ISSN 1055- 1425
August 2004
This work was performed as part of the California PATH Program of the
University of California, in cooperation with the State of California Business,
Transportation, and Housing Agency, Department of Transportation; and the
United States Department of Transportation, Federal Highway Administration.
The contents of this report reflect the views of the authors who are responsible
for the facts and the accuracy of the data presented herein. The contents do not
necessarily reflect the official views or policies of the State of California. This
report does not constitute a standard, specification, or regulation.
Final Report for Task Order 4104
CALIFORNIA PATH PROGRAM
INSTITUTE OF TRANSPORTATION STUDIES
UNIVERSITY OF CALIFORNIA, BERKELEY
Carlink II: A Commuter Carsharing
Pilot Program Final Report
UCB- ITS- PRR- 2004- 23
California PATH Research Report
Susan Shaheen, Kamill Wipyewski, Caroline Rodier,
Linda Novick, Molly Anne Meyn, John Wright
CALIFORNIA PARTNERS FOR ADVANCED TRANSIT AND HIGHWAYS
CARLINK II: A COMMUTER CARSHARING
PILOT PROGRAM
FINAL REPORT
Prepared for
California Partners for Advanced Transit and Highways
Memorandum of Understanding 4104
Prepared by
Susan Shaheen, Ph. D.
Kamill Wipyewski
Caroline Rodier, Ph. D.
Linda Novick
MollyAnne Meyn
John Wright
Partners for Advanced Transit and Highway ( PATH)
University of California, Berkeley
and
Institute of Transportation Studies- Davis,
University of California, Davis
June 2004
iii
ACKNOWLEDGMENTS
The authors would like to thank California Partners for Advanced Transit and Highways
( PATH), the California Department of Transportation ( Caltrans), American Honda Motor
Company, Caltrain, the Bay Area Rapid Transit ( BART) District, Lawrence Livermore National
Laboratory ( LLNL), Teletrac, and INVERS for their generous contributions to the CarLink I and
II programs. Each helped to make CarLink I and II testing and research possible.
We would also like to express appreciation to the CarLink I and II project partners who
supported the field test and pilot program, particularly: Clifford Loveland, Robert Justice,
Lindsee Tanimoto, Terry Parker, and William Tournay of Caltrans; Robert Uyeki of Honda
R& D North Americas, Incorporated; Robert Bienenfeld of Honda Motor Company; Janet
McGovern, Brian Fitpatrick, Gary Cardona and Steve Hanson of Caltrain; Amanda Jones of the
City of Palo Alto; Ramsey Shuayto of the Stanford Management Park; Victoria Nerenberg
( formerly of BART); Erma Paddack and Sal Ruiz of LLNL; Robert Tam ( formerly of PATH);
Stanley Polk ( formerly of Teletrac); and Uwe Latsch of INVERS. Linda Novick and Barbara
Bower of the Institute of Transportation Studies- Davis ( ITS- Davis), Dr. Richard Katzev, and
Bob Reese provided valuable assistance in managing the CarLink I and II programs in the field.
ITS- Davis and ITS- Berkeley faculty, staff, and students also deserve special credit for their
assistance with CarLink I and II ( between 1998 and 2003), including: Professor Daniel Sperling,
Joe Krovoza, John Wright, Mollyanne Meyn, Oliver Burke, Dimitri Loukakos, Rebecca Pearson,
Brian Hammons, Brian Hueng, Michael Paraiso, Amanda Eaken, Monica Bally- Urban, David
Dick, Jie Lin, Terrance Polen, Jennifer Ingersoll, Bryan Jones, Robin Owen, Monica Bally-
Urban, and John McCann. Additionally, thanks go to Shirley Long, Susan O’Bryant, and Candy
Clarke of ITS- Davis and Jim Fong of PATH for their indispensable assistance in managing the
CarLink I and II budgets.
This work was performed as part of the California PATH program of the University of California
in cooperation with the State of California Business, Transportation, and Housing Agency;
California Department of Transportation; the United States Department of Transportation; and
Federal Highway Administration.
The contents of this report reflect the views of the authors who are responsible for the facts and
the accuracy of the data presented herein. The contents do not necessarily reflect the official
views or policies of the State of California. This report does not constitute a standard,
specification, or regulation.
iv
v
TABLE OF CONTENTS
ACKNOWLEDGMENTS iii
EXECUTIVE SUMMARY ix
CHAPTER ONE: U. S. SHARED- USE VEHICLE FINDINGS: OPPORTUNITIES
AND OBSTACLES FOR CARSHARING & STATION CAR
GROWTH
• SECTION 1.0: INTRODUCTION 2
• SECTION 1.1: U. S. MARKET DEVELOPMENTS OF SHARED- USE VEHICLE
SERVICES 3
• SECTION 1.2: SHARED- USE VEHICLE MARKET GROWTH: OBSTACLES
AND OPPORTUNITIES 5
• SECTION 1.3: CONCLUSION 15
CHAPTER TWO: A FRAMEWORK FOR TESTING INNOVATIVE
TRANSPORTATION SOLUTIONS: A CASE STUDY OF
CARLINK— A COMMUTER CARSHARING PROGRAM
• SECTION 2.0: INTRODUCTION 20
• SECTION 2.1: CARLINK PROGRAM AND RESEARCH OVERVIEW 21
• SECTION 2.2: EARLY LESSONS LEARNED 24
• SECTION 2.3: CARLINK II USER & OPERATIONAL UNDERSTANDING 27
• SECTION 2.4: PILOT TRANSITION 33
• SECTION 2.5: CONCLUSION 34
CHAPTER THREE: TRAVEL EFFECTS OF A SUBURBAN COMMUTER-CARSHARING
SERVICE: A CARLINK CASE STUDY
• SECTION 3.0: INTRODUCTION 38
• SECTION 3.1: CARSHARING IMPACTS IN EUROPE AND THE UNITED
STATES 40
• SECTION 3.2: OVERVIEW OF CARLINK I AND II 41
• SECTION 3.3: RESEARCH AND DATA COLLECTION METHODOLOGY 44
• SECTION 3.4: EARLY ADOPTER MARKET PROFILE 46
• SECTION 3.5: CARLINK II TRAVEL EFFECTS 52
• SECTION 3.6: CONCLUSION 57
vi
CHAPTER FOUR: APPLYING INTEGRATED ITS TECHNOLOGIES TO
CARSHARING SYSTEM MANAGEMENT: A CARLINK CASE STUDY
• SECTION 4.0: INTRODUCTION 62
• SECTION 4.1: U. S. CARSHARING MARKET DEVELOPMENTS 62
• SECTION 4.2: CARLINK II: BUSINESS MODEL, TECHNOLOGY, AND
FINDINGS 64
• SECTION 4.3: CARLINK: A BRIEF OVERVIEW 64
• SECTION 4.4: CARLINK TECHNOLOGY 67
• SECTION 4.5: CARLINK II: TECHNOLOGY LESSONS LEARNED 68
• SECTION 4.6: CONCLUSION 72
CHAPTER FIVE: CARLINK— A COMMUTER CARSHARING MODEL:
CONDITIONS FOR ECONOMIC VIABILITY
• SECTION 5.0: INTRODUCTION 78
• SECTION 5.1: CARLINK: A COMMUTER CARSHARING MODEL 79
• SECTION 5.2: SCENARIO ANALYSIS: CONDITIONS FOR ECONOMIC
VIABILITY 84
• SECTION 5.3: RECOMMENDATIONS 88
• SECTION 5.4: CONCLUSION 91
APPENDIX 1: CARLINK II SURVEY INSTRUMENTS
• INITIAL QUESTIONNAIRE FOR CARLINK II USER 94
• END QUESTIONNAIRE FOR CARLINK II USER 105
• INITIAL QUESTIONAIRE FOR HOUSEHOLD 115
• END QUESTIONNAIRE FOR HOUSEHOLD 123
APPENDIX II: TRAVEL DIARY 128
APPENDIX II: CARLINK II INTERIM INTERVIEWS 132
APPENDIX IV: CARLINK II FOCUS GROUP SUMMARIES 150
vii
LIST OF FIGURES
• FIGURE 1.1: GROWTH IN STATION CAR AND CARSHARING
ORGANIZATIONS 4
• FIGURE 1.2: GROWTH IN CARSHARING MEMBERSHIP AND VEHICLES 5
• FIGURE 1.3: TECHNOLOGY LEVELS IN U. S. CARSHARING
ORGANIZATIONS 13
• FIGURE 2.1: SATISFACTION WITH CARLINK II FEATURES 28
• FIGURE 2.2: SATISFACTION WITH OTHER CARLINK II FEATURES 30
• FIGURE 3.1: THE CARLINK MODEL ( CONSISTING OF THREE USER GROUPS:
HOMEBASED USERS, WORKBASED COMMUTERS, AND
WORKBASED DAY USERS) 39
• FIGURE 3.2: GENDER OF CARLINK MEMBERS RELATIVE TO BAY AREA
( 2000 CENSUS) 47
• FIGURE 3.3: DISTRIBUTION OF NUMBER OF VEHICLES PER HOUSEHOLD
( CARLINK II VS. BAY AREA) 49
• FIGURE 3.4: DISTRIBUTION OF NUMBER OF VEHICLES PER HOUSEHOLD
FOR CARLINK II MEMBERS 49
• FIGURE 4.1: U. S. CARSHARING MEMBERSHIP AND VEHICLE FLEET
GROWTH 63
• FIGURE 4.2: THE CARLINK MODEL ( CONSISTING OF THREE USER GROUPS:
HOMEBASED USERS, WORKBASED COMMUTERS, AND
WORKBASED DAY USERS) 65
• FIGURE 4.3: USER SATISFACTION WITH CARLINK II FEATURES 69
• FIGURE 5.1: CARLINK II COST DISTRIBUTION 82
viii
LIST OF TABLES
• TABLE 1.1: RATING FACTORS FOR SHARED- USE VEHICLES 8
• TABLE 2.1: DIFFERENCES BETWEEN CARLINK I AND II 22
• TABLE 3.1: KEY DIFFERENCES BETWEEN CARLINK I AND CARLINK II 43
• TABLE 3.2: CARLINK II RESPONSE RATES BY GENDER AND USER
GROUPS 45
• TABLE 3.3: DISTRIBUTION OF CARLINK II PARTICIPANTS AND SURVEY
RESPONDENTS BY USER GROUP 45
• TABLE 3.4: AGE OF CARLINK MEMBERS RELATIVE TO BAY AREA
RESIDENTS ( 2000 CENSUS) 47
• TABLE 3.5: OCCUPATION DISTRIBUTION OF CARLINK PARTICIPANTS
RELATIVE TO BAY AREA RESIDENTS 48
• TABLE 3.6: PARTICIPANTS’ LEAST FAVORITE ATTRIBUTES OF
TRANSPORTATION MODES BEFORE CARLINK II 50
• TABLE 3.7: PSYCHOGRAPHIC SCALE SCORES FROM CARLINK II AND I 51
• TABLE 3.8: BEFORE AND AFTER COMMUTE MODE SHARES FOR
CARLINK II PARTICPANTS 52
• TABLE 3.9: BEFORE AND AFTER AVERAGE DAILY ROUND TRIP
COMMUTE VMT AND TRAVEL TIME ( MINUTES) FOR
CARLINK II PARTICIPANTS 53
• TABLE 3.10: THE EFFECT OF CARLINK II PARTICIPATION ON
COMMUTE STRESS 54
• TABLE 3.11: CHANGE IN CARLINK II NON- COMMUTE MODE SHARE 55
• TABLE 3.12: PERCENTAGE POINT CHANGE IN CARLINK I MODE
SHARE FOR ALL TRIP PURPOSES BY USER GROUP 55
• TABLE 3.13: USE OF PERSONAL VEHICLE( S) AFTER JOINING CARLINK II 56
• TABLE 4.1: DIFFERENCES BETWEEN CARLINK I AND II 66
• TABLE 4.2: COMPARISON OF CARLINK I AND II TECHNOLOGY 68
• TABLE 4.3: NUMBER OF TIMES RESERVED VEHICLE NOT AVAILABLE 70
• TABLE 5.1: ACTUAL CARLINK II COSTS 80
• TABLE 5.2: ACTUAL CARLINK II REVENUES 84
• TABLE 5.3: QUANTITATIVE RESULTS OF SCENARIO MODELS 87
ix
EXECUTIVE SUMMARY
CarLink II was a commuter- based carsharing pilot project administered by the Institute of
Transportation Studies at the University of California, Davis ( ITS- Davis) in conjunction with
Caltrans, American Honda Motor Company, and Caltrain. Partners for Advanced Transit and
Highways ( PATH) researchers conducted the evaluation. Pilot objectives included testing an
advanced carsharing system, understanding user response to this service, and testing its long-term
sustainability. From July 1, 2001 to June 30, 2002, the CarLink II program was deployed in
the field and continued the investigation of commuter- based carsharing originally explored in the
1998 CarLink longitudinal survey and the 1999 CarLink I field test. Lessons learned during the
CarLink I field test helped guide the project team’s design of the CarLink II project, resulting in
several differences and improvements. The table below summarizes the major differences
between CarLink I and II.
Differences between CarLink I and II
STUDY
CHARACTERISTICS
CARLINK I CARLINK II
Number of Vehicles 12 Vehicles 19 Vehicles
Primary Transit Partner BART Caltrain
Transit Station Location Dublin/ Pleasanton Palo Alto
Vehicle Type Compressed natural gas Honda
Civics
Ultra- low emission Honda Civics
Homebased Users Up to 10 households, pay $ 200
per month.
Up to 16 households, pay $ 300 per
month.
Workbased Commuters Up to 20 Lawrence Livermore
National Laboratory ( LLNL)
employees pay $ 60 per carpool
($ 30 each).
Up to 63 employees of businesses at
Stanford Research Park ( primarily),
share CarLink vehicles to carpool
to/ from work. Businesses pay $ 350
per month per vehicle ( a combined
fee) for Workbased Commuter and
Day Use services ( in contrast to
employees paying for this service
independently as in CarLink I).
Workbased Day Users Employees of LLNL pay $ 1.50
per hour and $. 10 per mile.
Up to 28 employees of Stanford
Research Park companies and other
nearby businesses have access to
vehicles for business and personal
use. Employers pay $ 350 per vehicle
per month to subscribe to the
combined Workbased Commuter and
Day Use services.
Total Users 54 107
Employer One: LLNL Six: Several private companies
at/ nearby Stanford Research Park
x
Technology In- vehicle tracking, smart key
kiosk at transit station, smart
cards, manual key boxes at
LLNL, and on- line scheduling
system at LLNL
In- vehicle tracking, automated data
collection, smart key fob entry, PIN-based
vehicle login, on- line
reservations, and in- vehicle
navigation system
Program Length Field test designed for limited
10- month duration
Pilot program with planned
transition to on- going carsharing
service
Research Goals Document demand for commuter
carsharing service and gauge
user satisfaction and needs
Continued analysis of commuter
carsharing ( in a new setting) with
greater statistical confidence ( i. e., a
greater sample size) and new
emphasis on technology testing, its
impact on cost reduction, and longer-term
program sustainability
CarLink I: Key Findings
The original CarLink I field test ran from January to November 1999, and featured 54
individuals sharing 12 natural gas powered Honda Civics. The vehicles were based at the Bay
Area Rapid Transit ( BART) District station in Dublin- Pleasanton, the eastern terminus of the
BART system. The shared cars were equipped with smart technologies including tracking,
communication, and reservation systems to facilitate access and logistics. The model
incorporated transit- based carsharing for traditional and reverse commute travel patterns, as well
as a day- use fleet application, tested at a major employment center  the Lawrence Livermore
National Laboratory ( LLNL). The CarLink model includes three separate user groups
( Homebased Users, Workbased Commuters, and Workbased Day Users), each of which used the
vehicles differently and paid fees accordingly. Key CarLink I study findings include:
• Even though CarLink users’ commutes took approximately 10 minutes longer on average,
they found them less stressful;
• The combination of CarLink, BART, and carpooling resulted in a net commute reduction of
approximately 20 vehicle miles ( or 32.2 kilometers) per commuter per day ( on average)
across the fleet;
• CarLink resulted in at least 20 new BART trips each day; and,
• Several Homebased Users stated that if CarLink became a permanent service, they would sell
one of their personal cars, which would greatly reduce their transportation costs.
CarLink II: Pilot Program Overview
Building upon the knowledge and experience gained in CarLink I, the CarLink II program
reflects several changes to the initial model. Chief among the differences were the decisions to:
1) transition the program to an ongoing service at the end of the pilot phase, and 2) test it in a
different location with a new transit provider and business partners. CarLink II launched in Palo
xi
Alto with Caltrain and several businesses located in the Stanford Research Park in summer 2001.
The main CarLink II user components were:
• Homebased Users: This group paid $ 300 per month to have access to vehicles on
evenings and weekends. They drove a CarLink vehicle to the California Avenue Caltrain
station each weekday morning and to home each evening.
• Workbased Commuters: Members of this group were employees of Stanford Research
Park businesses, who subscribed to CarLink, and drove the vehicles between the Caltrain
station and their worksites as part of their daily commute. Their employers paid $ 50 per
month per vehicle for this service.
• Workbased Day Use: Registered employees of businesses, which paid $ 300 per vehicle
per month, had access to the vehicles during the day for personal and company trips.
• Vehicles: Nineteen 2001 Ultra Low Emission Honda Civics
• Technology: CarLink II employed a seamless, customized system that coordinated
vehicle tracking, data collection, and reservations. Users reserved vehicles over the
Internet and accessed vehicles using smart key fobs.
A total of 107 individuals participated in the CarLink II program: 16 Homebased Users, 28 Day
Users, and 63 Workbased Commuters/ Day Users. Fifty- three percent of participants were female
and 47 percent were male. Sixty- four respondents completed the final questionnaire ( a response
rate of 60 percent). Respondents included nine Homebased Users ( five male, four female); 21
Day Users ( 9 male, 12 female); and 34 Workbased Commuters ( 14 male, 20 female).
While the principal goal of the CarLink I field test was to examine response to commuter
carsharing, CarLink II focused on commercial potential and technology assessment. The primary
tools used to investigate these topics were focus groups, questionnaires, travel diaries, data
collected automatically by in- vehicle technology, operational data, and feedback from the
CarLink management staff and project partners.
CarLink II Report Overview & Key Findings
This report consists of five chapters and four appendices ( survey instruments, travel diary,
interim program interviews, and final focus group summaries). Key findings for each chapter are
provided below.
Chapter One: U. S. Shared- Use Vehicle Survey Findings: Opportunities and Obstacles for
Carsharing & Station Car Growth
This chapter provides original research on the current market for shared- use vehicle services in
the U. S. during the timeframe of the CarLink II pilot project. It was also published in
Transportation Research Record, No. 1841, pp, 90- 98 ( permission was granted by the
Transportation Research Record to publish it in this final report).
xii
Principal findings include:
• In July 2002, there were 18 shared- use vehicle organizations: eleven carsharing
organizations; five station car programs; and two carsharing research pilots ( both in
California). As of July 2002, station car programs claimed approximately 163 members
and 121 vehicles, while carsharing programs collectively claimed approximately 12,098
members and operated 455 vehicles. Not surprisingly, the majority of carsharing
members ( 80 percent) live in the 25 most densely populated cities of the nation.
• A few organizations serve the majority of U. S shared- use vehicle program members in
multiple regions. For instance, City CarShare, Flexcar, and Zipcar each operate in several
cities. In July 2002, these organizations collectively served 92 percent of all U. S.
members and deployed 78 percent of the vehicle fleet.
• The reduced number of organizational launches in 2002, amidst continuing market,
membership, and fleet expansion, indicates that entry barriers likely exist and could be
increasing. The high fixed costs of vehicle lease/ purchase, technology development, and
insurance also are significant deterrents to a new organization’s market entry. In a survey
of planned organizations, 75 percent of respondents ( n= 7) ranked insurance and smart
technologies among their top three costs. While insurance was not listed as a major cost
or concern in the 2001 survey, by July 2002 researchers found increased insurance costs a
major challenge to expansion and sustainability for the vast majority of U. S.
organizations.
• Despite promising U. S. shared- use vehicle operational and membership growth rates, the
relative small scale of these organizations presents a challenge to: 1) obtaining affordable
insurance and 2) covering other high capital costs, including technology, vehicles, and
labor. While advanced shared- use vehicle technologies can help organizations to reduce
administrative costs and potentially lower insurance premiums, technology deployment
has typically required large private investments or public development grants. Second,
cooperation among shared- use vehicle organizations could also accelerate the
development of interoperable, customized technologies and continued innovation by
creating the scale necessary to attract technology providers. Additionally, a combined
insurance- technology cooperative strategy may be even more beneficial  lowering
insurance premiums, enhancing customer services and capabilities, and lowering capital
costs.
• The potential of new and existing shared- use vehicle service organizations to continue
expanding and serving new markets could be greatly enhanced through supportive
public- private partnerships. Policymakers and transit operators, for instance, should
continue to explore the social and environmental benefits of shared- use vehicle services
through grant making, preferential parking, supportive policies ( e. g., employer pre- tax
credits), and outreach/ marketing. Strong public- private partnerships are needed to
facilitate the on- going development and sustainability of viable U. S. shared- use vehicle
programs.
xiii
Chapter Two: A Framework for Testing Innovative Transportation Solutions: A Case Study of
CarLink— A Commuter Carsharing Program
This chapter outlines the CarLink model, technology, and early lessons learned. It also provides
an overview of CarLink II operational understanding, a synopsis of the pilot program transition,
and offers recommendations for future model development.
• The CarLink II pilot program built upon six key operational lessons learned from
CarLink I:
1. Streamlining Technology: Several technology shortcomings ( i. e., key management
and vehicle tracking systems) contributed to delays and necessitated program
modification. Technology should be integrated and customized to facilitate
carsharing use. A stand- alone “ smartcard” approach should be developed and tested
in which fixed key box lots are not needed. In this way, participants could access
vehicles with smartcards alone.
2. Limited Compressed Natural Gas ( CNG) Infrastructure: During CarLink I, two
CNG issues constrained operations: a limited number of CNG refueling sites and
slow CNG refueling pumps at Lawrence Livermore National Laboratory. The CNG
component of CarLink I restricted vehicle range and participation. Also, users did
not refuel vehicles as frequently as agreed. Use of CNG vehicles in the CarLink I
field test distracted from the shared- use vehicle evaluation. In the future, this model
should be tested with internal combustion engine ( ICE) vehicles and fuel cards.
3. Guaranteed Parking: Guaranteed parking at the Dublin- Pleasanton BART station
was a huge program incentive, as parking at this station filled up prior to 7AM at
the time of the program. In the future, carsharing programs should be sited in
locations where parking is costly and limited.
4. Vehicle Cleanliness: During CarLink I, participants and operations staff cleaned
and washed cars. Nevertheless, vehicle cleanliness continued to be a chronic
program issue. Consider hiring a third party to clean vehicles more frequently.
5. Employer Participation: Day Use participation in CarLink II was limited. In the
future, test an employer- focused carsharing service with multiple companies located
in a congested corridor with transit access and parking constraints.
6. Program Duration: CarLink I was a short- term demonstration project ( i. e., 10
months), which limited understanding of user adoption and behavior because of its
short timeframe. In the future, deploy CarLink as a pilot program with the potential
to transition to an ongoing operation after the research phase ends.
xiv
CarLink II user satisfaction highlights include:
• CarLink Staff: Thirty- nine percent of respondents to the final questionnaire were very
satisfied, and 45 percent were satisfied with the CarLink II staff. No respondents were
dissatisfied or very dissatisfied. During the interim program interviews, participants also
expressed satisfaction with CarLink II staff. Indeed, 68 percent of those responding were
very satisfied with CarLink II operations personnel. Members reported that when a
problem arose that CarLink II staff responded very quickly and kept them well informed
of relevant issues.
• Transit Costs: Transit costs ( primarily Caltrain) varied for individual members. All
CarLink II member companies contributed to the transit fares of their employees. As part
of the final questionnaire, 19 percent of respondents were very satisfied, 47 percent were
satisfied, and only five percent were dissatisfied with their transit costs. Ten percent of
respondents answered not applicable, since many Day Users carpooled, vanpooled,
bicycled, or walked to work.
• Member Coordination: CarLink II required all members to coordinate with each other to
ensure that vehicles reached designated locations at required times ( e. g., Caltrain during
AM and PM commute peaks). Approximately eight percent of respondents were very
satisfied, and 38 percent were satisfied with this process. Thirty- five percent were
neutral, indicating that the majority of participants adjusted easily to schedule
coordination. It is important to note that scheduling flexibility was accommodated with
additional “ pool” vehicles in the CarLink II fleet.
Similar to CarLink I, numerous lessons were amassed during CarLink II. Key findings include:
• Parking Impacts: Building on a principal CarLink I success factor, locations with limited
parking were emphasized during the CarLink II site selection in early 2000. At this time,
the parking lot at the California Avenue Caltrain station was at close to 90 percent
capacity. However, due to the subsequent economic downturn, lot utilization decreased to
less than 60 percent by the end of CarLink II ( July 2002). This change in parking
negatively impacted program recruitment, as guaranteed parking is a significant incentive
to carsharing use, particularly when parking is oversubscribed.
• Economic Impacts: CarLink II site selection was conducted in summer 2000. At that
time, the California economy had just begun to experience a downward turn, but the
extent of this decline was not yet apparent. Earlier, the strong economy had contributed to
increased highway congestion, and many transit lots were approaching or exceeding
capacity in the Bay Area. Employers were anxious about employee retention, and Palo
Alto was concerned about the impact of congestion on quality of life. At this time, there
was no reason to believe the economic strength of Silicon Valley would diminish enough
to affect CarLink II’s longer- term operation. Silicon Valley lost approximately nine
percent of its employment from the first quarter of 2001 to the second quarter of 2002
( i. e., the period of CarLink II operations). This impact affected user demand and
xv
willingness to pay during CarLink II, economic viability ( i. e., CarLink II was unable to
cover its costs), and long- term sustainability ( i. e., transition to a third- party operator).
• Participant Recruitment and Retention: Engaging potential participants is challenging.
Recruitment remains an ongoing effort due to member attrition ( business and individual)
due to changes in home, work, or business circumstances. During CarLink II, a wide
variety of recruitment strategies were utilized with varying levels of success ( e. g., the
CarLink II website, brochures/ postcards, a video, flyers at stations and in Caltrain bills,
flyers on trains, articles/ advertisements in local papers, community meetings, carpool
lists, a trial offer, etc.). The most effective tools included the trial offer ( as noted during
the CarLink longitudinal survey as a powerful recruitment device), flyers on trains,
recommendations from Stanford Research Park, word of mouth, and e- mail
communication. Least effective methods included flyers in the Caltrain bill and at the
stations, the carpool list, and the CarLink II video.
Chapter Three: Travel Effects of A Suburban Commuter- Carsharing Service: A CarLink Case
Study
This chapter presents market and behavioral data from CarLink II and contrasts results of
CarLink I where applicable.
Key demographic variables examined include gender, age, education, income, occupation, and
vehicle ownership. Findings include:
• Gender: Men and women were equally represented in CarLink II, which is consistent
with the distribution of men and women in the Bay Area. However, in CarLink I, male
participants were disproportionately represented. The difference in gender distribution
between CarLink I and CarLink II may be explained by the demographic or attitudinal
characteristics of employees at the respective worksites.
• Age: CarLink II participants tended to be younger than the general Bay Area population
and CarLink I participants. Participants 20 to 40 years of age represented approximately
79 percent of CarLink II members and over 20 percent were between 41 and 64 years of
age. The location of CarLink II in Silicon Valley, which tends to have a relatively young
employee base, may explain the lower relative age of participants in CarLink II. In
contrast, only 41 percent of CarLink I participants were between 20 and 40 years of age,
and 59 percent were between 41 and 64 years old. The LLNL worksite in CarLink I may
explain the higher relative age of participants ( i. e., employment may require more
advanced degrees).
• Education: Participants in both CarLink I and II possessed higher levels of education than
the general Bay Area population. Fifty- seven percent of CarLink I and 48 percent of
CarLink II participants had education levels of a bachelors degree or higher. This
compares to 14 percent of Bay Area citizens over the age of 25 with a bachelors degree
or higher.
xvi
• Income: The household income levels of CarLink participants were also relatively high
compared with the Bay Area population. Thirty percent of CarLink I members had
household incomes ranging from $ 80,000 to $ 99,999, while 16 percent had a household
income greater than $ 100,000. CarLink II members had fewer participants in the
$ 80,000-$ 99,999 range ( 19 percent), but more participants earning over $ 100,000 ( 47
percent). In CarLink II, the greatest portion of all user groups was in the $ 100,000 plus
income category. However, Homebased Users tended to have a relatively large
percentage of members in the lower income groups, and the reverse was true for
Workbased Commuters. Workbased Day Users tended to have a more even distribution
across the income categories than the other user groups.
• Occupation: With higher education and income levels, CarLink members were primarily
employed in the professional/ technical category ( 68 percent in CarLink I and over 64
percent in CarLink II). This is high relative to Bay Area residents. The distribution of
occupation types did not vary substantially among user groups in CarLink II relative to
CarLink I.
• Vehicles Per Household: CarLink II participants owned or leased an average of 1.75
vehicles per household at the start of the program. Overall, the number of vehicles per
household of CarLink II participants was similar to the Bay Area population.
CarLink attitudinal profiles follow:
• Current Modes: CarLink II Workbased Commuter and Day User attitudes were slightly
positive towards their current mode. Similarly, the CarLink I field test found that 77
percent were satisfied with their current mode. These results suggest that CarLink
participants did not join CarLink because of a general dissatisfaction with their current
transportation mode.
• Congestion: Participant’s least favorite negative transportation attribute, “ Spend too
much time in traffic,” suggests that traffic congestion may be a predictor of CarLink II
participation. CarLink I results also suggested that participants may be more sensitive to
congestion than other factors.
• Environmental Concern: CarLink I and II participants expressed concern for the
environment. These results indicate that reducing automobile effects on the environment
may have been an important motivating factor for joining CarLink.
• Transit: CarLink II participants were comfortable with transit. This tended to be most
strongly true for Homebased Users, most likely because of their lower vehicle ownership
rates, lower household incomes, and somewhat younger ages.
• Experimentation: Both CarLink I and II members indicted a similar comfort level with
respect to experimentation.
xvii
• Vehicle Hassle: All the CarLink II user groups tended to disagree that “ vehicles are a
hassle.” This result differs from CarLink I. These results indicate that CarLink II
participants may have been motivated more by a desire to get out of traffic ( as indicated
by their least favorite aspect of their current transport mode) as opposed to a desire to
reduce vehicle hassle.
CarLink travel effects follow:
• Commute Mode Change: In CarLink II, solo driving was reduced by 23 percent on
average for all members. Similarly, promising modal shifts were obtained for CarLink I
( over 43 percent reduction in drive alone for the commute travel). CarLink II showed a
slight reduction in carpooling. In CarLink I, carpooling increased by nearly five percent,
but this is likely because of carpooling requirements built into the program.
• Commute Vehicle Miles Traveled ( VMT) Change: In CarLink II, since the majority of
Homebased Users were previously Caltrain riders, this resulted in a slight net VMT
increase of 1.2 miles per day per person for this user group. However, the Workbased
group ( both Workbased Commuters and Day Users) reported a significant decrease of
27.2 VMT per day per person. Similarly, the CarLink I study found that the average
reduction in daily commute travel was 18.5 miles as a result of CarLink I participation.
• Non- Commute Travel Changes: Almost half of CarLink II participants indicated that
their transit use for non- commute trips did not change, while approximately 24 percent
stated that it increased, and over 17 percent responded that it greatly increased.
Approximately 27 percent of members reported a decrease in household vehicle usage, as
they relied less on their personal vehicles.
• Sold or Postponed Vehicle Purchase: Over half ( approximately 52 percent) of CarLink II
respondents reported no change in personal vehicle use after they joined CarLink. Eleven
percent of Homebased Users and five percent of Workbased Users ( Workbased
Commuters and Day Users) sold a personal vehicle or put it in storage. No one purchased
a personal vehicle, although over 51 percent said they would buy one in the next year in
the initial questionnaire ( i. e., at the time they joined CarLink II).
• Carpooling Effects: The average number of Workbased Commuters sharing a CarLink II
vehicle, including drivers, during commutes between the train station and their work sites
was 1.48 in both mornings and evenings.
• Parking Impacts: The overall parking benefit to employers resulted in one parking space
serving two CarLink II vehicles on average.
xviii
Chapter Four: Applying Integrated ITS Technologies to Carsharing and System
Management: A CarLink Case Study
This paper focuses on the role of technology in carsharing system management, lessons learned
from CarLink, and technology benefits to this nascent market. It was presented at the 10th World
Congress on Intelligent Transportation Systems hosted in Madrid, Spain in November 2003.
CarLink II technology findings include:
• Technology: Technology was a major aspect of CarLink II operations since it facilitated
user convenience, management tools, and program expansion. CarLink II technology
included: an in- vehicle navigation system for trip routing, vehicle access for all users,
refueling cards for maximum flexibility, and a reservation system for Day Use.
• In- Vehicle Navigation: The in- vehicle navigation system allowed users to route their trips
and receive visual and voice instruction. This was not a program requirement but rather
an additional feature that provided convenience for some trips. While 13 percent never
used the system, over 50 percent reported the system was very satisfying or satisfying to
use.
• Vehicle Access: Vehicle access is defined as unlocking the car with a key fob and
logging into the CarLink II computerized system with a personal identification number
( PIN), which released the ignition immobilizer and attributed trip activity to the user’s ID
number. Ninety- two percent of users were satisfied with vehicle access at the program’s
mid point. By the program’s end, only 60 percent were satisfied or very satisfied, and
nearly 20 percent were dissatisfied with the system. Homebased Users were the most
frustrated by the length of time ( three seconds) the fob took to unlock the vehicle, and
they felt that the location of the smart key reader ( rear windshield) was inconvenient if
holding a child, groceries, etc.
• Refueling: CarLink II vehicles each included a fuel card and a PIN associated with all
user. This system allowed individuals to refuel the cars at their convenience at local
stations. At the end of the program, 60 percent of respondents reported that they were
very satisfied or satisfied with refueling, and only seven percent were dissatisfied or very
dissatisfied. Throughout the program, participants indicated that the vehicles were
sufficiently fueled, although this was not always the case. Users also indicated that
incentives for individuals who frequently refueled the vehicles ( e. g., coupons for free
coffee, videos, etc.) would have provided more motivation for refueling consistently.
• Reservation System: At the end of the program, 44 percent of the respondents were
satisfied, and only eight percent were dissatisfied with the reservation system. However,
during interim program interviews, 28 percent were dissatisfied with the system. This
change likely reflects satisfaction with reservation system improvements made during the
program’s remainder. The primary reason for reservation system dissatisfaction was the
lack of a lockout system— guaranteeing that a reserved vehicle would be waiting for the
individual that requested it. Vehicle lockout was identified as an area for next generation
xix
technology development, as it was not addressed during the CarLink II pilot program due
to cost and time constraints.
• Integrated Carsharing Technology: As part of CarLink II, American Honda Motor
Company developed an integrated carsharing system that included: 1) vehicle access
( smart key fobs); 2) a reservation system ( Internet- based website); and 3) vehicle use and
tracking ( car location, vehicle miles traveled, fuel levels, user ID number, and time).
CarLink II also included a navigational system.
While the majority of participants were satisfied with the CarLink II technology, the
following improvements were recommended:
o A “ lockout” feature for reserved vehicles should be developed;
o The key fob door- release speed should be increased;
o The PIN entry screen process should be improved;
o The vehicle immobilizer should be integrated with the engine control unit to make
this feature much more secure;
o The online reservation page should be modified to improve scrolling and reflect the
correct time;
o The number of steps involved in making an online reservation should be reduced;
o A means to directly inform the reservation system that a trip is extending past the
reserved time period should be developed ( e. g., automated phone interface); and
o Reserved cars that are unused should be converted to “ available for use”
automatically on the reservation page after a 10 to 15 minute waiting period.
( Furthermore, users should be fined if they do not cancel a reservation in advance.)
Chapter Five: CarLink— A Commuter- Carsharing Model: Conditions for Economic Viability
This chapter provides an analysis of CarLink’s economic potential, drawing from data and
experience from CarLink I and II, with an emphasis on CarLink II. In addition to presenting an
economic analysis of the CarLink II experience, this paper explores the economic viability of
commuter carsharing under three different market scenarios.
Due to underutilization of fleet vehicles and sub- market pricing during the preliminary
implementation, CarLink II was unprofitable for its first year, recording costs of $ 342,002 and
revenues of $ 70,850. Research expenditures also account for some of the high costs. To obtain
meaningful results, the pilot project employed advanced data collection technologies that were
substantially more expensive than a commercial operator would require for basic carsharing
operations.
The three scenarios explore modifications to the CarLink model in terms of market demand,
costs, and revenues.
• Scenario One: Optimal Market Conditions. Scenario One assumes a high demand for
carsharing services. Modifications to the CarLink II figures are as follows: insurance
costs were increased to reflect market rates; in- vehicle technology costs were reduced to
xx
reflect off- the- shelf technologies likely to be employed by commercial carsharing
operators; marketing costs were reduced to reflect an emphasis on the most cost- effective
recruitment methods; monthly vehicle fuel costs were doubled to account for increased
use of vehicles; and salaries were slightly reduced to reflect operations of a typical
carsharing organization. All other costs remain the same as in CarLink II. The revenue
structure was modified to include short- term rentals on weekends to maximize use of idle
vehicles. Homebased User fees remained $ 300, and business customer fees were
increased to $ 960 per month to reflect willingness- to- pay. This scenario assumes 21
vehicles in Year 1 increasing by five in Year 2 and by another four in Year 3. Carsharing
becomes profitable in Year 2 and yields a profit of $ 98,614 by Year 3.
• Scenario Two: Sub- Optimal Market Conditions. Scenario two assumes similar
regional economic conditions, but reduced demand for the carsharing service, potentially
explained by economic, land- use, or demographic composition of a location. The second
scenario starts with just 17 vehicles to reflect slightly reduced demand, increasing by only
two each year. All costs and revenue patterns, adjusted for the smaller number of fleet
vehicles, remain as described in Scenario One. Economic viability is not achieved. The
major shortcoming appears to be a limited demand for commuter carsharing, which does
not allow the organization to achieve an economically viable size.
• Scenario Three: Revenue- Risk Sharing Under Sub- Optimal Market Conditions.
Scenario Three begins with market conditions and fleet growth rates similar to Scenario
Two; however, the carsharing operator shares the risk of losses with the transit operator.
Since transit will ultimately benefit from increased ridership as a result of carsharing, a
transit operator might reasonably assume some of the risks of starting up a carsharing
service. In this scenario, the transit operator purchases and maintains the reserve vehicles,
whereby reducing initial capital costs for the carsharing operator. Should carsharing
prove profitable, the transit operator receives a share of revenues. Under this scenario,
profitability is achieved in years two and three.
Scenario Three provides the best option for long- term viability because it is profitable even
under sub- optimal market conditions. The CarLink Scenarios One, Two, and Three provide
useful lessons for future carsharing programs:
Raise User Fees and Implement Short- Term Rentals: User fees are an important revenue stream.
Based on the high costs of vehicle ownership ($ 500-$ 550 monthly) and the services provided by
CarLink II ( e. g., includes insurance, fuel, and cleaning), user fees could be raised to $ 400 for
Homebased Users and to $ 1000 for business customers based on demand. A combination of flat
fees and usage- based fees for short- duration rentals can ensure a predictable income stream and
provide attractive user fees for short trip needs. Short- term weekend rentals can maximize use of
idle vehicles and increase profits.
Reduce Insurance Costs Through Innovative Techniques: To curb costs, a number of measures
can reduce insurance premiums. Driver screening can yield reduction in premiums given good
driving records and age limitations. Certain passenger safety- technologies can earn insurance
xxi
discounts; high deductibles can reduce monthly premiums; and non- profit insurers provide
reasonable rates to non- profit carsharing organizations.
Control Costs by Employing Standard Technology and Capitalize on Economies of Scale: To
achieve meaningful results, CarLink II employed an advanced technological system that proved
much more costly than that which would be required for a typical car- sharing organization. By
employing off- the- shelf technologies, substantial savings can be realized. With the addition of
new vehicles, economies of scale begin to materialize— all three scenarios support this assertion.
Use Most Cost- Effective and Proven Marketing Strategies: CarLink II focus groups and surveys
yielded that certain marketing techniques were substantially more effective than others in
recruiting participants. By targeting funds to these cost- effective strategies, commercial operators
can reduce marketing costs.
Employ Public- Private Partnerships to Share Revenues and Risks: The revenue- risk sharing
concept assumes that a carsharing service benefits third parties ( e. g., transit operators in the case
of a commuter carsharing program), and thus the third party might be willing to assume a share
of the program risk in exchange for a share of program profits. In Scenario Three, the transit
operator purchases and maintains the reserve vehicles, thus reducing initial capital costs for the
carsharing operator, and receives a share of profits as an incentive. The scenario analysis
( Scenario Three) indicates that the revenue- risk sharing could be feasible in practice.
CarLink II, which linked transit and employers with shared- use vehicles, could provide an
economically viable demand- responsive mobility option under specific conditions, provided
economic success factors gleaned from these experiences are incorporated into future carsharing
models.
Appendix I: CarLink II Survey Instruments
Appendix II: CarLink II Diary
Appendix III: CarLink II Interim Program Interviews
Appendix IV: CarLink II Final Focus Groups Summary
1
CHAPTER ONE
U. S. SHARED- USE VEHICLE SURVEY FINDINGS: OPPORTUNITIES AND
OBSTACLES FOR CARSHARING & STATION CAR GROWTH
Susan A. Shaheen, MollyAnne Meyn, and Kamill Wipyewski
Susan A. Shaheen, Ph. D.
Policy & Behavioral Research, Program Leader, California PATH &
Institute of Transportation Studies, University of California, Davis ( ITS- Davis)
California Partners for Advanced Transit & Highways ( PATH)
1357 S. 46th Street, Bldg. 452
Richmond, CA 94804- 4648
510- 231- 9409 ( O); 510- 231- 9565 ( F)
sashaheen@ path. berkeley. edu; sashaheen@ ucdavis. edu
ABSTRACT
Shared- use vehicle services provide members access to a fleet of vehicles for use throughout the
day, without the hassles and costs of individual auto ownership. From June 2001 to July 2002,
the authors surveyed 17 U. S. shared- use vehicle service organizations on a range of topics,
including organizational size, strategic partnerships, pricing strategies, insurance costs, and
technology applications. While survey findings demonstrate a decline in the number of
organizational starts in the last year, the rate of operational launches into new cities,
membership, and fleet size continue to increase. Several growth- oriented organizations in the
U. S. are responsible for the majority of this expansion. The authors explore several factors that
challenge shared- use vehicle growth, such as high capital investment ( or start- up costs), dramatic
hikes in insurance rates, and scarcity of cost- effective technologies.
The authors conclude that while early niche market findings are encouraging, the ability of this
emerging sector to actualize its total environmental, economic, and social goals may be limited
without the collective support of private industry ( e. g., automakers, insurance providers,
technology producers), public agents ( e. g., transit and governmental agencies), and shared- use
vehicle programs. Indeed, public- private partnerships and cooperation among shared- use vehicle
providers may play a key role in addressing insurance and technology costs and assuring the
long- term viability of this market.
Key Words: Shared- Use Vehicles, Shared- Vehicle Organizations, Carsharing, Station Cars,
Market Developments, Insurance, and Technology
MollyAnne Meyn
Assistant Research Specialist
California PATH
1357 S. 46th Street, Bldg. 452
Richmond, CA 94804- 4648
510- 231- 5707 ( O); 510- 231- 9565 ( F)
Kamill Wipyewski
Assistant Researcher
California PATH
2105 Bancroft Way
1357 S. 46th Street, Bldg. 452
Richmond, CA 94804- 4648
510- 231- 5707 ( O); 510- 231- 9565 ( F)
kamillw@ path. berkeley. edu
2
SECTION 1.0 INTRODUCTION
Travel choices are often limited by lack of connectivity among travel modes, such as transit and
highways. Shared- use vehicles ( linked to key activity locations and transit) can help to expand
the mobility options of individuals who use transit, walk, or cycle, but still require access to a
personal vehicle for a trip segment. The principle of shared- use vehicles is simple: Individuals
gain the benefits of private car use without the costs and responsibilities of ownership. Instead of
owning one or more cars, a household or business accesses a fleet of shared- use vehicles on an
as- needed basis. Members typically provide a deposit or monthly fee for access to a vehicle fleet
distributed throughout a region or concentrated at a transit station, activity center, or worksite.
Members also typically pay an hourly and per mileage fee that reflects their vehicle use.
Insurance, maintenance, vehicle repair, and reservations are included in the fixed cost of the
service ( 1, 2).
Many U. S. programs operate similarly to the majority of European carsharing organizations:
Individuals access cars from nearby neighborhood lots and return them to the same lot ( not
typically linked to transit). This European approach, which started as a grassroots, cooperative
effort in Switzerland, represents “ classical” carsharing ( 1). In contrast, “ station cars” are rooted
in the U. S. and serve transit/ rail commuters primarily, often using electric vehicles. Station cars
typically provide a demand- responsive extension to fixed- route rail services and may not be
shared by multiple individuals ( 1, 2, 3). Increasingly, the carsharing and station cars concepts are
“ merging” to include both elements: transit linkages that serve commuters and distributed lots
for spontaneous users ( 4, 5). One can envision a shared- use vehicle continuum, ranging from
carsharing to station cars, in which several new models fall in between these classifications. ( For
an in- depth discussion of this classification system, see ( 4).)
Common goals among shared- use vehicle organizations ( as reported in the authors’ 2001- 2002
survey) include:
• Facilitating more efficient land use ( e. g., shared- use vehicles reduce the number of
parking spaces needed);
• Providing cost savings since customers pay per use, sharing vehicle leasing costs,
maintenance, repair, and insurance;
• Increasing mobility options and connectivity among transportation modes; and
• Reducing pollution, if the vehicle links to an alternative travel mode— e. g., commuters
using transit can augment their travel with a shared car— or the fleet consists of “ clean
fuel” vehicles.
This paper provides an overview of shared- use vehicle system growth and market developments
from 1994 to present. From June 2001 to July 2002, the authors conducted a longitudinal survey
of all 13 operational carsharing programs and 4 station car programs to monitor trends and
developments. Operational organizations were surveyed through a combination of e- mail
questionnaires and telephone interviews three times throughout the year. Researchers also
updated data from each organization’s website, if available, and from press releases and news
articles. Researchers designed the initial questionnaire to collect baseline data on a range of
issues— organizational size, business costs and financing, operational model, technology
3
applications, marketing methods, and unexpected costs. A database was created to monitor
changes reported in subsequent surveys. The second questionnaire was implemented as a phone
interview in March 2002 and focused on new developments and membership. During this phase,
several organizations were added and ended, resulting in 17 total programs. The predominant
reason for termination was insurance rate increases following 9/ 11. The final survey, conducted
from June to July 2002, focused on membership and cost concerns, primarily insurance and
technology. In addition, researchers interviewed insurance providers and technology companies
to gain a better understanding of these issues and possible solutions. While a dozen planned
shared- vehicle efforts were identified throughout this study, only six developed detailed business
plans. Of these organizations, two were not available for comment, and four were surveyed as
part of the final survey.
This paper includes two main sections. The first is an overview of market growth for shared- use
vehicle programs since 1994, in which the authors discuss the emergence of more growth-oriented
organizations, total membership, and vehicle trends. In the second section, the authors
discuss several challenges facing organizations and explore opportunities to overcome them.
Finally, the authors provide a summary of key observations and conclusions following this
survey.
SECTION 1.1 U. S. MARKET DEVELOPMENTS OF SHARED- USE VEHICLE
SERVICES
Shared- use vehicle services were largely popularized in Europe in the late- 1980s. In its nascent
stages, shared- vehicle organizations in the U. S. have sought European guidance. Prior to the
1980s, European carsharing was experimental and limited to small, localized organizations. Two
influential carsharing organizations were formed in the late- 1980s— StattAuto in Berlin and the
Swiss program now called “ Mobility CarSharing Switzerland.” Mobility CarSharing Switzerland
has continued to grow— now claiming over 50,000 members— while StattAuto’s growth began
to stagnate in the mid- to late- 1990s ( 1). Switzerland’s success has been credited to a more
business- oriented approach, which has been emulated by many organizations. In North America,
carsharing was first successfully established in Canada in 1994 by a cooperative, which later
adopted a commercial model ( CommunAuto). By 1998, four non- profit organizations had
emerged in the U. S. In the mid- 1990s, U. S. rail transit operators, seeking to relieve parking
shortages at stations, also launched several electric station car programs ( 1). This section focuses
on U. S. shared- vehicle program growth.
Since early U. S. developments, the number of shared- use vehicle organizations has grown to 17:
11 carsharing organizations; four station car programs ( two are located in California; the others
are in New York and New Jersey); and two research pilots ( both in California). Approximately
11 U. S. carsharing deployments are planned. ( See Figure 1.1 below.) As of July 2002, station car
programs claimed approximately 148 members and 109 vehicles, while carsharing programs
collectively claimed approximately 12,195 members and operated 471 vehicles. Not surprisingly,
the majority of carsharing members ( 80 percent) live in the 25 most densely populated cities of
the nation.
4
While survey findings demonstrate a decline in the number of organizational starts between June
2001 to July 2002 ( see Figure 1.1 below), the rate of operational launches into new cities ( i. e.,
existing organizations replicate and enter new regions), total membership, and fleet size ( Figure
2 below) continue to increase. This indicates several possible trends:
1) An unmet demand for short- term vehicles that supplement existing transportation
networks;
2) Emergence of more growth- oriented organizations that can quickly enter new cities; and
3) A change in market forces ( e. g., initial fixed costs, such as vehicle leasing, and insurance
have increased or start- up grant funding has declined or both).
FIGURE 1.1: Growth in Station Car and Carsharing Organizations
As mentioned earlier, a few organizations serve the majority of U. S shared- vehicle program
members in multiple regions. For instance, City CarShare, Flexcar, and Zipcar each operate in
several cities. Collectively these organizations serve 92 percent of all U. S. members and deploy
78 percent of the vehicle fleet.
Almost exponential growth in U. S. shared- vehicle memberships further demonstrates that
existing organizations are developing effective strategies to attract a growing membership base.
For instance, carsharing membership grew by approximately 210 percentage points between
1998- 1999; 97 percentage points between 1999- 2000; 1,174 percentage points the following
year; and 127 percentage points to date ( between 2001 and July 2002). ( See Figure 1.2.) High
growth rates between 2000- 2001 are due to two organizational launches characterized by large
start- up funding, capital investments ( i. e., advanced technologies), and rapid growth rates. Total
carsharing fleet size reveals a similar trend ( see Figure 1.2).
0
2
4
6
8
10
12
14
1998 1999 2000 2001 2002
Carsharing Station Cars
5
FIGURE 1.2: Growth in Carsharing Membership and Vehicles
Shared- use vehicle programs continue to grow. However, just a few are responsible for most of
this expansion. Several growth- oriented organizations have emerged that are pioneering new
market segments and diversified rate structures, partnerships with the public and private sectors,
and advanced technology applications ( 6). However, such market developments may be
hindered. The majority of survey respondents reported increased insurance costs and scarcity of
cost- effective shared- vehicle technologies as their two greatest challenges. In the next section,
the authors describe these obstacles and explore possible solutions.
SECTION 1.2 SHARED- USE VEHICLE MARKET GROWTH: OBSTACLES AND
OPPORTUNITIES
The reduced number of organizational launches amidst continuing market, membership, and fleet
expansion indicates that entry barriers likely exist and could be increasing. The high fixed costs
of vehicle lease/ purchase, technology development, and insurance also are significant deterrents
to a new organization’s market entry. Indeed, in a survey of planned organizations 75 percent of
respondents ( n= 7) ranked insurance and smart technologies among their top three costs. While
insurance was not listed as a major cost or concern in the 2001 survey, by July 2002 researchers
found increased insurance costs a major challenge to expansion and sustainability for the vast
majority of U. S. organizations. This section includes a discussion of two main challenges ( and
opportunities for addressing them): access to affordable insurance and technology.
1.2.1 Insurance Coverage: A Brief History, Current Issues, and Potential Solutions
Since late- 2001, U. S. shared- vehicle services have found it more challenging to obtain affordable
insurance coverage. The authors interviewed existing and planned organizations to assess the
severity of sharp increases on business development. Since station car and shared- vehicle
0
2000
4000
6000
8000
10000
12000
14000
1998 1999 2000 2001 2002
Members Vehicles
560
0
80
160
240
320
400
480
Member
s
Vehicles
6
research programs are more insulated from these difficulties— as their insurance policies are
typically supported by agency relationships or partnerships with automakers and rental car
companies— this investigation was primarily focused on U. S. carsharing.
Researchers conducted expert interviews with several brokers and underwriters with carsharing
experience, as well as a literature review of insurance industry characteristics, underwriting
methods, and basic definitions. This discussion focuses largely on the roles of brokers and
underwriters. Underwriters determine insurance classifications and corresponding premium
prices. Brokers act as intermediaries between shared- vehicle organizations and underwriters. All
disclosed past and present carsharing insurers were contacted ( n= 4); another five were
undisclosed and not contacted. Research was partially hindered by the complexities of the
insurance industry ( e. g., a diverse range of approaches) and confidentiality of insurer- client
relationships. Based on these interviews, researchers identified several strategies that could lead
to lower insurance rates and attract additional providers. This section includes a discussion of
shared- vehicle insurance history, recent changes, and possible strategies.
1.2.1.1 Past Challenges to Recent Market Shifts
Early on, organizations experienced difficulty procuring shared- use vehicle insurance. Most
firms were only willing to provide minimum insurance and required members to provide
additional coverage— a scheme modeled after rental car insurance policies. Eventually, interest
in carsharing by VPSI ( a vanpool fleet provider) led to a contact at “ Insurance One” ( now called
HRH of Metropolitan Washington), a brokerage primarily serving vanpool fleets ( Dave Brook,
unpublished data). Via HRH, many U. S. organizations have acquired coverage. In addition,
several smaller organizations have obtained insurance through relationships with local providers,
who— according to policyholders— do not have an interest in writing new policies for shared-vehicle
services. Currently, there are approximately seven insurance companies actively carrying
carsharing policies.
Between 1998 and 2001, insurance premiums were within a range of $ 1,200 to $ 2,100 per
vehicle/ year. However, rates increased dramatically in 2001 across the entire insurance industry,
resulting in significant changes for the shared- vehicle market. For 2001, the insurance industry
reported a negative 2.7 percent rate of return— the worst in the insurance sector’s history. The
following factors have each contributed to the insurance industry’s tremendous losses: 1)
September 11, 2001 terrorist attacks; 2) high catastrophe losses; 3) rising medical costs; 4) high
litigation expenses; 5) the Enron debacle; 6) underpricing of insurance services during a soft
market in the late- 1990s; 7) economic recession; and 8) falling financial markets ( 7, 8).
Furthermore, several factors in particular contributed to rising automobile insurance costs. These
factors include: 1) medical cost inflation; 2) higher jury awards in auto liability cases; 3)
increased vehicle repair costs; and, 4) severe fraud problems in several states, such as New York,
Florida, and Massachusetts. Furthermore, a recent Georgia Supreme Court decision mandated
that insurance companies must compensate car owners for the diminished value of automobiles
involved in accidents. This applies even if the vehicle is repaired and fully functional ( 9). In light
of these market forces, insurance companies have become more conservative in selecting
markets to insure. Insurers are less likely to take on newer, undefined risks. Of the nine U. S.
shared- vehicle insurers, three have terminated coverage. Many insurers are simply unwilling to
7
insure carsharing at present. Indeed, one broker reported rejections by 250 companies in the last
four years ( Michael Boylen, unpublished data); another contacted nearly 20 insurance companies
before terminating a recent search.
1.2.1.2 Current Status and Market Barriers
During 2001- 2002, most shared- vehicle organizations reported a 50 percentage point increase
and higher in renewal rates. One organization even terminated operations due to a 500-
percentage point increase in premiums. As of July 2002, carsharing organizations reported
premiums ranging between $ 1,200 and $ 6,000 per vehicle/ year, reflecting a one million dollar
liability limit and $ 500 to $ 1,000 deductibles. ( Organizations on the lower end of this spectrum
have unique arrangements that are not available to others.) The majority of organizations pay
between $ 4,800 and $ 6,000 per vehicle/ year; this accounts for 20 to 48 percent of an
organization’s total fixed costs. On the high end, carsharing organizations reported
approximately 1.7 collision claims per 10 vehicles/ year— resulting in $ 600 to $ 900/ claim in
costs. No liability claims have been reported to date. Costs are also reflected in total staff hours
dedicated to obtaining affordable insurance coverage. A few organizations are committing up to
25 percent of high- level staff time to this problem.
At present, shared- vehicle services have not yet been assigned a risk class within the insurance
industry. Classification is the process of assigning a proposed party to a group or class of
“ insureds,” with approximately the same expected loss probabilities ( 10). To develop a premium
for a new class of insureds, an underwriter relies on credible historical data to characterize risks
across time and factors. Credible data require a large sample size over at least three years.
Significant data are generated from 10 years of experience and several millions in premiums
( Michael Boylen, unpublished data). When historical data do not exist, the underwriter can use
expert judgment to aggregate similar risks and infer applicable ones for a new class ( 11).
There are several disadvantages associated with an unclassified insurance status. First, policies
will vary widely among carriers, who interpret shared- vehicle risks differently, making it
difficult for carsharing organizations to predict their premium costs ( i. e., there is no standard).
Second, in the current market, insurers are less likely to explore new markets, so shared- vehicle
organizations have fewer options ( and less consumer power due to decreased competition).
Finally, premiums are raised to cover unknown risks and the expense of developing a new
classification category.
1.2.1.3 Possible Strategies: Lowering Insurance Premiums for Shared- Use Vehicles
Based on the authors’ examination, three strategies are recommended for addressing recent
insurance hikes experienced throughout the shared- use industry during the past year. They
include: 1) developing risk rating factors and actuarial tables for U. S. shared- vehicles, 2)
applying advanced technology applications, 3) pursuing captive self- insurance strategy, and 4)
investigating usage- based insurance approaches. Each is discussed below.
1.2.1.3.1 Risk- Rating Factors and Actuarial Tables Risk- rating factors provide a means for
appropriating and measuring driver and vehicle risk. Thus, they are important to carsharing
8
providers in calculating risks and corresponding premiums. Actuarial tables summarize risk-rating
factors and assign pricing structures for each. Determining accurate rating factors requires
the expertise of an actuarial analyst, who applies vehicle and driver statistics in assessing risks.
Developing a new class can be time consuming and costly for insurers.
As a first step in lowering premiums, risk- rating factors and actuarial tables must be developed to
help carsharing organizations minimize and manage risk exposure. To accomplish this, the
authors recommend:
1) Shared- vehicle organizations collaborate in helping underwriters establish standard risk-rating
factors through documentation of accurate data on accident history, vehicle and
driver profiles, fleet usage patterns, and preventive actions ( e. g., theft prevention
devices); and
2) Public assistance opportunities should be explored to establish carsharing risk- rating
standards.
Based on the authors’ review of the literature and interviews, the following rating factors
emerged as the most significant in affecting shared- use vehicle premiums ( See Table 1.1 below).
TABLE 1.1: Rating Factors for Shared- Use Vehicles
FACTOR DESCRIPTION & RESEARCH FINDINGS
Unauthorized
Drivers
This category was the most common concern among brokers and
underwriters. Advanced technologies present a strategy for addressing
this concern and minimizing risk exposure.
High
Member/ Vehicle
Ratio
The insurance company takes on the risk of covering multiple drivers on
one vehicle policy. Underwriters suggested that screening and
continuous monitoring of drivers slightly lowers risk.
Geographical
Location of
Fleet/ Vehicle
Underwriters perceive that shared- vehicle services are offered in
metropolitan areas primarily, where risk exposure is generally higher.
Underwriters were unclear regarding geographical driving restrictions.
One broker recommended that shared- use vehicles be limited to a 50-
mile radius of respective lots to reduce risks.
Fleet Scale The “ law of large numbers” means that the scale of the insured
contributes to policy profitability, data credibility, and price competition
among insurers. All insurers interviewed suggested that carsharing fleet
scale and potential market growth are key factors in a decision to offer
carsharing insurance. Approximations of ideal fleet size were not
provided; however, analogies were drawn to large car rental companies,
with fleets ranging between 148,000 and 486,138 vehicles ( 12).
Driving Records
and Personal
Driver Profiles
Underwriters were supportive of shared- vehicle efforts to screen drivers
and did not want to be responsible for this duty.
9
Claims History There was a difference in opinion among insurers regarding how much
claims data are need to provide credible evidence for a risk- rating factor
( e. g., between 3 and 10 years). Underwriters for a majority of carsharing
organizations were unaware of any significant claims in 2001- 2002.
Vehicle, Make,
Model, and Age
All of these vehicle attributes influence premiums. Car rental risk rates
are reduced when cars are between 1- 2 years old, as opposed to 3- 5
years, or older.
Usage Statistics The assumption that carsharing vehicles are driven more than average
vehicles contributes to higher premium prices. An underwriter for a U. S.
carsharing insurance provider ( as of July 2002) estimated that the
average carsharing vehicle is driven approximately 18,000 miles/ year.
According to reports, the majority of vehicles are driven between
12,000 and 14,400 miles per year. In comparison, the average operator
of a personal vehicle drives nearly 12,000 miles per year and pays $ 700
in total premiums per car/ year ( 13). Accurate usage statistics would
assist insurers in determining the appropriate premium.
At present, the small scale of the carsharing market deters insurers from investing time and
resources into analyzing appropriate rating factors for a classification. Data should be aggregated
across the carsharing industry to make the model as transparent as possible for insurers in
developing risk- rating factors and actuarial tables. Automated vehicle tracking technology could
also be used to accurately track and report actual vehicle usage for risk assessment.
1.2.1.3.2 Advanced Technologies A number of vehicle security technologies can also be
targeted to lower insurance costs. First, premiums for bodily injury and medical insurance can be
decreased by incorporating automated seatbelts into shared- use vehicles. Discounts can account
for up to 25 percent of total medical liability premiums ( 14). Medical liability coverage accounts
for up to 60 percent of total insurance premiums ( UC PATH shared- use vehicle survey,
unpublished data). Second, several anti- theft devices can earn discounts of up to 35 percent off
comprehensive coverage, including vehicle immobilizers that prohibit unauthorized users, smart
keys that facilitate vehicle access to specified keyholders, and vehicle tracking devices that
increase stolen vehicle recovery rates. While anti- theft devices impose a one- time initial cost,
they can provide increased security and insurance benefits over several years ( 14,15).
Comprehensive coverage premiums constitute between 10 to 15 percent of total insurance costs
in shared- use vehicle programs ( UC PATH shared- use vehicle survey, unpublished data).
Nevertheless, a cost- benefit analysis should be conducted before investing in anti- theft
technologies. Since vehicle type and geographical location are the two major factors used to set
comprehensive premiums, comprehensive insurance costs and the feasibility of anti- theft devices
can differ based on these factors.
Finally, a vehicle tracking system also allows for automated vehicle usage information and can
recognize misuses. Accordingly, insurers could receive automated vehicle- usage data for risk
assessment and quoting purposes. Indirectly, vehicle tracking can also positively influence user
treatment of vehicles. Despite these benefits, most insurers interviewed were unaware of vehicle
10
tracking benefits for carsharing and were unable to provide corresponding premium discounts.
To summarize, advanced technologies provide another opportunity for shared- vehicle
organizations to provide accurate risk- factor data to insurers, which could ultimately lower
insurance rates.
1.2.1.3.3 Captive Self- Insurance Strategy Self- insurance is another strategy to reduce high
carsharing insurance premiums. This strategy can be advantageous for many small- scale
organizations that could not otherwise self- insure and offers an alternative to commercial
policies. With self- insurance, the individual/ organization assumes financial risk directly, instead
of paying an insurance company to cover their risks ( 16). Another form of self- insurance is a
high deductible. Recently, one shared- vehicle organization began insuring their fleet through a
high deductible. In this case, the insurance company insures their liability only. The low
comprehensive and collisions claims history of U. S. shared- vehicle programs also suggests that
they may be good candidates for self- insurance.
Another strategy is to create a “ captive” self- insurance pool. This approach would entail four
steps. First, shared- vehicle organizations would agree to participate and identify a number of
operational standards. Second, organizations would attract private or public ( or both)
investments to create a “ risk retention pool” to cover risks ( approximately $ 4 to $ 6 million/ year,
based on current national fleet size), proportional to their fleet size. ( Investors would be
compensated from the profit pool in a manner negotiated.) Third, an “ association captive” ( a
non- insurer or insurer created/ owned by the group to underwrite collective risks) would be
established, who would administer claims, primarily those covered by premiums. Fourth, a
“ reinsurer” would be identified to cover costs from the risk retention pool fund. In exchange for
this role, the reinsurer would receive a portion of premiums and residuals. Accordingly, risks
would be spread among the pooled funds and reinsurer.
Based on estimates from the principal U. S. shared- vehicle broker and correspondence with
underwriters, the following price structures seem reasonable under this scheme:
• Premiums of $ 2,500 per car/ year. This figure is based on four years of experience in
brokering for shared- vehicle organizations and inference to vanpools and other fleet
applications.
• Approximately 42 percent of premiums collected in this model would cover the following
expenses: specific claim losses, claim settlement costs, payments to agents or brokers,
taxes, administrative costs, and initial acquisition expenses to establish the association
captive.
• Expected losses per year of approximately $ 750.00/ vehicle.
• Seven years is the anticipated timeframe before the premium surplus would sufficiently
support the risk pool ( and private investors are no longer needed) ( Michael Boylen,
unpublished data).
There are several potential benefits of the captive self- insurance approach. First, shared- use
vehicle service providers would have more influence over their risk- rating structure. The
formation of this shared- vehicle pool would facilitate collection of significant historical data
after a period of three to ten years. Pooling would also leverage the scale of respective fleets, so
11
that underwriting is streamlined and more profitable for insurers, resulting in lower premiums.
Furthermore, premiums could be lowered if organizations shared insurance overhead, agent
commissions, and tax costs. Finally, investment income would be generated from premiums paid
to insurance companies on reserves. Also, this approach could empower shared- vehicle providers
in determining the nature of their own risk exposure and refining business models to minimize
risk ( Michael Boylen, unpublished data).
Overall, many organizations reported an interest in self- insurance. However, they expressed
concern about generating a risk fund pool and maintaining a competitive edge, while
cooperating. Furthermore, the formation of a captive self- insurance scheme would require some
level of standardization in business practices. Clearly, such a scheme would benefit from the
guidance of an advisory board comprised of shared- use vehicle practitioners and experts,
insurers, and other public interest representatives.
1.2.1.3.4 Usage- Based Insurance Usage- based insurance is a concept piloted and copyrighted
by Progressive Casualty Insurance Company in July 2000. ( An agreement in January 2002
granted an U. K. auto insurer, Norwich Union, exclusive rights to offer usage- based insurance
( Progressive Public Relations, unpublished data). This system bases auto insurance rates largely
on when, where, and how much a vehicle is driven ( 17). While driving record, vehicle, and
location are still taken into consideration in this model, the customer’s rate varies largely due to
mileage. Progressive’s system, named “ Autograph,” employs a combination of Global
Positioning Systems ( GPS) and cellular technology to track vehicle usage. This concept is one
that resonates well with the shared- use vehicle model: Carsharing typically aims to assess fees
for variable vehicle use, and several large- scale operations have employed automatic vehicle
tracking systems to monitor usage.
A number of obstacles hinder the implementation of usage- based insurance; some of these
obstacles parallel those facing shared- vehicle organizations working to obtain affordable
coverage. From the insurer’s perspective, a shift to a new rating- structure is risky without a
guaranteed large consumer demand. Also, this new rating system requires deployment of low
cost in- vehicle technologies for monitoring usage. Additionally, there are numerous political
sensitivities ( e. g., privacy) associated with the use of on- board monitoring equipment and
mileage as a principal risk- rating factor. If a viable business case is developed for usage- based
insurance, the carsharing market should be considered for piloting and marketing this concept.
In the next section, the authors focus on another shared- vehicle challenge— advanced
technology— and opportunities for lowering access and cost barriers.
1.2.2 Advanced Technology: Accessing Cost- Effective Systems
Since several U. S. shared- use vehicle organizations have experienced rapid growth in
membership and fleet size, this creates a challenge to efficiently managing growth and more
complex operations ( 6). Advanced technologies are key to decreasing administration costs.
Interoperability, supported by some degree of standardization among technologies, can further
improve customer services ( e. g., use of smart cards among providers and transit). Existing
technologies already provide solutions for some shared- use vehicle challenges. However,
12
technological systems tailored to the needs of shared- vehicle providers ( e. g., data collection to
support insurance approaches, such as usage- based fees) could further enhance overall market
growth and operations.
Several organizations have taken an active approach in researching and developing new
technologies. Each system has been developed independently, limiting opportunities for
interoperability among systems, organizations, and transit services. Since development expenses
can easily exceed the budget of a single organization, partnerships among shared- vehicle
programs could leverage resources to attain necessary funding and attract technology
manufacturers. In addition, technology providers could become potential stakeholders in the
shared- use vehicle industry. This section explores challenges to accessing customized
technologies, current applications and benefits, and cooperative strategies for enabling the
development and distribution of tailored, cost- effective technologies.
1.2.2.1 The Challenges to Accessing Customized Technologies
The majority of shared- use vehicle providers surveyed— especially those with aggressive growth
objectives— agreed that advanced technologies are a driving force behind successful operations.
Organizations who did not recognize technology as essential typically had more limited growth
objectives. The majority of shared- vehicle providers reported that advanced technologies could
greatly enhance operations, serving larger and more diverse populations. For example, smart
technologies can facilitate one- way rentals ( i. e., members are not required to return their vehicles
to the same location), increasing market penetration. Furthermore, several providers also noted
the advantages of smart chip technologies in linking their services to other shared- vehicle
operators and transit systems.
Throughout the survey, many respondents expressed a need for more affordable technologies and
lacked sufficient scale to justify system acquisition. Not surprisingly, high technology
development costs have deterred most U. S. organizations from developing systems, with just a
few exceptions. Such developments have occurred independently through large start- up grants or
university- sponsored pilot programs in conjunction with the private sector. Additionally, these
efforts are continuously burdened to update their technology. In the next section, the authors
describe current technology applications and realized benefits.
1.2.2.2 Current Technology Applications and Realized Benefits
ITS technologies are currently used to improve three main areas of operation: 1) vehicles access
( smart cards), 2) reservations, and 3) data collection. Smart cards and key fobs communicate user
information to a centralized database that uses member IDs to track participant activities for
billing and security purposes. Smart card technologies have the potential to streamline
administration and fleet management and link to a variety of transportation services ( e. g.,
carsharing, smart parking, and transit).
Automated reservations via online, voice recognition, or touch- tone telephone systems enable
quick and convenient customer scheduling in addition to decreased labor and administration
costs. Finally, automated vehicle location technologies are used to track shared- vehicle fleets
13
over a cellular communications or radio frequency network. Improved member tracking can be
used for automated billing and increased security ( e. g., vehicle immobilization), hence
streamlining management and customer services.
Figure 1.3 ( below) illustrates the distribution of carsharing providers employing ITS
technologies, ranging from manual operations ( e. g., operator phone services, in- vehicle trip
logs), to partially automated ( e. g., automated reservations via touch- tone telephone or Internet or
both), to advanced operations ( e. g., smart card access, reservations, billing, automated vehicle
location, and cellular/ radio frequency communications). As illustrated, 50 percent of U. S.
organizations have advanced operations; 29 percent provide partially automated services; and 21
percent offer manual services. All four station car programs employ manual operations at
present.
FIGURE 1.3: Technology Levels in U. S. Carsharing Organizations
Although existing technologies can address many challenges of shared- vehicle services,
technological systems tailored to the needs of shared- use programs could further enhance market
growth and operations ( e. g., insurance data collection requirements). City CarShare, for example,
is currently developing software that will be licensed at no cost to other non- profit organizations.
This open source system will allow license- holders to modify and customize software to their
specific needs. Similar to the Linux concept— a free computer operating system available to
download and modify— this approach is based on the idea that collective input will accelerate
innovations at a lower cost. Furthermore, start- up organizations have recently begun leasing
customized technologies and services ( e. g., billing) from larger organizations. Both strategies
mentioned here support cooperation. In the final section, the authors explore benefits of a
cooperative approach to technology development.
1.2.2.3 Cooperation: Leveraging Expertise and Scale to Develop Cost- Effective Technologies
Cooperation among shared- vehicle organizations could be an effective means to decrease
technology development costs. There are essentially two possible approaches to furthering cost-
Advanced Operation Systems
Partly Automated Operations
Manual Operations
46%
23%
31%
14
effective, technological advances in shared- use vehicle systems. First, large, business- oriented
shared- use vehicle providers invest in researching new technologies and applications. When
applicable and feasible technologies are developed, they can be marketed to other shared- vehicle
organizations. Second, several shared- use vehicle organizations combine their research
investments to implement and develop technology for industry- wide operations.
The first approach partly resembles the current state of the industry. Due to high technology
development costs, however, sales to other operations only cover a small portion of initial
expenses. As a result, a developed system must prove extremely viable for a particular
organization. To date, development has been largely dependent on grant availability.
Innovations, even if determined essential, are dependent on subsequent grants or unique
opportunities ( e. g., private sector investment). To evaluate the second cooperative approach, a
more detailed analysis of customized technology systems is discussed below.
Customized technologies that address the specific needs of shared- vehicle programs can provide
a powerful tool in improving customer services and streamlining administration. Comprehensive
systems with on- board computers and vehicle tracking systems can process and transmit data on
vehicle usage and location for administrative and security purposes. Simultaneously, these
systems can offer customer- friendly features including phone access; directions and parking
information ( e. g., through GPS); reminders on low- fuel levels and rental- time limits; and
customer- specific features, such as preferred radio station, seating adjustments, and mapping
information. Furthermore, advanced on- board computers can be adapted to new customer
demands by adding or changing software or hardware. This is crucial when vehicle services
target new market segments and a more diverse customer base.
According to industry expert interviews, costs of a customized system, including on- board
computers and tracking devices, can range between $ 500 ( when several existing components are
incorporated) to $ 4,000 per vehicle for state- of- the- art system, with highly specified software
and hardware components and a high degree of upgrade flexibility. Installation typically does not
exceed two labor hours. Due to costs and long development times ( i. e., typically over six to
twelve months), technology providers are unlikely to design customized systems for small fleets.
When interviewed, technology developers reported that fleets of several hundred to one thousand
vehicles would be economically viable from a manufacturer’s perspective.
As the combined fleet of U. S. shared- use vehicle operators currently totals 580 vehicles,
cooperative efforts would appear beneficial in attracting the technology sector. Large- scale
partnerships among shared- vehicle providers could lead to necessary funding and fleet size to
attract technology manufacturers. This would also encourage continuous development efforts
due to larger, more sustainable industry relationships.
Finally, technology partnerships could lead to a greater degree of technology standardization and
facilitate interoperability among different shared- use vehicle organizations. Improved
interoperability can enhance customer service. For example, four carsharing organizations
located in four different cities in Ontario agreed that the members of a particular organization
should have the option to conveniently access the services of another when traveling. To
facilitate this partnership, the four organizations share the same insurance carrier, which helped
15
them overcome challenges related to insurance coverage. To summarize, this agreement
positively affected the market since overall carsharing vehicle use and customer satisfaction
increased.
SECTION 1.3 CONCLUSION
Despite promising U. S. shared- use vehicle operational and membership growth rates, the relative
small scale of these organizations presents a challenge to: 1) obtaining affordable insurance and,
2) covering other high capital costs, including technology, vehicles, and labor. While advanced
shared- vehicle technologies can help organizations to reduce administrative costs and potentially
lower insurance premiums, technology deployment has typically required large private
investments or public development grants.
Strategic cooperation among shared- use vehicle organizations could address these challenges on
several levels. First, limited cooperation could help organizations collectively address some
insurance issues ( e. g., development of risk- rating factors). Furthermore, more involved
cooperative efforts could support aggressive strategies, such as captive self- insurance.
Second, cooperation among shared- use vehicle organizations could also accelerate the
development of interoperable, customized technologies and continued innovation by creating the
scale necessary to attract technology providers. Additionally, a combined insurance- technology
cooperative strategy may be even more beneficial  lowering insurance premiums, enhancing
customer services and capabilities, and lowering capital costs.
Third, cooperation could also manifest highly desirable social and environmental benefits. The
potential of new and existing shared- use vehicle service organizations to continue expanding and
serving new markets could be greatly enhanced through supportive public- private partnerships.
Policymakers and transit operators, for instance, should continue to explore the social and
environmental benefits of shared- use vehicle services through grant making, preferential parking,
supportive policies ( e. g., high occupancy vehicle ( HOV) lane access), and outreach/ marketing.
Strong public- private partnerships are needed to facilitate the on- going development and
sustainability of viable U. S. shared- use vehicle programs. Thus, it will be important for the
public and private sectors to continue working together to monitor system designs and impacts
and to facilitate and encourage collective partnerships among shared- use vehicle organizations
particularly where tremendous synergies could be realized, such as insurance policies and
customized technologies.
ACKNOWLEDGEMENTS
The authors would like to acknowledge John Wright, Rebecca Pearson, Knute Ayhnes- Johnson,
and Rachel Finson of the University of California PATH for their assistance gathering shared-vehicle
program data. Thanks also go to the numerous shared- use vehicle programs that provided
survey responses, Michael Boylan of Insurance One, and Dave Brook of Flexcar. The authors
would also like to thank the California Department of Transportation, UC PATH, and
DaimlerChrysler for their generous contributions to this research. The contents of this paper
reflect the views of the authors and do not necessarily indicate acceptance by the sponsors.
16
This paper builds on an earlier article prepared for the Ninth World Congress on Intelligent
Transportation Systems to be held in Chicago, Illinois, October 14- 18, 2002.
REFERENCES
1) Shaheen, S. A., D. Sperling, and C. Wagner. Carsharing in Europe and North America: Past
Present and Future. Transportation Quarterly, Vol. 52, No. 3. Summer 1998, pp. 35- 52.
2) Shaheen, S. A. Dynamics in Behavioral Adaptation to a Transportation Innovation: A Case
Study of CarLink— A Smart Carsharing System. UCD- ITS- RR- 99- 16. Institute of
Transportation Studies, University of California, Davis, 1999.
3) Shaheen, S. A. Pooled Cars. Access Magazine. University of California Transportation Center
( UCTC), Berkeley. Number 15, Fall 1999, pp. 20- 25.
4) Barth, M. and S. Shaheen. Shared- Use Vehicle Systems: A Framework for Classifying
Carsharing, Station Cars, and Combined Approaches. Transportation Research Record.
Transportation Research Board, National Research Council, Washington, D. C., 2002, 19
pages.
5) Shaheen, S. A., J. Wright, and D. Sperling. California’s Zero Emission Vehicle
Mandate— Linking Clean Fuel Cars, Carsharing, and Station Car Strategies. Transportation
Research Record. Transportation Research Board, National Research Council, Washington,
D. C., 2002, 26 pages.
6) Shaheen, S. and M. Meyn. Shared- Use Vehicle Services: A Survey of North American
Market Developments. In 9th World Congress on Intelligent Transportation Systems
Conference Proceedings ( CD- ROM). Chicago, Illinois, October 2002, 12 pgs.
7) Hartwig, R. P. Special Report: Groundhog Forecast 2002. Insurance Information Institute
Website. http:// www. iii. org/ media/ industry/ financials/ groundhog2002/ content. print/.
Accessed July 1, 2002.
8) Hartwig, R. P. 2002— First Quarter Results. Insurance Information Institute Website.
http:// www. iii. org/ media/ industry/ financials/ 2002firstquarter/. Accessed July 31st, 2002.
9) Hartwig, R. P. What’s Behind the Rising Cost of Auto and Homeowners Insurance? Outlook
For The Auto and Homeowners Insurance: Second Half 2002 and Preview for 2003.
Insurance Information Institute Website.
http:// www. iii. org/ media/ hottopics/ hot/ 20022003outlook/ content. print/. Accessed July 1,
2002.
10) Teufel, P., T. Tongson, J. Rech. Insurance Risk 101. The Academy of Actuaries, July 9,
2001, pp. 23- 32. Available at http:// www. actuary. org/ briefings/ pdf/ risk101_ handout. pdf.
Accessed August 1, 2002.
17
11) Casualty Committee of the Actuarial Standards Board. Documentation and Disclosure in
Property and Casualty Insurance: Ratemaking, Loss Reserving, and Valuations. Actuarial
Standard of Practice No. 9. Adopted by the Actuarial Standards Board January 1991, pp. 7-
10. Available at http:// www. actuarialstandardsboard. org/ pdf/ asops/ asop9. pdf
12) Auto Rental News. U. S. Car Rental Market Statistics. 2001. http:// www. fleet-central.
com/ arn/ 01stat3. cfm. Accessed July 30, 2002.
13) Insurance Information Institute ( III). What determines the Price of My Auto Policy?
Available at http:// www. iii. org/ individuals/ auto/ b/ whatdetermines/ Accessed July 30, 2002.
14) Galvin, F. Yield! Information About Automobile Insurance Discounts. Citizen Information
Service. http:// www. state. ma. us/ sec/ cis/ cisyld/ yldidx. htm. Accessed July 30, 2002.
15) National Insurance Fraud Bureau Website. http:// www. nicb. org/ pd/ anti_ theft_ devices. pdf.
Accessed July 30, 2002.
16) Insurance Information Institute. Glossary of Terms. http:// www. iii. org/ media/ glossary/.
Accessed July 2002.
17) News Release. Progressive Awarded Second Patent for Usage- Based Auto Insurance Rating
System. July 13, 2000. http:/// www. progressive. com/ newsroom/ 2nd_ patent. asp. Accessed
July 30, 2002.
18
19
CHAPTER TWO
A FRAMEWORK FOR TESTING INNOVATIVE TRANSPORTATION SOLUTIONS:
A CASE STUDY OF CARLINK  A COMMUTER CARSHARING PROGRAM
Susan A. Shaheen and Linda Novick
Susan A. Shaheen, Ph. D.
Policy & Behavioral Research, Program Leader, California PATH &
Institute of Transportation Studies, University of California, Davis ( ITS- Davis)
California Partners for Advanced Transit & Highways ( PATH)
1357 S. 46th Street, Bldg. 452
Richmond, CA 94804- 4648
510- 231- 9409 ( O); 510- 231- 9565 ( F)
sashaheen@ path. berkeley. edu; sashaheen@ ucdavis. edu
Linda Novick
Research Specialist
California Partners for Advanced Transit & Highways ( PATH)
1357 S. 46th Street, Bldg. 452
Richmond, CA 94804- 4648
510- 231- 5602 ( O); 510- 231- 9565 ( F)
lnovick@ path. berkeley. edu
ABSTRACT
Transit accounts for just two percent of total travel in the U. S. One reason for low ridership is
limited access; many individuals either live or work too far from a transit station. In developing
transit connectivity solutions, researchers often employ a range of study instruments, such as
stated- preference surveys, focus groups, and pilot programs. To better understand response to
one innovative transit solution, the authors employed a number of research tools, including: a
longitudinal survey, field test, and pilot program. The innovation examined was a commuter
carsharing model, called CarLink, which linked short- term rental vehicles to transit and
employment centers. Over several years, researchers explored user response to the CarLink
concept, a field operational test ( CarLink I), a pilot program ( CarLink II), and a commercial
operation ( the pilot was turned over to Flexcar in summer 2002). This multi- staged approach
provided an opportunity for researchers to learn and adapt as each phase progressed. In this
paper, the authors outline the CarLink model, technology, and early lessons learned; describe
CarLink II operational understanding; provide a synopsis of the pilot program transition; and
offer recommendations for future model development.
Key Words: Carsharing, CarLink, User Response, Operations, Survey, Field Test, Pilot
Program, Sustainability
20
SECTION 2.0 INTRODUCTION
Although public transportation use is growing in the United States, it still accounts for only two
percent of total travel ( 1). In the San Francisco Bay Area, where there is an extensive public
transportation network, transit use is higher: twelve percent of commuters used public
transportation in 2002 ( 1). Congestion on freeways and surface streets, coupled with continuing
air pollution, requires the examination of more demand- responsive alternatives. According to a
nationwide report conducted in 2000, the San Francisco Bay Area averaged 92 hours of delay per
person per year during peak commute hours ( 2). Not surprisingly, transit access is a major
impediment to use; transit capacity often exceeds the number of people, living or working,
within walking distance  one quarter mile or less  of a station. If existing access methods are
augmented ( ranging from traditional fixed route transit to more demand- responsive solutions),
more individuals could use transit. Increased transit access would assist in reducing congestion
during peak travel periods, while also improving overall system efficiency.
Designing innovative solutions that increase transit access and ridership is challenging. This is
especially true in the context of altering long- term travel behaviors, particularly single
occupancy vehicle use. Furthermore, individuals are reluctant to try unfamiliar ideas, new
technologies, or both. Understanding how to change long- held travel patterns is one of the
greatest challenges faced by transportation professionals.
There are many complex issues associated with testing and implementing transportation
innovations. Significant data about an innovation’s impacts are typically needed to justify large-scale
deployment costs. There are several methods for gathering these data, such as simulation
modeling, stated- preference surveys, and controlled testing. As confirmed by CarLink I, much
can be learned from testing a transportation innovation in a real- world setting ( 3). Field tests and
pilot programs provide a framework for investigating complex relationships among system
efficiency, user acceptance/ impacts, economic viability, and other operational issues.
Usually, field tests operate for a predetermined length of time to evaluate a new concept/
technology. In contrast, pilot programs can extend beyond this initial “ proof- of- concept” phase
by focusing on program sustainability. Whether instituting a new concept, technology, or
regulatory framework, pilot programs can be beneficial to decision makers and participants.
Pilots enable new ideas to be tested, modified, and assessed with limited financial risk and no
ongoing obligation. At the same time, they can support program continuation and offer a cost-effective
alternative to exploring transportation innovations.
From July 1, 2001 to June 30, 2002, a carsharing pilot program, emphasizing transit and
employer access— CarLink II— was deployed in the San Francisco Bay Area. Pilot objectives
included testing an advanced carsharing system, understanding user response to this service, and
testing its long- term sustainability. This paper examines the CarLink technology, participant
response, and lessons learned from this multi- stage initiative. The authors first review the
CarLink model, technologies, and early lessons learned. Second, CarLink II operational findings
are examined. Third, a synopsis of the pilot program transition to a permanent service is
discussed. Finally, the authors conclude with opportunities for improving carsharing deployment
initiatives based on these findings.
21
SECTION 2.1 CARLINK PROGRAM AND RESEARCH OVERVIEW
Between 1998 and 2003, researchers deployed a three- phase carsharing research program in the
San Francisco Bay Area, CarLink, in conjunction with the California Department of
Transportation ( Caltrans), American Honda Motor Company, the Bay Area Rapid Transit
( BART) District, Caltrain, and Lawrence Livermore National Laboratory. During the first phase,
researchers conducted a longitudinal survey that examined CarLink concept response ( for more
information, see ( 4)). During the second phase, researchers assessed CarLink I— a demonstration
that examined user response and operations in a controlled setting. CarLink I was based at the
Dublin- Pleasanton BART station and operated for ten months during 1999 ( 3). In the final phase,
researchers examined the CarLink II pilot program, which ran from July 1, 2001 through June
30, 2002, and was based at the California Avenue Caltrain station in Palo Alto. The research
goals of this pilot project included testing advanced carsharing technologies, overall user
response, and economic sustainability.
Broadly defined, carsharing allows a group of individuals to share a vehicle fleet, paying for use
based on time and miles traveled ( e. g., City CarShare, Flexcar, and Zipcar). The most common
model is known as neighborhood carsharing, where a few vehicles are deployed in each of
several neighborhoods for easy member access. These vehicles are accessed from and returned to
the same lot. CarLink tested a commuter carsharing model that provided vehicle access at home
and work, as well as a transit linkage on either end of a commute. This section includes a brief
overview of the CarLink model, differences between CarLink I and II, and program pricing.
2.1.1 CarLink Model: A Brief Overview
Both CarLink I and II were based on the same commuter carsharing structure, involving three
sets of members: Homebased Users, Workbased Commuters, and Workbased Day Users
( described below). Both CarLink programs included a single, primary transit station that served
as a vehicle transfer point for Workbased Commuters and Homebased Users who commuted via
transit. CarLink provided a convenient transit linkage to and from home/ work via a shared- use
vehicle fleet. This same fleet was also shared by households and employers for tripmaking on
evenings and weekends and throughout the workday.
During CarLink I, Homebased Users would drive their CarLink vehicles to a selected transit
station each morning, park the car in a designated CarLink space, and ride transit to work. Next,
a Workbased Commuter would arrive at the same station via train in the morning, pick up a
CarLink car, and drive it to work, parking in a designated CarLink space at their work location.
Throughout the day, Workbased Day Users could reserve CarLink vehicles for business and
personal errands, returning the cars to a designated work lot after each trip. At the end of the
workday, Workbased Commuters drove the CarLink vehicles back to the transit station and
would take the train for the remainder of their trip home. After Homebased Users  riding the
train for the majority of their commute home  returned to the transit station, they would pick up
a CarLink vehicle and drive it home for personal use on evenings and weekends.
As mentioned above, the CarLink II pilot program is based on the same general model as
CarLink I. However, lessons gleaned from user feedback and recommendations from the
22
CarLink I staff and project partners ( i. e., Honda, Caltrans, BART District, and LLNL) suggested
several changes to improve the model and research focus. Overall, it was decided that more
could be learned by adapting the model to a new setting and attempting to create a permanent
enterprise. This section describes the CarLink II project components and how they differ from
CarLink I. Table 2.1, below, summarizes the major differences between CarLink I and II.
TABLE 2.1: Differences Between CarLink I and II
STUDY
CHARACTERISTICS
CARLINK I CARLINK II
Number of Vehicles 12 Vehicles 19 Vehicles
Primary Transit Partner BART Caltrain
Transit Station Location Dublin/ Pleasanton Palo Alto
Vehicle Type Compressed natural gas Honda
Civics
Ultra- low emission Honda Civics
Homebased Users Up to 10 households, pay $ 200 per
month.
Up to 16 households, pay $ 300 per
month.
Workbased Commuters Up to 20 LLNL employees pay
$ 60 per carpool ($ 30 each).
Up to 63 employees of businesses
at Stanford Research Park
( primarily), share CarLink vehicles
to carpool to/ from work.
Businesses pay $ 350 per month per
vehicle ( a combined fee) for
Workbased Commuter and Day
Use services ( in contrast to
employees paying for this service
independently as in CarLink I).
Workbased Day Users Employees of LLNL pay $ 1.50 per
hour and $. 10 per mile.
Up to 28 employees of Stanford
Research Park companies and
other nearby businesses have
access to vehicles for business and
personal use. Employers pay $ 350
per vehicle per month to subscribe
to the combined Workbased
Commuter and Day Use services.
Total Users 54 107
Employer One: LLNL Six: Several private companies
at/ nearby Stanford Research Park
Technology In- vehicle tracking, smart key
kiosk at transit station, smart cards,
manual key boxes at LLNL, and
on- line scheduling system at
LLNL
In- vehicle tracking, automated data
collection, smart key fob ( or
smartcard) entry, PIN- based
vehicle login, on- line reservations,
and in- vehicle navigation system
23
Program Length Field test designed for limited 10-
month duration
Pilot program with planned
transition to on- going carsharing
service
Research Goals Document demand for commuter
carsharing service and gauge user
satisfaction and needs
Continued analysis of commuter
carsharing ( in a new setting) with
greater statistical confidence ( i. e., a
greater sample size) and new
emphasis on technology testing, its
impact on cost reduction, and
longer- term program sustainability
2.1.2 CarLink Economics
Both CarLink I and II required members ( or their employers) to pay for vehicle use. Lessons
learned from carsharing programs in Japan informed this design decision. There, the program
lost participants when fees were implemented for services initially provided for free ( 5). Thus,
CarLink service fees were required to test the economic value of the service. For members, fees
covered all operational and vehicle maintenance costs, including fuel and insurance.
The fee structure was determined by a literature review, willingness- to- pay studies through focus
groups, discussions with employers, and by estimating operational costs. The fee structure was
below “ market value” for both the demonstration and the pilot program as this was a new
concept and users contributed to the research process. Participants provided feedback on the
program and technology, including completing surveys and participating in focus groups and
personal interviews.
CarLink I and II consisted of three user groups: Homebased Users, Workbased Commuters, and
Day Users. Homebased members paid a monthly fee for car use to commute to and from the
station and on evenings and weekends. CarLink I Homebased Users paid $ 200/ month; CarLink
II Homebased Users paid $ 300/ month. The payment structure for CarLink I and II differed for
the Workbased Commuter and Day Use portions of the model. In CarLink I, employees paid a
flat Workbased Commuter fee ($ 60/ month/ car), as well as usage fees ($ 1.50/ hour and
$ 0.10/ mile) for their personal CarLink vehicle use during the workday. Employers paid for
work- related trips. As part of CarLink II, the model was adapted slightly. Under the new
structure, employers paid a flat fee of $ 350/ month per car, which covered both the Workbased
Commuter and Day Use components. Employers joined CarLink II to provide the carsharing
service as an employee benefit. Potential benefits include: 1) promoting employee retention, 2)
reducing office parking demand, 3) encouraging transit use, and 4) substituting costly fleet
vehicle program operations with CarLink in some cases. Each business had specific, and
different, reasons for joining CarLink II.
24
SECTION 2.2 EARLY LESSONS LEARNED
The CarLink longitudinal survey and CarLink I field test were designed to test the commuter
carsharing concept. Proof of concept was the primary goal of CarLink I. Implemented as a
demonstration, CarLink I ceased operations at the close of the research project in late 1999. In
contrast, CarLink II was a pilot program designed to test integrated carsharing technology and
long- term sustainability. Pilots allow for a more realistic evaluation of user response, since
members understand that the program may become permanent. For instance, a member might
sell a car if she believes the program will continue. This section provides an overview of
CarLink longitudinal survey findings and CarLink I field test results, which informed the design
of CarLink II.
2.2.1 CarLink Longitudinal Survey
From June to October 1998, researchers collected response data on the CarLink concept from
302 individuals ( representing 212 households) in the Bay Area. These attitudinal and belief data
measured change in response, which helped to explain the innovation adoption process. The
survey consisted of a baseline ( or initial survey) and three identical questionnaires that followed
each of the informational media developed to explain the CarLink concept: an informational
brochure; video; and an interactive trial drive clinic with compressed natural gas ( CNG) Honda
Civics, smartcards, and a smart carsharing key management kiosk. An experimental group and a
control group were recruited for the study to evaluate informational media impacts on CarLink
response. Communication objectives emphasized the disadvantages of current modes, the
advantages and disadvantages of carsharing, and how the CarLink system works.
Participating households, for both the longitudinal survey and the CarLink I field test, included
four groups: 1) current BART commuters, 2) individuals who might use BART when carsharing
becomes available, 3) people who do not usually take transit but could take it to work, and 4)
individuals who live in neighborhoods with substantial BART ridership. These groups
represented potential CarLink participants.
The final sample population consisted of 207 experimental participants ( 154 households) and 95
control group participants ( 58 households). A total of 488 individuals ( i. e., both experimental
and control) received the initial questionnaire. Throughout this study, there were 186 dropouts
( 58 did not return the first questionnaire, and 128 individuals dropped out after returning the
second questionnaire). After the survey was completed, four focus groups were held with study
participants in October 1998, to further gauge participant perceptions and overall response to the
CarLink concept. The focus groups consisted of three experimental groups with a total of 28
participants and one control group session with nine participants.
Researchers found that CarLink response was influenced by the amount and type of exposure to
the concept, as predicted by social marketing and learning theories ( for more information on
these theories, see ( 4)). Specifically, participants who only read the CarLink brochure lost
interest over time ( interest dropped from 45 percent at the time of the initial questionnaire to 33
percent during the final questionnaire), while nearly 78 percent of those who read the brochure,
watched the CarLink video, and participated in the drive clinic reported that they would use
25
CarLink as part of the final questionnaire. In fact, many indicated that they would be interested
in joining the CarLink I field test ( i. e., 54 percent of the experimental group in contrast to 33
percent of the control) in the final questionnaire.
At the drive clinic, held in September 1998, participants used a smartcard to access a CarLink
vehicle and released the immobilizer, which blocked unauthorized users from starting the car,
and took a test drive, accompanied by a researcher who documented their observations,
questions, and concerns. The drive clinic offered participants a chance to see and try new
technologies, as well as to interact with study researchers. Each participant completed a 20-
minute exit interview with a researcher on his or her response to the CarLink system and
willingness to participate in such a service.
During the exit interview, over 90 percent of participants said “ Yes.” As a result of the clinic,
there was a 21 percent increase in the “ Yes” response category. Since control group respondents
did not participate in the clinic, there are no corresponding data for them. Thus, it appears that
the drive clinic was an effective tool for increasing positive awareness of the CarLink concept.
Nevertheless, this response appears to be overstated ( i. e., the social desirability effect or
tendency of participants to overstate a socially desirable position, especially in the presence of
researchers), as there was a 13 percent decrease ( from the exit interview) in the experimental
group’s response during the final questionnaire.
The CarLink program built on the longitudinal survey in three ways. First, researchers included
32 longitudinal survey participants in the CarLink I field test ( i. e., 15 percent of the experimental
population). Second, understanding about the value of multiple informational media was
integrated into CarLink recruitment strategies. Finally, a trial offer was added to the CarLink II
program ( i. e., an opportunity to try CarLink for a limited period of time prior to subscribing)
based on the success of the drive clinic.
2.2.2 CarLink I
The CarLink I field test provided an exploratory test bed for this carsharing model. During the
field test, many lessons were learned and success factors identified ( 3). Shortly after the CarLink
longitudinal survey was completed, researchers contacted individuals who indicated that they
would be interested in CarLink I field test participation. Individuals were able to enroll in
CarLink I, if they had a match with one or more of the following field test requirements,
including:
1) Homebased Use, those who could use the Dublin- Pleasanton BART Station to commute
to work;
2) Workbased Commuter Use, individuals who work at Lawrence Livermore National
Laboratory ( LLNL) and could commute via BART; and
3) Day Use, those who work at LLNL.
Researchers were unable to enroll individuals that did not match one of these user groups. Given
the restrictive participation requirements, a majority of interested participants did not meet the
criteria for program participation. Interestingly, no one from the control group joined the field
26
test. Thirty- two individuals or 28 percent of experimental respondents, who requested to be
contacted about field test participation, became members ( or 15 percent of the total experimental
population). These individuals ( i. e., from the longitudinal survey) represent 60 percent of the
field test population. Twenty additional individuals joined the field test ( i. e., not from the
longitudinal survey), primarily in the Homebased User and Workbased Commuter categories.
The field test was deployed in the Dublin- Pleasanton region from January to November 1999. As
part of the CarLink I evaluation, several participant feedback tools were employed, including
questionnaires, household interviews, and focus groups. A high percentage of users agreed to
participate in the study ( i. e., 73 percent response rate). This program enrolled 54 participants
throughout the 10- month field test with 38 active participants. Active participants drove the
vehicles frequently, whereas inactive members did not use the CarLink vehicles ( even though
they enrolled in the program). The participant pool was limited due to the short project duration,
program startup delays, and limited CNG infrastructure ( 3).
The CarLink II pilot program built upon six key operational lessons learned from CarLink I:
1) Streamlining Technology: Several technology shortcomings ( i. e., key management and
vehicle tracking systems) contributed to delays and necessitated program modification.
Technology should be integrated and customized to facilitate carsharing use. A stand-alone
“ smartcard” approach should be developed and tested in which fixed key box lots
are not needed. In this way, participants could access vehicles with smartcards alone.
2) Limited CNG Infrastructure: During CarLink I, two CNG issues constrained operations: a
limited number of CNG refueling sites and slow CNG refueling pumps at LLNL. The
CNG component of CarLink I restricted vehicle range and participation. Also, users did
not refuel vehicles as frequently as agreed. Use of CNG vehicles in the CarLink I field
test distracted from the shared- use vehicle evaluation. In the future, this model should be
tested with internal combustion engine ( ICE) vehicles and fuel cards.
3) Guaranteed Parking: Guaranteed parking at the Dublin- Pleasanton BART station was a
huge program incentive, as parking at this station filled up prior to 7AM at the time of the
program. In the future, carsharing programs should be sited in locations where parking is
costly and limited.
4) Vehicle Cleanliness: During CarLink I, operations staff and participants cleaned and
washed cars. Nevertheless, vehicle cleanliness continued to be a chronic program issue.
Consider hiring a third party to clean vehicles more frequently.
5) Employer Participation: Day Use participation in CarLink I was limited. In the future,
test an employer- focused carsharing service with multiple companies located in a
congested corridor with transit access and parking constraints.
6) Program Duration: CarLink I was a limited demonstration project ( i. e., 10 months),
which restricted understanding of user adoption and behavior because of its short
timeframe. In the future, deploy CarLink as a pilot program with the potential to
transition to an ongoing operation after the research phase ends.
27
SECTION 2.3 CARLINK II USER & OPERATIONAL UNDERSTANDING
During the CarLink I field test, the primary goal was narrowly defined— to study user response
to the commuter carsharing concept. In CarLink II, the research goals were broadened to
evaluate long- term program sustainability and to test an integrated smart carsharing system. The
California Avenue Caltrain station in Palo Alto was selected as the CarLink II transit hub after
evaluating a number of potential locations in the San Francisco Bay Area. The criteria for site
selection were: 1) located near a congested corridor, 2) significant number of commuters
traveling to and from the station, 3) concentration of employers near transit station ( i. e., within
five to ten miles of station), 4) supportive transit operator, 5) limited bus or shuttle services, 6)
transit parking at capacity, and 7) local governmental project support.
All sites evaluated for CarLink II had freeway congestion in both directions and commuters
traveling to and from the transit hub. Other locations evaluated included Santa Clara/ San Mateo
Counties, San Jose, and the Dublin/ Pleasanton area ( location of CarLink I). Based on the above
criteria Palo Alto was selected as the preferred location.
The following section includes an overview of CarLink II user satisfaction and operational
lessons learned.
2.3.1 CarLink II User Satisfaction
A total of 107 individuals participated in the CarLink II program: 16 Homebased Users, 28 Day
Users, and 63 Workbased Commuters/ Day Users. Fifty- three percent of participants were female
and 47 percent male. Sixty- four respondents completed the final questionnaire ( a response rate of
60 percent). Respondents included nine Homebased users ( five male, four female), 21 Day Users
( 9 male, 12 female), and 34 Workbased Commuters ( 14 male, 20 female).
Technology was a major aspect of CarLink II operations since it facilitated user convenience,
management tools, and program expansion. The CarLink II technology included: an in- vehicle
navigation system for trip routing, refueling cards for maximum flexibility, and a reservation
system for Day Use. Figure 1, below, provides user satisfaction data on four key program areas:
1) in- vehicle navigation, 2) vehicle access, 3) refueling, and 4) reservations.
28
FIGURE 2.1: Satisfaction with CarLink II Features
2.3.1.1 In- Vehicle Navigation System
The in- vehicle navigation system allowed users to route their trips and receive visual and voice
instruction. This was not a program requirement, but an additional feature that provided
convenience for some trips. Many users did not use it regularly, since their trips from the train to
home or work were identical each day. While 13 percent never used the system, over 50 percent
of respondents reported that the system was very satisfying or satisfying to use. It is interesting
to note that system use increased during the second half of the pilot program, particularly among
Homebased Users.
2.3.1.2 Vehicle Access
Vehicle access is defined as unlocking the car with a key fob and logging into the CarLink II
computerized system with a personal identification number ( PIN), which released the ignition
immobilizer and attributed trip activity to the user’s ID number. Ninety- two percent of users
were satisfied with vehicle access at the program’s mid point. By the program’s end, only 60
percent were satisfied or very satisfied, and nearly 20 percent were dissatisfied with the system.
Homebased Users were the most frustrated by the length of time ( three seconds) the fob took to
unlock the vehicle, and they felt that the location of the smart key reader ( rear windshield) was
inconvenient if holding a child, groceries, etc.
2.3.1.3 Refueling
CarLink II vehicles each included a fuel card and a PIN associated with each user. This system
allowed individuals to refuel the cars at their convenience at local stations. Members were
required to refuel a vehicle if the fuel level fell below 1/ 4 tank or a $ 10 fine was imposed. At the
end of the program, 60 percent of respondents reported that they were very satisfied or satisfied
with refueling, and only seven percent were dissatisfied or very dissatisfied. Throughout the
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
In- Vehicle Navigation
System
Vehicle Access Refueling Reservation System
Very Satisfied Satisfied Neutral Dissatisfied Very Dissatisfied Not Applicable
29
program, participants indicated that the vehicles were sufficiently fueled, although this was not
always the case. Homebased Users tended to fuel more frequently since they used the cars more
often and for longer trips. Users also indicated that incentives for individuals who frequently
refueled the vehicles ( e. g., coupons for free coffee, videos, etc.) would have provided more
motivati