, one by one, they are unable to compete
and the market is revolutionized.
The features discussed in this thesis represent sustaining improvements, not dis-
ruptive ones. They are, therefore, silent on the issue of the results that a large,
pattern-breaking innovation in the automotive industry might achieve. Examples of
major disruptive innovations have been notably absent from the automotive market
and despite vast growth in capabilities, passenger cars and the infrastructure in which
they are sold, serviced and operated remain fundamentally similar to that of a century
ago.
[Christensen et al., 2002] set forth a brief list of tests that are helpful in identifying
common themes for disruptive innovations that will create a new market:

• Test #1: Does the innovation target customers who in the past haven’t been
able to ”do it themselves” for lack of money or skills?
• Test #2: Is the innovation aimed at customers who will welcome a simple
product?
• Test #3: Will the innovation help customers do more easily and effectively what
they are already trying to do?

and another set of tests for identifying if a technology is capable of disrupting a

market from the bottom up:

• Test #1: Are prevailing products more than good enough?

• Test #2: Can you create a different business model?

The beginnings of disruptive innovation may now be starting not with vehicles
themselves, but the vehicle ownership model. Just as automotive leasing has grown

81
to a major portion of the new passenger car market, new business models are forming
that further distance a vehicle operator from purchase of the vehicle.
Companies such as Better Place, founded in 2007, envision a model in which
battery packs and electric power are purchased as a service. Discharged battery packs
are exchanged for charged ones at service stations much as barbecue users exchange
empty propane tanks for full ones.
Car sharing services such as ZipCar take this concept a step further, allowing
subscribers to use vehicles themselves as a service. While this business model is
not inherently different from traditional car rental, careful placement of vehicles in
dense urban locations, hourly rentals and online reservations have made use of service
sufficiently convenient that some ZipCar members now use the service in lieu of private
cars.
Battery and car sharing exemplify the traits identified by Christensen as typical
of disruptive technologies. With sufficient users, car sharing services could alter ve-
hicle sales and usage dynamics, disrupting current sales models. Such services could
also fundamentally change demand for vehicle technology, perhaps pushing for faster
deployment of comfort/convenience features and safety features, or perhaps a large
number of cost-conscious car-sharing users could strengthen demand for fuel-saving
technology.
However, disruptive technologies are notoriously hard to predict in advance. Only
in retrospect may we be able to point to specific causes of innovation.

5.5 Extensions

As discussed in earlier chapters, required and optional features features add signifi-
cantly to the mass of a vehicle. However, some features impact vehicle efficiency in
other ways. Electrical features necessitate the use of a larger alternator and charging
system and impose higher parasitic losses. Vacuum and hydraulic assistance also sap
power from a vehicle’s engine. These effects are not captured in this thesis but are
potentially quantifiable.

82
 Chapter 4 of this thesis has focused exclusively on characterizing historical de-
ployment rates. However, this document does not attempt to identify the limits of
feasibility and with enough investment deployment rates could be pushed higher.
A more complete analysis would also characterize the marginal cost curve of faster
deployment of vehicle technology and features.

83
84
Appendix A

Tables

For features where take rate data is unavailable, 2010 take rates have been estimated
using availability data for top-10 selling passenger cars in early 2010 as described in
Chapter 2. These features and associated availability is shown in the table below.
For features that have been in existence for a long time but exact introduction date
is unknown, i.e. fog lamps, the year 1938 is used.

Table A.1: Common vehicle features and availability on 2010 pas-

senger cars.
Introduction Date

Chevrolet Malibu
Chevrolet Impala
Toyota Corolla
Toyota Camry

Honda Accord

Nissan Altima

Toyota Prius
Honda Civic

Ford Fusion

Ford Focus

Feature
Trip Computer / Digital Clock 1958 S S S S S S S S S S
Fold-down seats 1965 S S S S S S S S S S
Remote Fuel / Trunk Release 1988 S S S S S S S S S S
12V Power Outlets 1938 S S S S S S S S S S
Cup Holders 1938 S S S S S S S S S S
Readings Lights 1938 S O S S S S S S S S
Center Console / Armrest 1938 S S S O S S S S N S
Aux / iPod input 2001 S S S O O S S S O S
Auto On/Off Headlamps 1952 S N S N O O S S N S
Continued on next page

85
 Table A.1 – continued from previous page

Introduction Date

Chevrolet Malibu
Chevrolet Impala
Toyota Corolla
Toyota Camry

Honda Accord

Nissan Altima

Toyota Prius
Honda Civic

Ford Fusion

Ford Focus
Feature
Rear Headrests 1968 S N S N N N S S O S
15” Wheels 1938 N S N S N N N N S S
16” Wheels 1938 S O S O S S N N O N
17” Wheels 1938 O O O O O O S S O O
18” Wheels 1938 N N O N N O O O N N
19” Wheels 1938 N N N N N N N N N N
20” Wheels 1938 N N N N N N N N N N
Wood / Metal Interior Trim 1938 O O O N S O S S O N
Illuminated Entry 1983 O O N N N O S S O S
Body Kit 1938 O O O O O O O O O S
Rear Spoiler 1938 O O O O O O O O O S
Fog Lamps 1938 O O O O O O O O O O
Multi-Beam / HID headlamps 1991 S N N N O S N N N O
Overhead Console 1938 O N N N S N N N N S
Universal Garage Door Opener 1983 O N O N O N O O N O
Trunk Organizer 1938 S N N N N N N N N N
Rear Sunshade 1938 O N N N N N O O N N
Active Cruise Control 1995 N N N N N N N N N O
Lane Departure Warning 2004 N N N N N N N N N O
Blind Spot Detection 2007 N N N N N O N N N N
Rain-Sensing Wipers 1996 N N N N N N N N N N
Night Vision Camera 2007 N N N N N N N N N N
Semi-Automatic Gearbox 1997 N N N N N N N N N N
Power Rear Seats 1979 N N N N N N N N N N
Variable Ratio Steering 1969 N N N N N N N N N N
Heated Steering Wheel 1999 N N N N N N N N N N
Active Headlamps 2003 N N N N N N N N N N
Soft-Close trunk 1991 N N N N N N N N N N
Soft-Close doors 1991 N N N N N N N N N N
Continued on next page

86
 Feature
Power Adj. Wheel
Side-View Cameras
Rain-Sensing Wipers

Introduction Date
1996
2007
1965

N
N
N

Toyota Camry

87
N
N
N

Toyota Corolla
N
N
N

Honda Accord
N
N
N

Honda Civic
N
N
N

Nissan Altima
N
N
N

Ford Fusion
Table A.1 – continued from previous page

N
N
N

Chevrolet Impala
N
N
N

Chevrolet Malibu
N
N
N

Ford Focus
N
N
N

Toyota Prius
88
Appendix B

Mass Decompounding Studies

Table B.1: Literature that cites the concept of secondary mass savings, uses it in an
analysis of vehicle lightweighting or attempts to quantify it. Updated and expanded
from [Bjelkengren, 2008]
Mass De- Method Used Reference
compound-
ing Coeffi-
cient
23% - 56% Audi experience [International Aluminum Institute, 2007,
European Aluminum Association, 2008]
140% Renault engineer opinion [Daniels, 1987]
50% industry expert opinion [Office of Technology Assessment, 1995]
50%-70% industry rule of thumb [Patton and Edwards, 2002]
50%-100% referenced literature [Lloyd and Lave, 2003]
50% referenced literature [Lorenz, 2005]
50%-80% referenced literature [Asnafi et al., 2003]
50% industry rule of thumb [Das, 2005]
25% physical modeling of VW Lupo [Van Acker, 2009, J. R. Duflou et al., 2009]
1.4 model lightweighted with
Carbon Fiber Reinforced Plastic
(CFRP)
52% physical modeling of prototype [Stodolsky, F. et al., 1995]
Ford Mercury Sable
64%-68% physical vehicle modeling [Bull et al., 2008,
Aluminum Transportation Group, 2006]
54% statistical regression [Artinian and S. L. Terry, 1961]
108%-134% statistical regression [Adams, D. G. et al., 1975]
72%-212% statistical regression [GM Cost of Weight Task Force, 1975,
Padovini and GM Corporate Mass Core Group, 1981]
123%-147% statistical regression and calcula- [Kato and Shiroi, 1992]
tion of loads
70%-180% statistical regression [Malen and Reddy, 2007]
82%-129% statistical regression [Bjelkengren, 2008,
Bjelkengren, C. et al., 2008, Lee, 2010]

89
90
Appendix C

Safety Features: Take Rates and

Regression

91
92
Appendix D

Powertrain Features: Take Rates

and Regression

93
94
Appendix E

Comfort and Convenience

Features: Take Rates and
Regression

95
96
97
98
Appendix F

Other: Features with Limited Data

Due to limited data availability the deployment trends shown here are not included
in analyses in Chapter 4. However, some of these features are used for mass analysis
in Chapter 2.

99
100
101
102
Appendix G

Regression Data Points

103
Table G.1: This table includes data points and regression values for features logistic
regressions analyzed in Chapter 4.
Feature Category Required? Introduction Max Growth Max Growth
(year, approx.) (year, fit) (%/year, fit)
Automatic Trans- Comfort/Convenience No 1940 1961 2.2%
mission
Power Steering Comfort/Convenience No 1951 1970 3.2%
Intermittent Confort/Convenience Yes 1963 1983 6.1%
Wiper/Washers
Rear window defog- Comfort/Convenience Yes 1938 1979 4.2%
ger
Air Conditioning Comfort/Convenience No 1939 1973 2.8%
Power Windows Comfort/Convenience No 1940 1990 3.2%
Tilt/Telescopic Comfort/Convenience No 1965 1983 3.5%
Wheel
Power Locks Comfort/Convenience No 1956 1988 3.6%
Power Mirrors Comfort/Convenience No 1946 1988 3.1%
Keyless En- Comfort/Convenience No 1983 1998 7.7%
try/SmartKey
Cruise Control Comfort/Convenience No 1958 1984 2.9%
AM/FM Radio Comfort/Convenience No 1932 1964 2.1%
CD/Cassette Comfort/Convenience No 1965 1988 5.3%
Premium Sound Comfort/Convenience No unknown 2011 1.9%
Power Front Seats Comfort/Convenience No 1946 2007 1.1%
Advanced Climate Comfort/Convenience No 1954 2024 1.2%
Control
Moonroof Comfort/Convenience No 1956 1996 1.0%
Satellite Radio Comfort/Convenience No 2002 2010 11.6%
Rear DVD Player Comfort/Convenience No 1998 2019 0.8%
Front Wheel Drive Powertrain Partial 1929 1983 8.7%
Four Wheel Drive Powertrain No 1905 2036 2.4%
Fuel Injection Powertrain Yes 1955 1985 13.4%
Multivalve Powertrain Yes 1922 1995 4.3%
VVT Powertrain Yes 1970 2005 6.6%
Dual Master Cylin- Safety Yes 1921 1966 22.5%
ders
Front Disc Brakes Safety Yes 1949 1971 19.1%
Power Brakes Safety Yes 1928 1969 4.5%
Anti-Lock Brakes Safety Pending 1971 1998 4.0%
Driver Front Safety Yes 1981 1990 23.9%
Airbag
Dual Front Airbags Safety Yes 1987 1994 23.9%
Side Impact Beams Safety Yes unknown 1972 19.2%
Side Airbags Safety No 1995 2005 9.0%
Curtain Airbags Safety Pending 1998 2006 12.5%
Traction / Stability Safety Pending 1971 2004 5.9%
Control
Alarm / Immobi- Safety Yes 1920 2000 5.3%
lizer

104
Table G.2: Relevant parameters for Z curve fit on feature time series data. CC =
Comfort/Convenience, PT = Powertrain, SAF = Safety.
Feature Category Regulated? Develop Deploy Ratio
Automatic Trans- CC No 1 40 0.0
mission
Power Steering CC No 3 31 0.1
Intermittent CC Yes 12 16 0.7
Wiper/Washers
Rear window defog- Comfort/Convenience Yes 29 24 1.2
ger
Air Conditioning CC No 16 36 0.5
Power Windows CC No 32 31 1.0
Tilt/Telescopic CC No 4 29 0.1
Wheel
Power Locks CC No 18 28 0.7
Power Mirrors CC No 26 32 0.8
Keyless En- CC No 9 12 0.7
try/SmartKey
Cruise Control CC No 10 33 0.3
AM/FM Radio CC No 8 48 0.2
CD/Cassette CC No 14 19 0.7
Premium Sound CC No unknown unknown N/A
Power Front Seats CC No 16 91 0.2
Advanced Climate CC No 28 83 0.3
Control
Moonroof CC No 19 41 0.5
Satellite Radio CC No 4 8 0.5
Rear DVD Player CC No 11 21 0.5
Front Wheel Drive PT Partial 49 10 5.1
Four Wheel Drive PT No 110 42 2.6
Fuel Injection PT Yes 26 7 3.5
Multivalve PT Yes 62 22 2.8
VVT PT Yes 17 15 1.2
Dual Master Cylin- SAF Yes 43 4 9.6
ders
Front Disc Brakes SAF Yes 19 5 3.7
Power Brakes Safety Yes 30 22 1.3
Anti-Lock Brakes SAF Pending 15 25 0.6
Driver Front SAF Yes 9 4 2.1
Airbag
Dual Front Airbags SAF Yes 5 4 1.2
Side Impact Beams SAF Yes unknown unknown N/A
Side Airbags SAF No 5 11 0.4
Curtain Airbags SAF Pending 4 8 0.5
Traction / Stability SAF Pending 25 17 1.4
Control
Alarm / Immobi- SAF Yes 73 13 5.4
lizer

105
106
Bibliography

[Adams, D. G. et al., 1975] Adams, D. G. et al. (1975). High strength materials and
vehicle weight reduction analysis. SAE Transactions, 1975(750221):797.

[Aluminum Transportation Group, 2006] Aluminum Transportation Group (2006).

Aluminum as a cost effective technology. Technical report, The Aluminum As-
sociation, Inc. Available online at http://www.autoaluminum.org/main/global-
aluminum-content/altg-media-kit.

[Amey, 1995] Amey, R. G. (1995). Automotive component innovation: development

and diffusion of engine management technologies. Technovation, 15(4):211–223.

[An and DeCicco, 2007] An, F. and DeCicco, J. M. (2007). Trends in technical ef-
ficiency trade-offs for the U.S. light vehicle fleet. SAE Technical Paper Series,
01(1325).

[Artinian and S. L. Terry, 1961] Artinian, L. and S. L. Terry (1961). The total cost
of weight. SAE National Automobile Week. SAE Transactions: Detroit, MI.

[Asnafi et al., 2003] Asnafi, N., T. Nilsson, and G. Lass (2003). Tubular hydroforming
of automotive side members with extruded aluminum profiles. Journal of Materials
Processing Technology., 142(1):93.

[Bandivadekar, 2008] Bandivadekar, A. P. (2008). Evaluating the Impact of Advanced

Vehicle and Fuel Technologies in the U.S. Light-Duty Vehicle Fleet. PhD disserta-
tion, Massachusetts Institute of Technology, Technology, Management and Policy.

[Bass, 1969] Bass, F. M. (1969). A new product growth model for consumer durables.
Management Science, 15(5):215–227.

[Bjelkengren, 2008] Bjelkengren, C. (2008). The impact of mass decom-

pounding on assessing the value of vehicle lightweighting. Techni-
cal report, Massachusetts Institute of Technology. Available online at
http://dspace.mit.edu/handle/1721.1/44384.

[Bjelkengren, C. et al., 2008] Bjelkengren, C. et al. (2008). Predicting mass decom-

pounding opportunities during the vehicle development process. GM Internal Re-
port CL-08-302MPL, Detroit, MI.

107
[Bull et al., 2008] Bull, M., R. Chavali, and A. Mascarin (2008). Bene-
fit analysis: Use of aluminum structures in conjunction with alternative
powertrain technologies in automobiles. Technical report, IBIS Associates,
Inc for the Aluminum Association, Waltham, MA. Available online at
http://aluminumintransportation.org/downloads/IBIS-Powertrain-Study.pdf.

[Cheah, 2010] Cheah, L. W. (2010). Cars on a Diet: The Material and

Energy Impacts of Passenger Vehicle Weight Reduction in the U.S.
PhD dissertation, Massachusetts Institute of Technology, Engineering
Systems Division. Available online at http://web.mit.edu/sloan-auto-
lab/research/beforeh2/files/LCheah PhD thesis 2010.pdf.

[Chon and Heywood, 2000] Chon, D. M. and Heywood, J. (2000). Performance scal-
ing of spark-ignition engines: Correlation and historical analysis of production
engine data. Technical Report SAE Paper No. 2000- 01-0565, SAE Interational.

[Christensen et al., 2002] Christensen, C. M., Johnson, M. W., and Rigby, D. K.

(2002). Foundations for growth: How to identify and build disruptive new business.
MIT Sloan Management Review, 43(3):22–31.

[Clark and Fujimoto, 1991] Clark, K. B. and Fujimoto, T. (1991). Product Devel-
opment Performance: Strategy, organization, and management in the world auto
industry, chapter 1.2. Harvard Business School Press.

[Daniels, 1987] Daniels, J. (1987). Researching recovery. Autocar, (4):73–78.

[Das, 2005] Das, S. (2005). Life cycle energy impacts of automotive liftgate inner.
Resources, Conservation, and Recycling, 43:375.

[DeCicco, 2010] DeCicco, J. M. (2010). A fuel efficiency horizon for U.S. automo-
biles. Technical report, University of Michigan, School of Natural Resources and
Environment. Report prepared for The Energy Foundation. Available online at
http://energy.umich.edu/info/pdfs/Fuel%20Efficiency%20Horizon%20FINAL.pdf.

[DOT, 2004a] DOT (2004a). Cost and weight added by the fed-
eral motor vehicle safety standards for model years 1968-2001 in
passenger cars and light trucks. Technical Report DOT HS 809
834, U.S. Department of Transportation. Available online at
http://www.nhtsa.dot.gov/cars/rules/regrev/evaluate/pdf/809834Part1.pdf
and http://www.nhtsa.dot.gov/cars/rules/regrev/evaluate/pdf/809834Part2.pdf.

[DOT, 2004b] DOT (2004b). Lives saved by the federal motor vehicle safety
standards and other vehicle safety technologies, 1960-2002. Technical Report
DOT HS 809 833, U.S. Department of Transportation. Available online at
http://www.nhtsa.dot.gov/cars/rules/regrev/evaluate/pdf/809833Part1.pdf and
http://www.nhtsa.dot.gov/cars/rules/regrev/evaluate/pdf/809833Part2.pdf.

108
[EIA, 2010] EIA (2010). Annual energy review, 2009. Technical Report
DOE/EIA-0384(2009), Energy Information Administration. Available online at
http://www.eia.doe.gov/emeu/aer/pdf/aer.pdf.

[Ellison et al., 1995] Ellison, D. J., Clark, K. B., Fujimoto, T., and suk Hyun, Y.
(1995). Product development performance in the auto industry: 1990s update.
Technical report, International Motor Vehicle Program.

[EPA, 1999] EPA (1999). U.S. EPA tier 2 vehicle emissions standards final rule
making. Regulatory impact analysis, U.S. EPA Office of Mobile Sources. Available
online at http://www.epa.gov/tier2/frm/ria/chiv.pdf.

[EPA, 2010] EPA (2010). Light-duty automotive technology, carbon diox-

ide emissions, and fuel economy trends: 1975 through 2010. Techni-
cal Report EPA-420-R-10-023, Office ofTransportation and Air Qual-
ity, U.S. Environmental Protection Agency. Available online at
http://www.epa.gov/oms/cert/mpg/fetrends/420r10023.pdf.

[EPA, 2011] EPA (2011). Mobile source emissions - past, present, and future: Mile-
stones. EPA Office of Transportation and Air Quality website. Available online at
http://www.epa.gov/otaq/invntory/overview/solutions/milestones.htm.

[European Aluminum Association, 2008] European Aluminum Association

(2008). Aluminum in cars. Technical report. Available online at
http://www.eaa.net/upl/4/en/doc/Aluminium in cars Sept2008.pdf.

[Glennan, 2007] Glennan, T. B. (2007). Strategies for managing vehicle mass

throughout the development process and vehicle lifecycle. SAE Technical Paper
Series 2007-01-1721, General Motors Corporation, Detroit, MI.

[GM Cost of Weight Task Force, 1975] GM Cost of Weight Task Force (1975). Re-
port of the cost of weight task force. Technical report, General Motors, Detroit,
MI.

[Heavenrich, 2006] Heavenrich, R. M. (2006). Light-duty automotive technology and

fuel economy trends: 1975 through 2006. Technical Report EPA 420-R-06-011,
U.S. EPA Office of Transportation and Air Quality.

[Heywood and Welling, 2009] Heywood, J. B. and Welling, O. Z. (2009). Trends

in performance characteristics of modern automobile si and diesel engines. SAE
International Journal of Engines, 2(1):1650–1662.

[International Aluminum Institute, 2007] International Aluminum Institute (2007).

Improving sustainability in the transport sector through weight reduction and
teh application of aluminum. Technical report, European Aluminum Asso-
ciation, The Aluminum Assocation. Available online at http://www.world-
aluminium.org/cache/fl0000172.pdf.

109
[J. R. Duflou et al., 2009] J. R. Duflou et al. (2009). Environmental impact analysis
of composite use in car manufacturing. CIRP Annals - Manufacturing Technology,
58(1):9.

[Kato and Shiroi, 1992] Kato, K. and Shiroi, Y. (1992). The effect of growth fac-
tor on weight reduction. International Symposium on Automotive Technology and
Automation, page 920422.

[Knittel, 2009] Knittel, C. (2009). Automobiles on steroids: Product at-

tribute trade-offs and technological progress in the automobile sector.
The American Economic Review. In Submission. Available online at
http://web.mit.edu/knittel/www/papers/steroids latest.pdf.

[Kramer and Haigh, 2009] Kramer, G. J. and Haigh, M. (2009). No quick switch to
low-carbon energy. Nature, 462(3):568–569.

[Lee, 2010] Lee, T. (2010). Overview of the mass decompounding methodoloft and
model using updated gm and competitor midsized vehicles. GM Internal Report
CML-056, Detroit, MI.

[Lloyd and Lave, 2003] Lloyd, S. and Lave, L. (2003). Life cycle economic and en-
vironmental implications of using nanocomposites in automobiles. Environmental
Science & Technology, 37.

[Lorenz, 2005] Lorenz, D. (2005). Modeling and analysis of integrated hotforming

and quenching process. Advanced Materials Research.

[MacKenzie, 2009] MacKenzie, D. W. (2009). Trends and drivers of the per-

formance fuel economy tradeoff in new automobiles. Master’s thesis, Mas-
sachusetts Institute of Technology, Technology and Policy Program, Engi-
neering Systems Division. Available online at http://web.mit.edu/sloan-auto-
lab/research/beforeh2/files/MacKenzie-SM%20Thesis-TPP-2009.pdf.

[Malen and Reddy, 2007] Malen, D. and Reddy, K. (2007). Preliminary vehicle mass
estimation using empirical subsystem influence coefficients. Technical report, The
FGPC-Mass Compounding Project Team Auto/Steel Partnership., Ann Arbor, MI.
Available online at http://www.a-sp.org/publications.htm.

[MIT, 2008] MIT (2008). On the road in 2035: Reducing transportation’s

petroleum consumption and ghg emissions. Technical Report Report No. LFEE
2008-05, Massachusetts Institute of Technology Laboratory for Energy and the
Environment, Cambridge, MA. Available online at http://web.mit.edu/sloan-auto-
lab/research/beforeh2/otr2035/On%20the%20Road%20in%202035 MIT July%202008.pdf.

[Nakicenovic, 1986] Nakicenovic, N. (1986). The automobile road to technological

change: Diffusion of the automobile as a process of technological substitution.
Technological Forecasting and Social Change, 29(4):309.

110
[NHTSA, 2010] NHTSA (2010). 2016 corporate average fuel economy final regulatory
impact analysis. Technical Report 7, Office of Regulatory Analysis and Evaluation
and National Center for Statistics and Analysis.
[NHTSA, 2011] NHTSA (2011). Preventing backovers in america’s driveways.
http://www.nhtsa.gov/people/injury/pedbimot/ped/BackoversTry2/index.htm.
[Office of Technology Assessment, 1995] Office of Technology Assessment (1995). Ad-
vanced automotive technology: Visions of a super-efficient family car. Tech-
nical Report OTA-ETI-638, GPO stock 052-003-01440-8. Available online at
http://www.princeton.edu/ ota/disk1/1995/9514/9514.PDF.
[Padovini and GM Corporate Mass Core Group, 1981] Padovini, M. and GM Corpo-
rate Mass Core Group (1981). Mass influence coefficients. Internal report, General
Motors, Detroit, MI.
[Patton and Edwards, 2002] Patton, R. and Edwards, M. (2002). Estimating weight
reduction effects of material subsitution on joints with constant stiffness. Technical
Report 2002.
[Rogers, 2003] Rogers, E. M. (2003). Diffusion of Innovations, 5th Edition. Free
Press, New York, NY, fifth edition.
[Safe Kids USA, 2011] Safe Kids USA (2011). Safety in and around cars.
http://www.usa.safekids.org/our-work/programs/buckle-up/in-and-around-
cars.html.
[Schafer et al., 2006] Schafer, A., Heywood, J. B., and Weiss, M. A. (2006). Future
fuel cell and internal combustion engine automobile technologies: A 25-year life
cycle and fleet impact assessment. Energy, 31(12):2064–2087.
[Sterman, 2000] Sterman, J. (2000). Business dynamics: Systems thinking and mod-
eling for a complex world., page 631. Irwin/McGraw-Hill, Boston.
[Stodolsky, F. et al., 1995] Stodolsky, F. et al. (1995). Life-cycle energy savings po-
tential from aluminum-intensive vehicles. Total Life Cycle Conference & Exposition
- SAE, page 47.
[Van Acker, 2009] Van Acker, K. (2009). Lightweight materials for the automotive:
environmental impact analysis of the use of composites. Selected papers from the
3rd International Seminar on Society & Materials 12, SAM3 Freiberg.
[Ward’s, 2010a] Ward’s (2010a). Factory-installed equipment on imported cars and
light trucks, 1975 - 2010. Technical report, Ward’s Automotive Group, a Division
of Penton Media Inc.
[Ward’s, 2010b] Ward’s (2010b). Factory-installed equipment on U.S. domestic cars
and light trucks, 1975 - 2010. Technical report, Ward’s Automotive Group, a
Division of Penton Media Inc.

111