Mark Fagan, Daniel Comeaux, and Benjamin Gillies
March 2021
Autonomous Vehicles Are Coming:
Five Policy Actions Cities Can Take Now to Be Ready
Mark Fagan, Daniel Comeaux, and Benjamin Gillies
March 2021
Autonomous Vehicles Are Coming:
Five Policy Actions Cities Can Take Now to Be Ready
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
iii
about the authors
Mark Fagan is a Lecturer in Public Policy at Harvard Kennedy School. He leads the
School’s Autonomous Vehicles Policy Initiative through the Taubman Center for State
and Local Government. Mr. Fagan teaches Operations Management, Policy Devel-
opment and Delivery, and Systems Thinking and Supply Chain Management in the
degree program. In executive education programs, he teaches about service delivery
and cross boundary collaboration. Mark has consulted to management in the public
and private sectors on strategy issues for more than 30 years.
Daniel Comeaux is an Associate Transportation Policy Analyst at the Chicago Met-
ropolitan Agency for Planning, working to improve access to mobility in the Chicago
region. Before joining CMAP, Daniel received an MPP from the Harvard Kennedy School
of Government. At HKS, he worked with local officials to solve transportation policy
challenges, including as a Research Assistant for the Autonomous Vehicles Policy Ini-
tiative. He previously worked for the Civic Consulting Alliance in Chicago and received
his undergraduate degree from the University of Chicago.
Benjamin Gillies is the founder and president of the Winnipeg Trolley Company, which
offers tour and transportation services in Winnipeg, Canada, and co-founder of Fools
& Horses Coffee, a small chain of coffee shops focused on deep environmental and
social sustainability. After graduating with a Master in City Planning degree from MIT
and a Master in Public Policy degree from Harvard, Benjamin worked on a number of
transportation policy projects for organizations including NACTO, the City of Toronto,
the City of Winnipeg, and the Canadian Centre for Policy Alternatives. He also writes
freelance articles on Canadian urban and transportation policy, which have been pub-
lished in newspapers including the Globe and Mail, Toronto Star, and Montreal Gazette.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
iv
about the taubman center for state and local government
The mission of the Harvard Kennedy School’s Taubman Center for State and Local
Government is to support current and future public sector leaders in improving the
governance of states, counties, metropolitan regions, and cities through research,
teaching, programs, and convenings.
The Taubman Center works to:
• Develop the next generation of state and local government leaders
• Generate big ideas and solutions to state and local government challenges
• Help state and local government implement and scale solutions
The Taubman Center focuses on urban policy issues, including economic development,
transportation, education, public infrastructure, land use, social services, public sec-
tor technology and data utilization, procurement, and performance management.
This paper is copyrighted by the author(s). It cannot be reproduced or reused without permission. This
paper is available to the public at https://www.hks.harvard.edu/centers/taubman free of charge.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
Contents
I. About This Paper 1
II. Introduction 2
III. Autonomous Vehicles: A Quick Primer 6
IV. AVs Are Coming: The Need for Proactivity 14
V. Five Recommended Policy Actions for Cities 24
Policy Action 1: Foster Mobility as a Service (MaaS) 24
Policy Action 2: Rethink Curb Design and Street Space Allocation 32
Policy Action 3: Manage and Reduce Congestion 41
Policy Action 4: Establish Data-Sharing Guidelines and Agreements 50
Policy Action 5: Reposition Revenues 56
VI. Pilot Project Guidelines 65
VII. Acknowledgements and Contact Information 67
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
1
i. about this paper
Over the last three years, the Autonomous Vehicle Policy Initiative (AVPI) at the Taub-
man Center for State and Local Government at Harvard University’s Kennedy School of
Government has been helping cities and states in the United States and abroad pre-
pare for the policy issues they’ll face as AVs enter the urban environment. Faculty, staff,
and students at Harvard have worked directly with Boston, Toronto, Kansas City, Rhode
Island, Buenos Aires, and Detroit to help them prepare for AVs though policy scrums.
These sessions begin with briefings on AV technology and policy, then provide struc-
tured sessions to identify AV objectives, alternatives, and recommendations. Through
policy scrums, cities can make real progress on relevant policy issues, and develop and
involve a broader AV stakeholder network for diverse inputs and improved outcomes.
This document is designed to be a resource for policymakers. In our work, we
have learned city and state officials are both excited and nervous about the advent of
AVs. They believe this technology can have a positive impact in helping them address
some of the challenges present in their transportation systems, but they also recall
the tumultuous arrival TNCs had, and do not want to face such an unplanned disrup-
tion again. Yet, with so much activity taking place in the AV space, they find them-
selves struggling to determine where to start in crafting the policy to prepare for AV
arrival. Our purpose is to assist them in making sense of this fast-moving field in order
to better craft and implement policies.
The first two sections of this paper offer a glance at the specifics of the technology,
and how companies are testing new AV use cases across the country and abroad. Nei-
ther section is meant to be comprehensive, but they do offer several examples readers
can investigate further based on their interest. The third section provides a series of five
policy action areas we believe cities should be looking at immediately, to lay the foun-
dation for future AV policy. While we would welcome a city government willing to tackle
all five areas, we recognize most local governments will choose one initial area of focus
given practical time, resource, and political constraints. We therefore encourage policy-
makers to consider the most pressing challenges their transportation systems face, and
where those challenges align with the five action areas and the steps we propose. While
these steps can help guide local AV deployment, they can also improve a transportation
system dominated by conventional human-driven vehicles.
Beyond the five policy actions we offer, we encourage cities to develop pilot proj-
ects to test use cases in their jurisdiction. Local context and objectives will drive such
experiments for each city, but our fourth and concluding section provides some gen-
eral guidelines for conducting these pilots.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
2
ii. introduction
On a warm fall day in 2019, four people wait at Providence Station in Rhode Island for
a shuttle bus. Soon, the six-person electricity-powered vehicle that arrives takes them
to Olneyville, a neighborhood that, until May of that year, offered no public transpor-
tation options. That change occurred when the Rhode Island Department of Trans-
portation (RIDOT) and May Mobility, a Michigan-based automated shuttle start-up,
introduced Little Roady, the state’s first self-driving shuttle service. After winning the
contract, the company established a local office and operations center in Providence
and hired more than 50 employees from the region.
1
RIDOT noted that Little Roady’s
debut put Rhode Island at the forefront of mobility testing, while filling a gap in the
state’s public transportation network.
2
In May 2019, a fleet of 12 six-seat electric autonomous vehicles (AVs)—that resem-
ble miniature buses—began to offer free rides along a 5.3-mile route between Provi-
dence Station and Olneyville Square.
3
The vehicles were loaded with sensors, radar
and LiDAR—a technology that uses laser light to perceive the surrounding landscape
in three dimensions—which facilitated their travel along a pre-programmed route.
Although able to operate autonomously, each vehicle was staffed by a safety driver
who could take control if conditions warranted. For instance, rain, speed bumps, con-
struction work zones, potholes, and aggressive drivers could each pose a potential
challenge that might encourage the attendant to switch from autonomous operation
to manual control.
4
Both RIDOT and May Mobility considered the $1.2 million program a research
project devised to study the community impact of a small-scale automated transit ser-
vice. Little Roady presented RIDOT with an opportunity to better understand the risks
and rewards that came with integrating this new technology into Rhode Island’s trans-
portation planning.
5
With the help of academic research and data analytics, as well as
social innovation consultants, policymakers, the community, and other stakeholders,
1 Rhode Island Department of Transportation. Little Roady Autonomous Shuttle Pilot Project, http://www.dot
.ri.gov/projects/trip/LittleRoady.php. Accessed September 11, 2019.
2 Carey, Christopher. “Uber resumes AV testing in San Francisco.” Cities Today, March 17, 2020, https://cities
-today.com/uber-resumes-av-testing-in-san-francisco/. Accessed January 23, 2021.
3 Comery, Beth. “Free Ride – Little Roady Passes The Test.” Providence Daily Dose, July 12, 2019, https://
providencedailydose.com/2019/07/12/free-ride-little-roady-passes-the-test/. Accessed September 11, 2019.
4 Faulkner, Tim. “Driverless Little Roady Shuttle Hits a Few Speed Bumps.” ecoRI News, August 3, 2019, https://
www.ecori.org/transportation/2019/8/2/little-roady-shuttle-reaches-milestone-hits-speedbumps. Accessed
September 11, 2019.
5 Rhode Island Department of Transportation. Little Roady Autonomous Shuttle Pilot Project.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
3
RIDOT hoped to use the experiment to explore mobility solutions, ridership trends,
workforce opportunities and impacts, customer satisfaction, environmental impacts,
and state and local policies.
6
The test pilot program engendered public enthusiasm;
by July 2019, the service had picked up its 9,000th passenger.
7
“Done right, it will be much better than the status quo.”
—former National Transportation Safety Board Chair Christopher Hart,
on the deployment of AVs
Three years earlier, in June 2016, Christopher Hart, then-chairman of the National
Transportation Safety Board (NTSB), predicted that AVs would eventually enjoy wide-
spread adoption and would be the best way to improve road safety. “Done right, it will
be much better than the status quo,” he remarked, even as he noted the technology
presented significant challenges.
8
Indeed, Hart predicted the industry would have to weather an inevitable spate
of fatal crashes before the technology would be perfected.
9
Two years later, on March
18, 2018, Elaine Herzberg became the first known case of a pedestrian killed by a
self-driving vehicle. Herzberg was walking her bicycle across a four-lane road in Tempe,
Arizona late in the evening when she was hit by an automated Uber test vehicle, which
was operating in self-drive mode with a human safety driver sitting in the driver’s seat.
Uber immediately discontinued its testing of self-driving vehicles and Arizona
Governor Doug Ducey suspended all testing on public roads. The incident also
caused other companies to temporarily suspend their tests as well. After the inves-
tigation was concluded, the National Highway Traffic Safety Administration and the
American Automobile Association flagged nighttime driving as an area for safety
improvement.
10
In December 2018, Uber received permission to resume testing in
6 Ibid.
7 Faulkner, Tim. “Driverless Little Roady Shuttle Hits a Few Speed Bumps.” ecoRI News, August 3, 2019, https://
www.ecori.org/transportation/2019/8/2/little-roady-shuttle-reaches-milestone-hits-speedbumps. Accessed
September 11, 2019.
8 Laris, Michael. “NTSB chief foresees transition to life-saving driverless cars that includes deaths.” The Washing-
ton Post, June 30, 2016, https://www.washingtonpost.com/news/dr-gridlock/wp/2016/06/30/ntsb-chief
-sees-deadly-transition-to-life-saving-driverless-cars/. Accessed September 11, 2019.
9 Ibid.
10 National Transportation Safety Board. “Preliminary Report Released for Crash Involving Pedestrian, Uber
Technologies, Inc., Test Vehicle.” National Transportation Safety Board Office of Public Affairs, May 24, 2018,
https://www.ntsb.gov/news/press-releases/Pages/NR20180524.aspx. Accessed September 9, 2019; and
Garcia, Uriel J. and Karina Bland. “Uber crash called ‘unavoidable’ for any driver.” The Arizona Republic,
March 20, 2018, https://www.azcentral.com/story/news/local/tempe/2018/03/20/tempe-police-chief-fatal
-uber-crash-pedestrian-likely-unavoidable/442829002/. Accessed September 9, 2019.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
4
Pittsburgh, and in March 2020 California also re-issued the company’s permit to test
again in San Francisco.
11
Early AV Development Timeline
June 2016: National Transportation Safety Board (NTSB) Chairman Christopher Hart pre-
dicts AVs will enjoy widespread adoption and improve road safety.
March 2018: Elaine Herzberg is killed by a self-driving Uber test vehicle. The resulting
negative publicity is a significant setback for AV development.
December 2018: Satisfied it had addressed the issue, Uber recommences AV testing.
May 2019: Rhode Island DOT and May Mobility begin operating the Little Roady electric
AV shuttle, putting the state at the forefront of mobility testing while extending its public
transportation network.
July 2019: The Little Roady shuttle is met with public enthusiasm, picking up its 9,000th
passenger.
* * *
The Little Roady pilot and Uber’s testing tragedy provide snapshots of a rapidly-evolving
field, one moving inexorably toward fruition. As testing continues, we are now coming
to recognize the challenges and impacts these vehicles may have on congestion, curb
usage, parking revenues, and other areas of interest to local government officials—
issues we address in detail in section IV of this paper. State and local governments
must develop appropriate public policies if they are to yield public benefits from the
inevitable arrival of AVs in their jurisdictions. We are now at a critical juncture, as AVs
are poised to ride the same evolutionary road that Transportation Network Companies
(TNCs)
12
such as ride-hailing services Uber and Lyft did in the 2010s. The impending
arrival of AVs offers states and municipalities a chance to shape this industry with
policies that will deliver public value.
Policymakers must consider many factors when weighing how best to regulate
AVs. As with all change, AVs will have positive and negative impacts on municipalities.
Deployment of AVs with the right policies could potentially have significant positive
impacts on mobility, congestion, public finance, land use, logistics and the environ-
ment in cities across the country. But the challenges facing the industry are daunt-
ing as it struggles with issues surrounding technology, implementation costs, data
security, and adoption and ridership. Cities will have to address the potential loss
11 Carey, Christopher. “Uber resumes AV testing in San Francisco.” Cities Today, March 17, 2020, https://cities-today
.com/uber-resumes-av-testing-in-san-francisco/. Accessed January 23, 2021.
12 A transportation network company, sometimes known as a mobility service provider, is a company that matches
passengers with vehicles, via websites and mobile apps. TNCs for automobiles are commonly referred to as
ride-hailing services.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
5
of parking and other revenues as well as likely job losses as professional drivers are
displaced. Adopting the right policies will position cities to maximize the upside of AVs
while mitigating the downside.
COVID-19 Commentary
In early 2020 as we were synthesizing the work from our research for this paper, the
United States—and the world—began to confront the unprecedented COVID-19. Unfor-
tunately, the early evidence suggests COVID-19 may remain a part of our lives for a
long time. Even with a vaccine, people may begin to change their habits, such as how
they work and where they live, to reduce the risk of infection in a future pandemic. Like
experts in many fields, we were left asking: How will COVID-19 specifically, and the
potential of large-scale pandemics more generally, impact AV development?
It is still early on in determining the interplay between AVs and the spread and
implications of disease spread. Nevertheless, we think it important to acknowledge
this profound new reality, so where appropriate, we have added notes on COVID-19 in
the text. Drawing on the emerging research about the urban planning and transpor-
tation impacts of coronavirus, empirical evidence on American travel more generally,
along with our knowledge of the benefits and limitations of AVs, we offer some initial
ideas about the role AVs can play as citizens and governments react to this pandemic.
Notwithstanding the uncertainties, we are confident that AV policy making is still
a priority for city and state officials as they respond to the evolving AV mobility tech-
nology and respond to COVID-19.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
6
iii. autonomous vehicles: a quick primer
What Does It Mean to Be Autonomous?
Most automobile drivers have had some experience with autonomous driving fea-
tures. According to the Society of Automotive Engineers,
13
there are several levels of
automation:
Society of Automotive Engineers Levels of Vehicle Automation
Level 0: Manually controlled, with no automated features.
Level 1: A human driver controls an automated feature (for instance, cruise control,
which allows drivers to set a constant vehicle speed).
Level 2: A human driver must steer, brake or accelerate to maintain safety, but some
driving function is automated (for instance, lane centering).
Level 3: A vehicle is self-driving but a person must sit in the driver’s seat and, in event of an
imminent crisis, disengage the vehicle from autonomous operation and drive it manually.
Level 4: A vehicle is autonomous but may be limited by certain weather, lighting or road
conditions, or may be restricted to operating within a pre-defined area or route.
Level 5: A vehicle operates autonomously in all driving conditions and situations.
Most AV operators are focused on developing services that can operate on levels
4 or 5 to remove the need for—and costs associated with—human intervention. While
many are currently testing Level 4, Level 5 is at least 10 years—and possibly several
decades—away from mass-adoption feasibility. Even so, designers are imagining—
and creating—the future: first- and last-mile shuttles, land-based AV delivery vehicles
and cities filled with cars, buses, and even trucks that drive themselves.
How Does Technology Enable AVs?
14
When an AV company wants to operate in a geographic area, its workers drive the
vehicle over the terrain repeatedly, each time gathering information about the operat-
ing environment including the landscape, traffic behavior, and pedestrian behavior.
AVs operate based on rules, and those rules are in part governed by maps built and
13 The U.S. Department of Transportation uses the Society of Automotive Engineers’ five levels of automation for
on-road motor vehicles in its Federal Automated Vehicles Policy and the document became a de facto global
standard adopted by stakeholders in the automated vehicle technology sphere. Source: Society of Automotive
Engineers International website, https://www.sae.org/news/press-room/2018/12/sae-international-releases
-updated-visual-chart-for-its-“levels-of-driving-automation”-standard-for-self-driving-vehicles, accessed Sep-
tember 12, 2019.
14 For more information on AV technology, please see “Self-driving car technology: When will the robots hit
the road?” Available at: https://www.mckinsey.com/industries/automotive-and-assembly/our-insights/
self-driving-car-technology-when-will-the-robots-hit-the-road/de-de.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
7
updated using this data. These maps, devised from data gleaned after operating on a
roadway hundreds of times, provide “virtual guardrails” for an AV’s operation.
How do AVs collect the data necessary
to construct these complex maps? AVs use
outward-facing cameras to provide images
of the surrounding environment, and technol-
ogies including light detection and ranging
(LiDAR) to perceive the landscape in three
dimensions. These data are fed into a power-
ful computer that coordinates and acts on the
information provided by the peripherals.
Through the mid-1980s, robots used sensors to gather information about their
environment, which was then used to plan and control their movements. This was
referred to as the Sense, Plan, Act paradigm. Building on this paradigm, AVs See,
Think, and Act: they gather information about their surroundings, such as obstacles
or “landmarks;” they form a plan to respond to the sensor data using an existing strat-
egy (i.e., if there is an object in its path it moves to the left); and finally, they carry out
the actions devised in their plan. These operations are essentially performed in real
time. One technology company, ZF, claims its automotive supercomputer is capable of
performing the equivalent of 150 trillion operations per second.
15
The technology already exists to allow AVs to engage in V2V and V2I communica-
tion, and in the future, AVs will also communicate with the wider world around them
using V2X functionality to operate more safely and efficiently.
Forms of AV Communication
1. Vehicle-to-Everything (V2X): Information passes from a vehicle to any person,
device, entity, landmark or object that may affect the vehicle and vice versa.
2. Vehicle-to-Vehicle (V2V): Information passes from one vehicle to another.
3. Vehicle-to-Infrastructure (V2I): Information passes from a vehicle to the surrounding
infrastructure.
Initially, the AV industry was focused on V2I, which involves the exchange of
safety and operational data between individual vehicles and the transportation
infrastructure that supports them. In addition to stop signs and crosswalks, other
connected components could include RFID readers embedded in overhead signs on
overpasses, as well as connected cameras, lane markers, traffic lights, and parking
15 Weitbrecht, Lars. “Autonomous Driving: The Formula for Success.” ZF Vision Magazine, January 7, 2019, https://
www.zf.com/site/magazine/en/articles_13440.html. Accessed September 13, 2019.
HOW AVS WORK
1. Outward-facing cameras provide
images of the environment sur-
rounding the vehicle.
2. Vehicles use laser light tech-
nology, or LiDAR, to perceive
landscape in three dimensions.
3. A powerful computer coordinates
and acts on the information pro-
vided by cameras and LiDAR.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
8
meters. However, the complexity and costs of interacting with roadside infrastructure
has called into question how quickly the technology will be implemented.
16
Indeed,
a 2015 U.S. Government Accountability Office report found that every individual V2I
hub could cost as much as $52,000, leaving municipalities to wonder how V2I tech-
nology would be funded.
17
How could municipalities gain taxpayer support for such
investments—particularly if the benefits accrue primarily to private AV fleet operators?
And who specifies the parameters of the smart infrastructure? These are only a few of
the issues that need to be addressed from a public perspective if V2I interaction is to
become a requirement for AVs.
While V2V was also seen as an attractive option for improving safety, it will likely
be decades before all vehicles on the road are capable of communicating with each
other. While these and other V2X technologies may become a critical component to
the future of AVs, they are likely years away from being realized.
AV Development
Many companies, from traditional car manufacturers to
tech start-ups, are currently working on AVs capable of
Level 4 autonomy. Traditional manufacturers including
GM, Ford, and Nissan-Renault already have AVs on the
road. Tesla has announced plans to roll out extensive AV
coverage in the early 2020s. Alphabet (best known as
Google’s parent company) subsidiary Waymo is heavily
invested in AVs with around 600 cars in service, and
plans to deploy 20,000 through its ride-hailing service,
Waymo One in the next few years.
There are also tech startups in this space. For instance, nuTonomy, now part of
Aptiv, is an MIT spinoff that has been testing vehicles capable of level 4 autonomy for
more than three years in Singapore and Boston.
18
16 Forrest, Conner. “The X-factor in our driverless future: V2V and V2I.” ZD Net, February 1, 2018, https://www
.zdnet.com/article/the-x-factor-in-our-driverless-future-v2v-and-v2i/. Accessed October 8, 2019.
17 United States Government Accountability Office. “Intelligent Transportation Systems: Vehicle-to-Infrastructure
Technologies Expected to Offer Benefits, but Deployment Challenges Exist.” United States Government Account-
ability Office Report to Congressional Requesters, September 2015, https://www.gao.gov/assets/680/672548
.pdf. Accessed October 8, 2019.
18 Sawers, Paul. “The world’s first public self-driving taxi service hits Singapore roads today.” Venture Beat,
August 25, 2016, https://venturebeat.com/2016/08/25/nutonomy-singapore-self-driving-car-service/.
Accessed October 8, 2019; and WBUR News and Wire Services. “Buckle Up, Boston: Testing of Self-Driving Cars
Begins Tuesday.” WBUR, January 2, 2017, https://www.wbur.org/bostonomix/2017/01/02/self-driving-car
-testing-begins. Accessed October 8, 2019.
KEY AV DEVELOPERS
Aptiv
Argo AI
Blackberry
Drive.Ai
Ford
GM
May Mobility
Nissan-Renault
Optimus Ride
Tesla
Aurora
Waymo
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
9
Another set of providers are offering autonomous shuttle buses capable of trans-
porting between four and eight passengers. Companies such as May Mobility, Opti-
mus Ride, and Argo are designing, building, and operating these vehicles.
Today, most AVs are standard driver-based automobiles that have been modified
to include AV technology. For example, Waymo uses the Chrysler Pacifica, and Uber
relies on Volvo XC90 SUVs.
19
In early 2020, however, the NHTSA gave Nuro the first
permit to produce vehicles specifically designed as automated: Designed for roads,
but without windshields or mirrors.
20
This is a major milestone, as federal officials
acknowledge and permit construction of vehicles that will never have a driver behind
the wheel.
While many autonomous car companies currently put self-driving technology
in and on conventional gasoline internal combustion engine vehicles, companies
developing shuttles are designing and building electric from scratch. In addition,
many AV providers are planning for their vehicles to be electric-powered when they
operate at scale.
The rate of AV adoption, however, will be predicated on whether the cost of the
technology used to convert a traditional vehicle to an AV will continue to decline. Early
test passenger AV vehicles cost as much as $300,000 each but over the next 10 to 15
years, costs will likely be less than $50,000 as manufacturers scale for production.
Nevertheless, some cities are currently paying high annual leasing fees for AV tests
in their municipalities. For instance, French AV shuttle company EasyMile is charging
$27,000 per month per shuttle. Arlington, Texas contracted with Drive.ai to fund a
year-long test with a trio of AV shuttles for $435,000.
21
How Might AVs Arrive in Your City?
AV developers and operators are determining the initial use cases for how AVs will be
deployed. Given how aggressively companies are testing the technology, one of the first
use cases has been ride-hailing services operating within urban areas—essentially,
driverless taxis and TNCs. Several factors make this model highly attractive, including
the high likelihood of consumer adoption, based on existing consumer acceptance of
taxis and TNCs. Indeed, judging by the explosive growth of TNC rides, the urban
19 Gardner, Greg. “Waymo to Convert Chrysler Minivans and Jaguar I-Paces in Detroit plant.” Forbes, April 23, 2019,
https://www.forbes.com/sites/greggardner/2019/04/23/waymo-to-convert-chrysler-minivans-and-jaguar-i
-paces-in-detroit-plant/?sh=787e2f632c08. Accessed October 8, 2019; and O’Kane, Sean. “Uber debuts a new
self-driving car with more fail-safes.” The Verge, June 12, 2019. Accessed October 8, 2019.
20 https://www.autonews.com/mobility-report/milestone-driverless-delivery.
21 Waddell, Kaveh. “The true cost of autonomous cars.” Axios, December 9, 2018, https://www.axios.com/who
-will-pay-for-autonomous-vehicles-11702624-0fbf-4b5f-9d3a-bf5c4cd3ed8a.html. Accessed September 16, 2019.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
10
population is eager to adopt a new model for mobility that is convenient to use. Second,
assuming the TNCs pass along to their riders the cost savings of operating without driv-
ers, consumers will likely be eager to embrace a less expensive transportation option.
Finally, because mapping the environment is costly and time consuming, limiting the
scope of operation to a relatively small geographic area decreases time to market. AV
TNCs are most likely to be implemented within dense urban cores, ensuring high utiliza-
tion, which will make the economics of AVs more viable.
A second emerging use case is AV shuttle
service. The initial applications have been pro-
viding fixed-route service on short routes as a
first/last mile solution. For example, Drive.Ai
operates an AV shuttle in Arlington, Texas that
connects passengers between various down-
town “last mile” destinations.
Another likely early use case is providing
on-demand mobility within specified areas,
such as retirement communities. For example, Via and EasyMile currently provide a
flexible, on-demand transit service to residents of the Marian Grove Retirement Village
in New South Wales, Australia.
22
AVs are a particularly attractive service for this and
other senior-focused communities because of the otherwise limited mobility some
residents might face.
Beyond carrying passengers, land-based AVs can be used to substitute for or aug-
ment existing food and package delivery services. As demand for home and office deliv-
ery has grown, delivery services are keen to find a lower-cost way to get their goods to
consumers. Several land-based delivery AVs are currently being tested. Nuro, for exam-
ple, uses AVs to deliver groceries in Scottsdale, Arizona, and Houston, Texas. Amazon is
testing its own delivery robot service, Scout, in Snohomish County, Washington.
While these AVs could offer attractive economics, they face the challenge of
delivering their goods from the vehicle to the recipient’s door—the so-called “last 100
feet” problem. To solve this problem, Agility Robotics has partnered with Ford to test
a land-based AV delivery program using a new robot, Digit. When a Ford AV arrives
at a delivery location, Digit “unfolds” itself from the back of the vehicle, selects the
appropriate package (or pizza, for instance) to deliver then walks it to the recipient’s
door. Digit, which was designed to look and walk like a human, is able to climb up and
22 Intelligent Transport. “Via unveils its first autonomous vehicle service pilot.” Russell Publishing Ltd, July 15,
2019, https://www.intelligenttransport.com/transport-news/84115/via-first-autonomous-vehicle-service/.
Accessed February 3, 2020.
AV USE CASES
• Ride-hailing services (e.g.,
taxis and TNCs)
• Shuttle services (e.g., short
first/last-mile transit carriers)
• On-demand mobility services
(e.g., retirement communities)
• Delivery services (e.g., food,
office supplies, packages)
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
11
down stairs carrying packages up to 40 pounds.
23
A Digit-like solution will be part of
an early use case because if and when the last 100 feet problem is solved, it will allow
the realization of the holy grail of automated land-based delivery: 24/7 operation with
no humans.
Which of these services arrives first in your community will depend on a number
of factors, including the investment decisions of private operators, projected demand
and even climate. As we discuss below, local and state governments also have a role
to play in inviting or limiting companies to test in their jurisdictions, meaning the deci-
sions you make will also influence the first use case your citizens experience.
The Regulatory Climate
Autonomous vehicles will be regulated at the federal, state and municipal level. The
National Highway Traffic Safety Administration is charged with setting safety stan-
dards, though the agency has not yet established rules for the industry. In the absence
of federal regulations, states and cities are stepping in. Some states, like Massachu-
setts, have devised an overarching set of regulations while allowing municipalities
to set their own local rules. At the other end of the spectrum, some states have pre-
empted local authority; the California Department of Motor Vehicles, for instance, is
setting statewide rules for AV operations. Some analysts believe the federal govern-
ment will eventually set safety standards, leaving states and municipalities responsi-
ble for devising regulations for testing and operations.
Seeing work at the state and federal level, some local officials may wonder what
authority they have in AV policy development. We argue city governments can and
should have a significant influence on how automated vehicles will arrive in their juris-
dictions. Through their transit agencies, taxi/TNC regulations, land use planning and
streetscaping priorities, parking authorities, Vision Zero initiatives, congestion pric-
ing, fleet purchasing power, and speed limit measures, local governments have many
levers to pull. While not every city will have every tool (in some parts of the country,
transit and TNCs are managed by the state, for instance), we encourage all policymak-
ers to look across their departments and see how they can shape the local rollout of
AVs to the benefit of the community.
23 Yan, Michelle. “Watch Ford’s delivery robot that walks on two legs like a human.” Business Insider, May 23,
2019, https://www.businessinsider.com/ford-delivery-robot-digit-walks-on-two-legs-like-human-2019-5.
Accessed September 22, 2019.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
12
Challenges to Operationalizing AVs
Five challenges— (1) technological barriers; (2) data privacy and security concerns; (3)
moral decision-making concerns; (4) impact on jobs; and (5) public trepidation—must
be overcome for the potential of AVs to be realized. While not all of these issues are
under the authority of local policymakers, city and regional officials must understand
these topics and concerns in order to engage effectively with citizens and stakeholders.
Technology
AV technology needs further refinement to improve passenger and pedestrian safety.
While the frequency of safety-driver interventions (i.e., disengagements) are declining
(indicating that the technology is becoming more reliable), safety drivers are still nec-
essary, and it is unclear how soon AVs will be able to operate at levels 4 or 5—in any
use case scenario.
For efficient AV operation, there most be continual communication between the
vehicles and the operators data management platform. As such, the availability of
5G—fifth-generation cellular network technology—is another technology issue. With
speeds as much as one hundred times faster than 4G, 5G may offer AV operators
speeds of up to 100 gigabits per second—speeds AV operators will need to meet the
communications requirements of their fleets. While four major carriers—AT&T, Sprint,
T-Mobile and Verizon—had announced plans to roll out 5G in 2019, by July 2020 only
Los Angeles, Phoenix, Houston, Dallas, and Atlanta are live with 5G on all four net-
works and most cities have only one, or none. While 5G is a “nice to have” within the
current testing environment, it will be a necessity as AVs are implemented in earnest.
Data Privacy and Cybersecurity
AVs capture an enormous amount of data about vehicle operation, passengers and
the surrounding environment. Several questions arise from the collection of the data.
AVs and Data Privacy Questions
1. Who owns the information generated by AVs?
2. How can and should this information be used?
3. Can this information be transmitted and stored in a secure way?
4. Should users have to opt in or are AV operators able to legally and ethically harvest
data without user consent?
An associated technology issue is cybersecurity. AVs are vulnerable to hacking.
Given the potential size of AV operators’ fleets, ransom demands after a successful
hack could emerge as a significant concern. There is also a risk that hijacked AVs could
be used as terrorist weapon.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
13
Moral Decision-Making
Today we rely on a vehicle’s driver to make moral decisions about safety while driv-
ing. For instance, if a child runs into the road and the only way to avoid the child is to
swerve and hit a wall, the vehicle driver makes that decision—and is responsible for
the consequences. With AVs, the software is responsible for decision-making when
faced with this moral dilemma. Will passengers be comfortable “delegating” this deci-
sion to AV operators’ software programmers?
MIT is collecting data on human perspectives on moral decisions and applying it
to machine intelligence—such as self-driving cars. MIT’s Moral Machine project web-
site asks visitors to react to scenarios such as one in which a self-driving car will either
hit a barrier wall, killing all five passengers, or swerve to avoid the barrier, killing five
pedestrians. The website asks, “What should the self-driving car do?”
The MIT project synthesized website visitor responses—40 million decisions from
residents of more than 200 countries—to produce aggregate information on consen-
sus morality.
24
While the project identified much commonality among respondents’
views—for instance, most preferred to save humans over animals and younger people
over older people—there was a lack of consensus over other moral decisions, such as
whether to prioritize pedestrian versus passenger lives.
25
We cede control to human drivers when riding on a bus or when a family member
is at the wheel, but will passengers be willing to cede control over these decisions to
AV operators? It is unclear how AV operators will choose to make these decisions and
what role passengers will expect to play. This will also likely lead to major changes in
insurance, as the companies operating the AVs and/or writing the code behind their
decision-making become liable for the results of the decisions the vehicles make and
the actions they take.
Impact on Jobs
In the United States, more than 4 million Americans drive professionally for compen-
sation. AVs are expected to significantly impact these jobs; if TNCs adopt AVs at scale,
their driving workforce will no longer be needed. We have already seen significant
negative impact on taxi drivers as a result of TNC growth. AVs would cause further
job displacement, a concern for municipalities, especially given added uncertainty
caused by the coronavirus pandemic.
24 Awad, Edmond, Sohan Dsouza, Richard Kim et al. “The Moral Machine experiment.” Nature 563, 59–64, 2018,
https://www.nature.com/articles/s41586-018-0637-6.
25 The Economist. “Whom should self-driving cars protect in an accident?” October 27, 2018, https://www
.economist.com/science-and-technology/2018/10/27/whom-should-self-driving-cars-protect-in-an-accident.
Accessed October 8, 2019.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
14
On the positive side of the ledger, AVs offer a variety of economic opportunities
for communities, ranging from the development and fabrication of AVs and software
development to staffing local AV operating sites and building out 5G networks. How
the positives and negatives net out will be a function of each city’s intrinsic position-
ing for AVs and policies that are adopted by its government.
Consumer Trepidation
In early 2019, a study by the American Automobile Association (AAA) found that 71%
of respondents to an annual survey on self-driving cars report being afraid to ride in a
completely autonomous vehicle. Even fewer—only 19%—said they would be willing to
allow a child or another family member to ride in an AV. In prior years, AAA had found
more public acceptance for AVs, but press accounts of crashes and deaths during AV
testing appears to be eroding public support.
26
Even so, 55% of respondents expect
self-driving vehicles to become the norm by 2029. The challenge for AV operators is to
build public trust and overcome current skepticism.
A 2019 AAA study found that 55% of respondents expect self-driving
vehicles to become the norm by 2029.
The timing of AV deployment is currently uncertain, but when the industry is able
to develop effective responses to these challenges, AVs will move from concept and
testing to becoming part of the mobility ecosystem.
iv. avs are coming: the need for proactivity
Policymakers, faced with many pressing current challenges regarding traffic safety,
congestion, and sustainability, may wonder why they should give time and attention
to future automated travel. We believe it is worth their while for two major reasons.
Firstly, as we discuss below, the measures we advocate serve to address many exist-
ing transportation issues, while also preparing cities for the AV age. Secondly, while
AV companies still have a number of hurdles to overcome, there are literally thousands
of self-driving vehicles on the road across the country and around the world today, and
some are coming very close to commercialization.
26 Eisenstein, Paul. “Three-quarters of Americans ‘afraid’ of fully autonomous vehicles.” NBC News, March 14,
2019, https://www.nbcnews.com/business/autos/three-quarters-americans-afraid-fully-autonomous-vehicles
-n983091. Accessed October 8, 2019.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
15
In the early 2010s, a number of startups made bold claims around how soon we
would be riding in automated cars. Their optimism was muted by technological chal-
lenges and especially the death of Elaine Herzberg in Arizona. This tragedy led to a
period of self-reflection among many AV industry players, which has resulted in better
testing protocols.
We will discuss a number of these accomplishments below, but perhaps the most
noteworthy recent development was a simple email sent by Waymo in October 2019.
Entitled “Completely driverless Waymo cars are on the way,”
27
the message informed
customers of its ride-hailing service, Waymo One (currently operating only in the Phoe-
nix East Valley region), that their next trip might not have a human safety driver behind
the wheel, and as of October 2020, 5–10% of Waymo’s rides are completely driverless.
Certainly, there are many parameters around these Phoenix-area trips. The point is
that policymakers should not see AVs as relegated to some far-off future: there are
commercial ride-hailing trips underway, today, in self-driving cars.
Moreover, municipal governments have a key role to play in how AV deployment
takes place. AVs have the potential for improving safety, mitigating congestion, stim-
ulating urban vibrancy, and improving municipal finances; however, these enticing
outcomes are not guaranteed. Companies have invested billions of dollars in AV devel-
opment, and have their own expectations around how to maximize their use. While
many stated company objectives align with urban policy goals, city officials should
not assume companies will deploy their automated fleets in a manner that makes
their communities more equitable, vibrant, sustainable, and safe. City officials should
ensure AVs arrive in a manner that works for the people of their communities, rather
than in a way that aligns with the objectives of their operators but may run counter to
broader city objectives.
Policymakers should not see AVs as relegated to some far-off future: there are
commercial ride-hailing trips underway, today, in self-driving cars.
One need only look to the arrival of TNCs or micromobility vehicles like scooters
to understand the headaches that come when cities are unprepared for new technol-
ogy, such as the scramble to regulate an industry that has already arrived and built a
loyal customer base. As such, local governments should be taking immediate steps
to understand the existing use cases of this technology, and preparing the policy
27 Korosec, Kirsten. “Waymo to customers: ‘Completely driverless Waymo cars are on the way.’” Tech Crunch,
October 9, 2019, https://techcrunch.com/2019/10/09/waymo-to-customers-completely-driverless-waymo
-cars-are-on-the-way/. Accessed February 3, 2020.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
16
structures and built environments of their communities for the arrival of these vehi-
cles—an arrival that, relatively speaking, may not be all that far down the road. Before
delving into what our research shows are the critical starting points for policymakers,
we examine the applicable use cases that those policymakers may need to prepare
for in their cities.
Existing Use Cases
Current and Near-Term AV Use Cases
• Automated ride-hailing (similar to today’s TNCs)—picking riders up at a requested
point and dropping them off directly at their destination
• Automated ride-hailing with shared rides, using algorithms to match riders going in
similar directions into a combined trip for a lower price
• Small fixed-route shuttles to provide transit-type service in lower-demand contexts
• Fully automated conventional transit vehicles, such as buses and streetcars
• Automated land-based delivery vehicles, including the possibility of vehicles that
run on the street, the sidewalk, or some combination of the two
Potential use cases—different ways users can interact with AV technology—are
broad in their range. In general, our use cases may best be broken down into two cat-
egories: moving people and moving goods.
Moving People
Point-to-Point Services
From the operators’ economic perspective, the most attractive use case will likely be
similar to today’s ride-hailing services. In this model, AVs would carry a single passen-
ger (or passengers traveling as a group) from their origin directly to their destination
for a pre-set rate. This model is familiar and comfortable to the consumer, as the ser-
vice is already ubiquitous today and the network is relatively easy to manage. Further,
potential operators would like to take advantage of the attractive economics of lower
operating costs and the potential to earn additional revenue by selling their operating
data. Your TNC operator could, for instance, sell your travel data to a fast food com-
pany so they could send you coupon notifications because it knows when you are
going by one of their restaurants.
This use case would likely first emerge in a virtually defined geographic area (via
what’s known as geo-fencing), such as a city center. Since AVs improve the efficiency
and reduce the uncertainty of their mapped routes by repeatedly traveling over those
routes (while analyzing the movement of other vehicles, pedestrians, congestion
times, etc.), operating within a restricted geographic area also serves to lower oper-
ating costs and improve safety. Waymo One is the leading example of this use case.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
17
While still in the testing phase, their intention is to eventually cover the 600-square-
mile Phoenix area with a fleet of hundreds of cars. Waymo launched its ride-hailing ser-
vice in the Phoenix suburbs in 2017, first as a free testing program and, as of 2019, as
a commercial service.
28
All Waymo One rides are in autonomous vehicles with a safety
driver present and ready to take over if the AV is unable to safely operate in a given
situation. At this time, the program is only available to a small group of Phoenix-area
residents. Waymo is also partnering with Lyft to offer rides in the Phoenix area, using
a small number of the same vehicles as deployed in Waymo One.
Aptiv has partnered with Lyft to provide AV ride-hailing services in Las Vegas and
has provided at over 100,000 rides to date. Company officials have noted that this is
the largest AV ride-hailing service in the world open to the general public. Mercedes
and Bosch have also launched their Urban Automated Driving project, which is run-
ning a similar automated ride-hailing trial in San José.
29
Passengers have generally been receptive to the experience of riding in an AV.
However, in the wake of the pedestrian fatality in Arizona, there has also been some
public unhappiness with AVs in Arizona, including vandalism on Waymo’s vehicles.
Nevertheless, more services are likely to launch in the coming months, including
those by Waymo, AutoX, Zoox, and Pony.ai. Each of these firms have applied for and
received permits from California to launch pilot AV ride-hailing services.
30
Most current deployments of AVs use a staged approach, where a company will
develop high-quality maps of a specific neighborhood, test in this area, then expand
to the next. In Boston, for example, Aptiv began testing in the South Boston water-
front and has since grown to mapping larger portions of South Boston. It is important
for policymakers to ensure that as companies enlarge their geographic footprint, they
do not ignore lower-income communities. When issuing permits, city officials should
understand the timeline for future deployment, so a company does not expand into only
high-income areas, then cease mapping so as to avoid having to provide service to other
parts of the city. Companies should also not be permitted to prioritize pick-ups in some
communities or inflate wait times to discourage use among certain residents.
28 Krafcik, John. “Waymo One: The next step on our self-driving journey.” Medium, December 5, 2018, https://
medium.com/waymo/waymo-one-the-next-step-on-our-self-driving-journey-6d0c075b0e9b. Accessed February
3, 2020.
29 Davies, Chris. “Mercedes launches autonomous cars in first US ride-hailing pilot.” SlashGear, December 9,
2019, https://www.slashgear.com/mercedes-benz-daimler-bosch-us-ride-hailing-autonomous-car-pilot
-launches-09602611/. Accessed February 3, 2020.
30 Hawkins, Andrew J. “Waymo gets green light in California to pick up passengers in self-driving cars.” The Verge,
July 3, 2019, https://www.theverge.com/2019/7/3/20680938/waymo-self-driving-cars-passengers-permit
-california-pick-up. Accessed February 3, 2020.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
18
We do not assume companies will employ these strategies, but there are some
examples of technology companies that have limited their services based on neigh-
borhood. Transportation is an essential service, and it is a priority for officials to
proactively ensure everyone enjoys access to the technology of the future. Cities
or states not should simply rely on the goodwill of companies; they should have
clear policies outlining performance expectations to operate in their jurisdictions.
This will also mean contemplating how to serve those without smartphones or credit
cards. We offer additional thoughts on equity throughout our policy recommenda-
tions, but this is an overarching lens through which you need to view all AV policy as
this technology matures.
Shared Rides
The second use case is similar to the first but would focus on shared rides instead of
serving only one origin and destination per trip. TNCs followed this evolution in their
own development, starting with single rides before adding a shared-ride trip-matching
capability. A shared-ride system would match two or more riders traveling in a similar
direction into one combined trip, with a lower price to compensate for the potential
increase in (and uncertainty of) travel time. If the shared-ride use case is the dominant
one, it could have positive effects in terms of lowering congestion and emissions, par-
ticularly compared to the single-rider use case.
Shared rides in an autonomous context do present new challenges: passenger
safety and security will need to be addressed, as passengers may be uncomfortable
riding in a shared AV with no “norm setter” (i.e., driver) to provide reassurance or
to regulate passenger behavior. To overcome these obstacles, some companies have
discussed wholesale shifts in vehicle design, for instance to allow for partitions
between passengers. At least in the early stages, however, this challenge, as well as
the broader economic attractiveness of single-rider trips, makes it more likely that
single-rider AV trips will predominate in the absence of public policy intervention.
Fixed-Route Services
AVs could also provide service along pre-set routes, similar to shuttle and bus ser-
vices that operate today. In particular, low-speed shuttle services, similar to the Little
Roady pilot project, are likely to be another one of the early use cases. In this deploy-
ment model, AVs would run on fixed routes in relatively small vehicles. This would
reduce the complexity of operations, since running on a fixed route would reduce the
scope of detailed mapping required and minimize the number of choices that the AV
operating system needs to evaluate. Such services could provide so-called first- and
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
19
last-mile connections to other transportation hubs. For example, Little Roady’s route
begins at Providence station, the city’s Amtrak and commuter rail station, allowing
passengers to transition from train to bus to complete their journey.
Beyond Providence, May Mobility has launched shut-
tle services in three cities: Detroit and Grand Rapids, Mich-
igan, and Columbus, Ohio. These shuttles are operating in
relatively dense, urbanized settings, with regularly sched-
uled headways and a fixed route. Denver’s regional transit
agency, RTD, is partnering with EasyMile to operate an AV
shuttle between the Denver Airport and a nearby suburb.
This represents the first time a U.S. transit agency has
used AVs in lieu of a regularly scheduled bus route. Optimus Ride recently launched a
shuttle service in the Brooklyn Navy Yard in New York City. The shuttle is designed as
a “last-mile” connection for commuters from a ferry terminal to the thousands of jobs
located within the complex.
AV shuttles could also provide transit-type services in other contexts, such as
retirement communities and college campuses, where shuttle operations are already
common today. Eventually, AVs will likely be integrated into the operations of mass
transit more broadly, such as through automated bus systems in urban cores. Auto-
mated technology already exists on some urban rail systems, which benefit from a
completely separated right-of-way. The introduction of AV technology could allow
the benefits observed through automated trains, such as precision stopping and the
ability to redeploy vehicle drivers to more customer-oriented roles, to be extended to
other mass transit vehicles. However, any large-scale transit deployment is likely to
come after the successful deployment of smaller vehicles in less difficult contexts.
Moving Goods
Local officials will also need to contend with heightened commercial interest in auto-
mated land-based delivery vehicles. The growth of home delivery—of everything from
meals to mattresses—has caused the commercial sector to search for more efficient
and cost-effective ways to deliver their goods. This impact will only become more
prominent in the years to come, as internet-based deliveries skyrocket; in 2009, there
was one daily Internet delivery for every 25 Americans, while today that number has
jumped to one for every eight, and experts believe it will double once more by 2023.
31
31 Sisson, Patrick. “How your online shopping snarls traffic on city streets.” Curbed, January 10, 2019, https://
www.curbed.com/2019/1/10/18177399/amazon-delivery-traffic-online-shopping-e-commerce. Accessed Febru-
ary 3, 2020.
CITIES WITH AV
SHUTTLE SERVICE
Columbus, OH
Denver, CO
Detroit, MI
Grand Rapids, MI
New York, NY
Providence, RI
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
20
In New York City alone, average daily deliveries tripled between 2009 and 2017, to
1.1 million.
32
It is thus not surprising that researchers at the Texas A&M University
Transportation Institute found trucks cause 28 percent of all U.S. congestion, despite
making up just seven percent of the nation’s total vehicle fleet.
33
Land-based AV delivery might be an easier early use case as the risk of personal
injury to riders is eliminated. John Krafcik, Waymo’s CEO, has stated that “driverless
delivery likely has a better chance of catching on early versus passenger transpor-
tation.”
34
Cities might therefore find themselves contending with these services well
before they have to address automated ride-hailing in their communities.
The primary city use case for land-based delivery AVs in freight is replacing
human-operated vehicles. Self-driving delivery company Udelv has already partnered
with Draeger’s Market in San Mateo, California, Esperanza stores in Oklahoma City,
and Walmart in Surprise, Arizona.
35
As of late 2019, the company had completed around
6,000 automated deliveries (with a safety driver at the helm).
36
While Udelv and Argo’s
vehicles look similar to vans on the road today, Nuro has designed land-based delivery
vehicles that do away with seating to offer only space for parcels. Nuro launched its
first trial with Kroger in Scottsdale, Arizona, and in 2019 announced a partnership with
Domino’s in Houston.
37
The company uses Toyota Priuses to map the communities in
which it will operate, then driverless vehicles employ the maps to run in these areas.
Commercial Delivery
Commercial drop-offs present an even more promising use case. Unlike food or pack-
age delivery routes to private homes, which require a different route every day and can
have up to 150 stops, commercial services—such as running stock to a restaurant or
store, or packages to an office tower—only have about ten to 15 consistent drop-offs.
It is possible the same vehicle could operate on the same route to the same dozen
32 Haag, Matthew and Winnie Hu. “1.5 Million Packages a Day: The Internet Brings Chaos to N.Y. Streets.” The New
York Times, October 27, 2019, https://www.nytimes.com/2019/10/27/nyregion/nyc-amazon-delivery.html.
Accessed February 3, 2020.
33 Sisson, “How your online shopping snarls traffic on city streets.”
34 Laury, Daniel. “Autonomous Delivery: The Thesis.” Medium, November 12, 2019, https://medium.com/@udelv
/autonomous-delivery-the-case-for-fixed-repeatable-autonomous-delivery-routes-19044c29177. Accessed
February 3, 2020.
35 Hawkins, Andrew J. “Robot delivery startup Udelv partners with Walmart and Baidu.” The Verge, January 8,
2019, https://www.theverge.com/2019/1/8/18173776/udelv-self-driving-delivery-walmart-baidu-ces-2019.
Accessed February 3, 2020.
36 Laury, “Autonomous Delivery: The Thesis.”
37 Murphy, Mike. “Dominos is about to start delivering pizzas with autonomous robots.” Quartz, June 17, 2019,
https://qz.com/1644476/nuro-will-deliver-dominos-pizza-with-its-robots-in-houston/. Accessed February 3,
2020.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
21
commercial customers every day. This would make it easier to map the route and
address obstacles, remove the risk associated with carrying human beings, and even
eliminate the difficulty of having someone pick up the delivery from the curb. These
advantages make it one of the most achievable early deployments of AV technology,
and one that Udelv has already begun to explore.
“Driverless delivery likely has a better chance of catching on
early versus passenger transportation.”
—Waymo CEO John Krafcik
Home Delivery
For home delivery, land-based AV delivery operators still have to tackle the “last thirty
feet” problem. How will a package, or a pizza, get from the vehicle to the customer’s
door? Would delivery vehicles still need to be staffed as a consequence?
This issue remains unsolved. However, there are several different models that
could help to address the problems of automated freight delivery. Some AV companies
are now testing sidewalk-based delivery robots that could bring goods closer to the
front door. Others are developing mobile kiosks that would park nearby, requiring cus-
tomers to come pick up their order. Some grocery stores have even proposed mobile
storefronts, where customers would be able to shop for fresh produce and more right
at the curb. It is likely that some combination of these will emerge, but which model
predominates will vary based on the specifics of the local built environment.
The most developed tests are those that focus on small land-based delivery
robots that can traverse sidewalks and curbs. FedEx reports more than 60 percent
of its retail delivery customers live within three miles of the stores from which they
have ordered, which is a very manageable range for a small, wheeled robot.
38
The
delivery company has launched its SameDay bot, traveling at ten miles per hour while
couriering packages between the company’s offices in Memphis. With multiple sets
of wheels, it can climb steps and curbs to reach even difficult destinations.
39
Using its
own six-wheeled robot, Europe’s Starship Technologies has completed over 25,000
deliveries worldwide, with trials in nine U.S. states. Many of these trials have been
food deliveries on college campuses, such as the University of Pittsburgh and George
Mason University, where the services have been wildly popular. Deliveries require the
38 Vincent, James. “FedEx unveils autonomous delivery robot.” The Verge, February 27, 2019, https://www.theverge
.com/2019/2/27/18242834/delivery-robot-fedex-sameday-bot-autonomous-trials. Accessed February 3, 2020.
39 Ibid.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
22
recipient to enter a passcode in order to open the robot and receive the delivery. Other
companies, including Amazon Scout and Postmates’ Serve, have launched similar
projects designed to bring their packages as close as possible to a customer’s door.
Cities are already contending with the issue of dockless bicycles and scooters on
their sidewalks. Are they ready for the challenge of land-based delivery robots rolling
through traffic, running beside pedestrians, or using bike lanes? These vehicles raise
questions around speed, right of way, congestion, and privacy, and very few cities
have even begun to design policy for these bots. While it could be advantageous for
companies to “right-size” their delivery fleets instead of having large trucks moving
through dense communities, an army of small bots interacting with humans may not
be ideal either. Cities will therefore want to pay particular attention to developments
in the delivery sector. In preparation for these and other such developments, we turn
now to our recommended policy actions for cities.
COVID-19 Commentary
We see three potential impacts of COVID-19 on the initial AV use cases. First, the virus has
caused a significant increase in online shopping and food deliveries. The resulting deliv-
ery vehicle congestion is a serious concern especially as the economy and traffic return
from the shutdown. Given the continual progression in delivery technology, land-based
delivery vehicles could well be the first prominent use case of automated technology.
We do not currently see any evidence to suggest concerns regarding COVID-19 have sig-
nificantly altered the development trajectory of automated freight delivery.
Second, facing a world shaken up by the coronavirus, where we are encouraged
to social distance and avoid places that other people might have contaminated, it may
be easy to be skeptical of some of the scenarios we describe above. Certainly, we are
under no illusions the virus may deter some people from traveling in shared spaces
such as a TNC or an automated transit vehicle. That said, AVs may also offer a number
of benefits in our new world even for passenger service. Even without a risk of infec-
tion, a number of transportation experts pointed out that AVs might need separated
pods for shared rides so each passenger can enjoy better security. It would not be a
stretch to think that TNCs could incorporate a system that disinfects each pod after
a customer gets out, so it is clean and ready for the next rider. Moreover, while today
travelers might be concerned about getting into a car (whether a TNC or a taxi) with a
human driver who has spent the day being exposed to all previous passengers, they
may be more comfortable hailing a ride with no driver present making AV-TNCs more
attractive than conventional ones.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
23
Third, cities across the world face a major challenge of maintaining safety and
ridership on fixed-route public transit. Thankfully for those concerned about the pub-
lic transit fallout of the pandemic, early evidence shows public transit was not a major
cause of viral spread. Studies from Paris and Austria, for example, found no evidence
of infection clusters on transit vehicles, while cities that have reopened their transit
systems “have not seen subsequent infection spikes.”
40
A bus—even a busy one—
where riders silently browse their phones with their mouths closed may be a much
less risky place to be than a restaurant, bar, or concert venue.
Transit agencies, however, will still need to implement a number of policies to pro-
tect public health (such as ensuring frequent cleaning and perhaps mandating mask
usage), and AVs can be helpful in limiting the risk of spread. First, one strategy for reduc-
ing overcrowding on buses and trains is to increase frequency. Employee costs are a
major limiting factor in putting more vehicles on the roads or rails, so using automated
vehicles will reduce the cost of providing more buses, in order to reduce the number of
travelers per vehicle. Additionally, COVID-19 has taken a significant toll on transit work-
ers. As of mid-April, 100 American transit workers had lost their lives due to the virus,
41
and across the Atlantic, 29 bus drivers died from COVID-19 in London alone.
42
As drivers
spend a long time on a bus and in interaction with more people, they are more exposed
to the virus. Switching to AVs will mean buses can still run without putting drivers at risk.
It is still too soon to know exactly what impact COVID-19 will have on the AV
industry, but even if it delays deployment by months or years, we remain confident
that addressing the five policy actions outlined below will remain relevant for cities
looking to encourage sustainable, efficient, equitable transportation networks in the
years to come.
40 Sadik-Khan, Janette and Seth Solomon. “Fear of Public Transit Got Ahead of the Evidence.” The Atlantic, June
14, 2020, https://www.theatlantic.com/ideas/archive/2020/06/fear-transit-bad-cities/612979/. Accessed
January 23.
41 Beckett, Lois. “Revealed: nearly 100 US transit workers have died of Covid-19 amid lack of basic protections.”
The Guardian, April 20, 2020, https://www.theguardian.com/world/2020/apr/20/us-bus-drivers-lack-life
-saving-basic-protections-transit-worker-deaths-coronavirus. Accessed January 23, 2020.
42 “Coronavirus: London bus driver deaths to be reviewed.” BBC News, May 21, 2020, https://www.bbc.com
/news/uk-england-london-52752022. Accessed January 23, 2020.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
24
v. five recommended policy actions for cities
Given the current AV landscape, how should you prepare your city for the arrival of
AVs? Using insights we derived from our series of AV policy scrums and academic
research, we conclude that there are five policy action areas cities can take on now to
position themselves for successful adaptation to the AV environment.
Policy Action 1: Foster Mobility as a Service (MaaS)
The Concept
Over the last few years, cities across the country have published growth plans—such
as Boston 2030, Philadelphia 2035, or Louisville’s Plan 2040—meant to provide a
roadmap for development in the first half of the 21st century. These ambitious plans
often adopt a holistic perspective to urban growth, recognizing that transportation is
a critical component to providing citizens with access to education, healthcare, afford-
able housing and other resources. Moreover, city officials now recognize that commu-
nities built on car-centric transportation networks are not only environmentally and
financially unsustainable, but also limit opportunities to those who cannot afford or
otherwise choose not to own a car.
In the last decade, transportation mode choices for urban travelers have multi-
plied. The traditional options—private cars, bus and rail transit, taxis, personal bikes,
and walking—have been joined by TNCs, shared bikes, and shared scooters, modes
that have grown from zero to hundreds of millions of trips per year in a matter of only
a few years.
This expanded set of options gives travelers more choices when planning their
journeys, including some that might be better than what existed before—cheaper,
quicker, more convenient, and even more fun. However, the ever-growing number of
available modes can make selecting the “best” option a daunting task, particularly
when payment and trip planning requires interacting with several different services
independently: one transit card for the bus, another for the subway, a key fob to unlock
a shared bike, and a smartphone app to call a TNC.
A traveler’s preferred travel modes are also rarely static. Cycling in colder-weather
climates is a more attractive option on a beautiful day in May than a blustery one
in February. Transit might be a more convenient way to get to the store than return-
ing, when laden with a week’s worth of groceries. Traveling alone versus traveling
with young children may significantly influence mobility decisions. More broadly,
travelers are searching for a way to create a dynamic travel plan responsive to their
evolving needs.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
25
Mobility as a Service (MaaS) is a nascent concept that would combine trip plan-
ning, booking, and payment within a single platform. The likeliest implementation,
although not the only possible one, would be through a smartphone application. After
downloading such an app, a user would enter their journey origin and destination.
The MaaS platform would then provide information on all the best available transpor-
tation options, most significantly time, reliability and cost. This would also include
journeys that combine multiple modes, such as a subway ride connecting to a shared
bike rental for the “last mile,” or a TNC connecting to a commuter rail station. The
MaaS platform would also provide the ability to pay for any multi-modal trip in a single
transaction, or at the very least would provide a unified payment process, such as tap-
ping the same smartphone for payment at each leg of the journey. Alternatively, some
early MaaS proposals have dispensed with payments per trip altogether, moving to a
subscription-based approach where users pay a monthly fee in exchange for a set of
mobility options, such as unlimited transit and bikeshare plus ten TNC rides.
MaaS proponents promise to make planning and completing multimodal jour-
neys simple, fast, reliable and convenient. Indeed, they argue these systems will make
multimodal transportation so user-friendly that travelers may be persuaded to make
less use of private vehicles and more use of alternative transportation modes. As cit-
ies contemplate the potential of MaaS, they should view it as not only a transportation
tool, but one that accomplishes broader planning objectives, linking more people to
as many resources as possible.
None of the components of a MaaS
system are particularly novel. Consum-
ers have become accustomed to plan-
ning trips using mobile platforms like
Google Maps. Transit agencies across
the world have embraced real-time
data-sharing using the General Transit
Feed Specification (GTFS) data stan-
dard, which allow platforms like Google
Maps to provide directions not only for
driving and walking but also transit. In
the United States, apps like CityMapper
and Transit allow users to plan multi-
modal journeys with real-time arrival, departure, and availability data. Furthermore,
TNCs have clearly demonstrated the merits of app-based trip requests and payments.
Lyft, for example, acquired other mobility companies like Motivate and have begun
MAAS CASE STUDY: WHIM
• Early multi-modal transportation app
• Offers one-app trip payments,
real-time trip planning
• Launched by MaaS Global in Helsinki,
December 2017
• 70,000 registered users by 2019
(10,000 use regularly)
• Promoted as “one app for all your
transport needs”
• Pay-as-you-go and subscription mod-
els available
• Expanding to Antwerp, Belgium,
Birmingham (UK), Miami, Vancouver,
and Chicago
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
26
to incorporate the ability to reserve, unlock, and pay for vehicles like e-bikes and
e-scooters into their applications.
MaaS platform innovation will come primarily from the combination of all these
features, rather than the development of a host of new features—and that combina-
tion has the potential to be tremendously powerful. Travelers would have better infor-
mation on ways to reach their destinations, more available options to get there, and
an easier way to plan and pay for trips. Transportation operators could increase the
visibility and attractiveness of their services by combining their various modes into
journeys that might not have made sense alone. And public officials could tweak the
structure of any MaaS platform to better align with local mobility goals and priorities,
such as shifting travelers away from single-occupancy vehicles, making transit more
attractive, and ensuring equitable access to transportation services.
A MaaS system would not require the development of AVs. It could—and in some
places, does—exist even today, with current technology. But the arrival of AVs will
likely change the economics of TNC-type trips, making them relatively more attractive.
This would be true not only for individual rides but also for trips linked into a broader
multimodal trip platform like a MaaS system. This might attract travelers who currently
rely on their own personal vehicles to shift to a more multimodal MaaS-type system,
but could also incentivize travelers who currently rely predominantly on transit to shift
some of their trips toward this cheaper alternative mode. The pricing, structure, and
modal inclusion in a MaaS deployment could incentivize travelers to choose shared
AV rides, or transit, or single-occupancy vehicles. Which of these outcomes ultimately
occurs will depend on the specific implementation of a MaaS system. Moreover, cities
can influence passenger choice through financial incentives to choose public transit
or shared rides over single-occupant vehicles.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
27
The Landscape
MaaS Global was the first to launch a MaaS-type sys-
tem, known as Whim. Launched in December 2017 in
Helsinki, Finland, Whim still appears to be the most
developed MaaS deployment, with 70,000 registered
users in 2019, of which 10,000 regularly used the
app.
43
MaaS Global promoted Whim as the “one app
for all your transport needs,” and “the carefree way to
travel.”
44
Users in Helsinki can choose between a subscription-based model and a
pay-as-you-go service. For a monthly fee, subscribers get access to unlimited tran-
sit, bikeshare, and depending on the subscription tier, either discounted or unlimited
access to taxis and rental cars. Payment and ticketing are handled through the app.
Similarly, users who select pay-as-you-go options can use the app to plan, book, and
pay for trips across transit, taxi, and rental car services.
The service has since expanded to several other European cities, including
Antwerp, Belgium and Birmingham, United Kingdom, and the company expects to
launch in Miami, Vancouver, and Chicago in the near future.
45
MaaS Global’s CEO
noted that once launched, it would take three years for Whim to achieve profitabil-
ity within a city, provided that three to five percent of the eligible population use
the app.
46
While Whim was an early mover, perhaps the most ambitious companies in the
MaaS space are TNCs Uber and Lyft. Both companies have claimed that their overarch-
ing goal is to defeat the popularity of the private car, and in that vein, they want to make
it much easier for their customers to find the alternative mode that best suits them.
Travelers can now not only choose a car to pick them up, but also unlock bike-share
and e-scooters owned by Uber and Lyft through their apps. Touting the success of the
new initiative, Lyft president John Zimmer noted one in eight Lyft riders chose a bike
or scooter in the summer after the app redesign, in the cities where these services
43 Reid, Carlton. “Netflix-Of-Transportation App Reduces Car Use and Boosts Bike and Bus Use, Finds MaaS Data
Crunch.” Forbes, March 28, 2019, https://www.forbes.com/sites/carltonreid/2019/03/28/netflix-of-transportation
-app-reduces-car-use-and-boosts-bike-and-bus-use-finds-maas-data-crunch/?sh=36464e474923. Accessed Sep-
tember 17, 2019.
44 Shieber, Jonathan. “Whim, the all-in-one mobility app for ridesharing, public transit and rentals is coming to the
US.” Tech Crunch, January 24, 2019, https://techcrunch.com/2019/01/24/whim-the-all-in-one-mobility-app
-for-ride-sharing-public-transit-and-rentals-is-coming-to-the-us/. Accessed September 17, 2019.
45 Pohjanpalo, Kati. “Mitsubishi Invests in an App That Wants to Replace Cars.” Bloomberg, September 4, 2019,
https://www.bloomberg.com/news/articles/2019-09-04/mitsubishi-invests-in-mobile-app-whim-that-wants
-to-replace-cars. Accessed February 3, 2020.
46 Reid, “Netflix-Of-Transportation App Reduces Car Use and Boosts Bike and Bus Use, Finds MaaS Data Crunch.”
EXISTING MAAS APPS
CityMapper
Lyft
Moovel
Transit
Uber
Whim
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
28
were offered. And North America’s largest bikeshare system, New York’s Citi Bike, hit
a 100,000 daily ridership record in September 2020.
47
Going further, both apps now allow riders to assess real-time transit information
in the app directly. By employing transit agency route data and predictive software,
Lyft’s app shows where and when transit vehicles will arrive, and Lyft says 80 percent
of U.S. transit routes, along with their fares, are available in the app.
48
In Denver, Uber
allows customers to pay their transit fare through the smartphone application.
49
By
integrating transit, bikes, e-scooters, and cars, customers can see time and cost com-
parisons of the different services, helping them choose the service that works best
for them.
A number of other organizations have begun to roll out similar services. Transit
offers its customers the ability to book and pay for bike-share within its app. CityMap-
per offers its London customers a combined pass for unlimited transit and bike-share
usage, with an integrated fare card that allows the seamless use of both. Moovel offers
integrated payment options through its app for multimodal journeys in Stuttgart and
several other German cities.
Seeing the power of making multi-modal travel more convenient, some munici-
palities and transit agencies are working on their own MaaS platforms. In Maryland,
transit riders can search, book, and pay for their subway, light rail, and commuter
train fares on a Moovel-operated application known as CharmPass. The objective is
to make it more convenient for commuters in the Baltimore and Washington regions to
transfer between services. As the service runs on a smartphone, it is also more flex-
ible than traditional ticketing systems; for example, users can purchase a weekly or
monthly pass based on a specific start date, rather than on a fixed calendar.
50
Municipalities are also developing their own platforms. In May 2019, Louis-
ville, Kentucky’s transit authority became the first U.S. transit agency to release a
nearly-complete MaaS platform. Louisville’s app integrates transit, ride-hailing,
bike-share and e-scooters to encourage multimodal trips. For now, the app only allows
47 Dwyer, Matthew et al. “Shifting Gears: As Citi Bike Ridership Spikes, Share of Women Users Increases.” Sam-
Schwartz.com, October 12, 2020, https://www.samschwartz.com/staff-reflections/2020/10/7/shifting-gears
-citi-bike-demographics-change-with-ridership-increases. Accessed February 23 2021.
48 Ibid.
49 Pyzyk, Katie. “There’s an app for that: Transit agencies tackle MaaS platform development.” Smart Cities Dive.
June 26, 2019, https://www.smartcitiesdive.com/news/theres-an-app-for-that-transit-agencies-tackle-MaaS
-platform-development/557234/. Accessed September 17, 2019.
50 Moovel. “Moovel and MDOT MTA Power First Mobile Ticketing App for Baltimore Region.” September 28, 2018,
https://www.moovelus.com/moovel-and-mdot-mta-power-first-mobile-ticketing-app-for-baltimore-region/.
Accessed February 3, 2020.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
29
for trip planning, but the agency intends to eventually accept and process payments.
51
Other municipalities have explored providing data to and contracting directly with the
private sector to deliver these services, with developers taking a small percentage fee
from each ride.
Louisville, KY has developed its own app that integrates transit, ride-hailing,
bike-share and e-scooters to encourage multi-modal trips.
COVID-19 Commentary
As citizens worry about crowded places, a MaaS system can help spread travel both
temporally and geographically. MaaS platforms of the future might not just highlight
mode, travel time, and price, but also person-to-person congestion on a bike path,
bus, or (in some places like Manhattan or Central London) a busy sidewalk. When a
traveler enters their destination, the platform could suggest a mode, route, or alter-
nate time that would be less crowded. We understand many people will not be able to
take a private car for various reasons (and for a host of reasons we should be encour-
aging less automobile travel, not more), so giving citizens information as to how they
can travel safely will be a big benefit of a robust MaaS platform. While this platform
can exist without AVs, these new vehicles can make a multi-modal lifestyle more
affordable and reliable
The combination of MaaS and COVID-10 could also spur the use of active mobil-
ity. Studies show the most effective time to alter commuters’ travel habits is during
a period of change such as when they move to a new job. The pandemic has been a
period of change for Americans, and already urbanists are keen to capitalize on new
trends. Bicycle sales have risen during the pandemic. This provides an opportunity for
municipal governments to implement active transportation corridors, which might to
encourage longer-term shifts in how citizens get around. Adding a strong MaaS sys-
tem onto physical infrastructure can nudge travelers away from cars and on to bikes
and scooters by showing them convenient bike (and pedestrian) corridors. AVs will
assist with the transition, as people will be confident that they can find a ride later,
should a rainstorm arrive unexpectedly
Getting Started
As Whim’s early successes show, MaaS deployments will not require the arrival of AVs.
Cities can, and we believe should, pave the way for MaaS-type services to arrive even
51 Pyzyk, Katie. “There’s an app for that: Transit agencies tackle MaaS platform development.”
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
30
now. However, the arrival of AVs will likely make MaaS service offerings even more
attractive, particularly if the economics of TNC-style AV fleets allow transportation ser-
vice providers to significantly lower prices. This could increase mobility options, but it
could also lead customers to substitute ride-hailing for trips currently made on transit.
Considering these opportunities and concerns, there are a number of actions
that public officials can take today, both to allow for the development of MaaS and to
prepare for the responsible integration of AVs into a MaaS deployment.
Step 1: Build or Foster Data-Sharing and Interoperability Requirements
A successful MaaS system relies on data-sharing and interoperability among different
transportation providers. For example, Whim is able to offer one-app payments and
a real-time trip planning experience for transit, taxis, bike-share, and rental cars only
because these operators either make their data publicly available or because MaaS
Global has secured agreements with operators. Such agreements are not guaranteed,
however. In recent months, it has emerged that Lyft is no longer allowing the Transit
app to provide in-app bookings of New York’s Citi Bike (which is operated by Lyft’s
subsidiary, Motivate), an exclusion that could be a sign of broader actions and exclu-
sions to come. To ensure that MaaS systems can emerge, public officials could require
data-sharing and broader interoperability, including payments and bookings, as a
condition of operating in a city or region. Showing potential riders travel information
in a third-party app similar to many transit agencies’ trip planner platforms will liter-
ally put the bus on the map for prospective riders. Especially in areas where transit is
not as prevalent, this can show customers where transit trips are more convenient and
cost-effective, with a goal of increased ridership. Similarly, for city-run bikeshares,
providing information on station locations and—as much as possible—real-time avail-
ability reminds citizens of this option and makes it easier for them to choose a bike.
Municipalities can prepare for interoperability now by making sure their own data is
ready for a partner MaaS operator. (See Policy Action 4 for a more extended discus-
sion of data-sharing).
Step 2: Improve Coordination of Existing Public Transportation Services
While many of the modes that MaaS promises to integrate are operated by private firms,
many others, such as buses, urban rail, commuter rail, and more, are offered by pub-
lic sector organizations. In many U.S. cities and regions, these services are not coordi-
nated, requiring different fare cards, payment schedules, trip planning software, etc.
For example, in the San Francisco Bay Area, there are 27 different agencies that pro-
vide transit services, with inconsistent fare structures and confusing transfers yielding
a system that often operates in silos, rather than as one regional whole. The Bay Area is
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
31
not alone. From Chicago to New York to Los Angeles, public transportation services are
often provided by a variety of different agencies, leaving riders unable to find (or afford)
the trip that might be best for them. Thus, while a MaaS system would offer significant
benefits to consumers, public sector transportation providers can start realizing some
of those gains by coordinating their own offerings. Doing so would also pave the way for
private providers to participate in any larger regional coordination.
Step 3: Ensure MaaS Aligns with Regional Mobility Goals
A common concern regarding MaaS is that offering existing transit riders a taxi or TNC
option might divert them from public transportation. If this effect were sufficiently
pronounced, it might stymie efforts by public officials to improve mobility and reduce
overall traffic congestion and pollution. We recommend that policymakers consider
their broader regional mobility goals when setting up, or enabling, any MaaS system.
If increasing transit ridership is a regional priority, then MaaS systems should be
encouraged, or even regulated, to monitor overall mode-shares and tweak the basket
of available mobility options if it appears too many users are diverting from transit.
Or, if policymakers want to reduce greenhouse gas emissions due to transportation,
MaaS systems could be designed accordingly, to prioritize low-carbon activities like
biking and/or to disincentivize single-use car rides as opposed to shared rides. A MaaS
system could incentivize both, or neither, of these outcomes, or any others. It will be
incumbent on public officials to weigh the tradeoffs between consumer responsive-
ness and overall mobility goals.
Policymakers also need to consider the equity impacts of any MaaS system.
These could be significantly positive, by providing better transportation access in
places that might not previously have had it, but that is not guaranteed. For example,
although we anticipate the majority of travelers will access a MaaS app through their
smartphones, smartphone ownership is not universal. This could leave some trav-
elers—particularly elderly and low-income residents—unable to participate, unless
other options like a “Dial-a-ride” system are made available. As with AV-TNC opera-
tions more broadly, MaaS providers must offer equitable geographic coverage of ser-
vice and pricing.
Public officials will also need to be aware of smaller equity impacts. For instance,
poverty advocates have noted that ‘monthly passes’ are unfair to lower-income tran-
sit riders, because wealthier commuters can save money overall by having enough
money to buy up front, whereas those who must pay on a ride-by-ride basis enjoy no
similar discount. Positively, transit fare card technology (such as London’s Oyster card
and Boston’s CharlieCard) has improved such that it is now possible to cap charges
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
32
once a person has paid the equivalent cost of a monthly pass. Any MaaS subscription
model should have a similar pricing structure. Additionally, should your transit agency
already offer a low-income transit pass, the MaaS platform will need to be able to
accommodate that provision.
Quick Notes
The Concept
• Travelers are frustrated as they try to piece together travel options to find the lowest
cost and/or fastest combination of travel modes.
• MaaS offers a way for travelers to plan and pay for their trips quickly, easily, flexibly
and through a single vendor.
• MaaS will make multimodal transportation so user-friendly that some travelers will
eschew private vehicles in favor of alternative transportation modes.
• MaaS builds on the success of the TNCs’ ride-hailing software.
• MaaS apps can spur use of micro-mobility, which may help reduce vehicle
congestion.
The Landscape
• The first MaaS system, Whim, launched in 2017; by 2019, it was available in four
European cities.
• Other platforms are appearing in Germany and England.
• Some municipalities and U.S. transit agencies are developing their own MaaS platforms.
• Louisville, Kentucky is the first transit authority to launch MaaS in the U.S. integrat-
ing transit, ride-hailing, bike-share and e-scooters.
Getting Started
1. Build or foster data-sharing and interoperability requirements.
2. Improve coordination of existing public transportation services.
3. Ensure MaaS aligns with regional mobility goals.
Policy Action 2: Rethink Curb Design and Street Space Allocation
The Concept
For most of human history, streets were for everyone—pedestrians, cyclists, and
those riding in horse and carriage. In the twentieth century, however, this changed. A
growing number of pedestrian fatalities had U.S. automobile companies realize that
to keep everyone safer (while not banning cars or cutting their permitted speed), they
needed to separate people from vehicles. The raised curb was born. Over time, cities
have installed some separated bus and bike lanes, but the principle of road use sep-
aration and automobiles’ supremacy remains predominant. The lane beside the curb
has become the “parking” lane, providing a space for car storage out of the way of
traffic, with cities sometimes charging a nominal fee for this service.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
33
Now, several pressures—including traffic congestion, increased demand for curb
use and the proliferation of delivery vehicles and TNCs—are making city planners
rethink the use of the curb. Indeed, TNCs are a particular pain point as they have led
to increased traffic volume while simultaneously redefining the use of the curb as a
pick-up and drop-off (PUDO) point for their services. Reacting to this change in travel
demand, already cities have begun to institute designated TNC PUDO zones. As we
consider a shift to AVs, cities can leverage the advantages of the technology to imple-
ment restrictions on where to stop and have confidence that AVs will respect the rules
regarding assigned use of street space.
In addition, it is reasonable to expect that these vehicles, which will be con-
stantly scanning their environment, could be programmed such that if one “sees” an
opening in the curb that’s available, it could alert other vehicles that may be searching
for a drop-off point. Alternatively, as they are connected, they could all communicate
with one another or with a central system, booking space along the curb in an orderly
manner (a measure that would be especially useful for deliveries). The combination of
these factors has planners rethinking and redesigning curb space usage. For instance,
to discourage the use of personal cars in the urban core, cities are increasing the cost
of parking, which serves as a financial disincentive for personal auto use. Some are
also reducing the number of parking spaces, purposely making it more difficult for
drivers to park to discourage them from driving into the city center.
AVs will accelerate the opportunity to rethink both how the curb is used and
how our street space is allocated. As AVs require less space, automobile lanes can be
removed or narrowed in favor of buses, bikes, and wider sidewalks, or even vibrant
public spaces such as parklets and patios. In some areas, cities can prohibit cars out-
right or drastically reduce speed limits. This becomes easier with AVs, both because of
the vehicles’ requirement to obey these laws and because the expectation is AV riders
will be able to occupy their time while in the car, meaning they will be less put out by
slower travel speeds or a need to reroute around a dedicated bus- or pedestrian-only
street. Moreover, if shared ride fleets become an initial use case, many existing park-
ing spaces can be repurposed.
AVs also raise a challenge for municipalities. If the vision of AVs replacing private
vehicles becomes a reality, then parking revenues associated with private cars will
decline. In many cities, street parking is a multi-million-dollar revenue source. Initial
ideas on replacing parking revenues center on shifting the idea of charging a fee to
warehouse cars along the road to charging a pick-up/drop-off access fee for using the
curb. Shifting the mindset of drivers, delivery services and local residents from inex-
pensive street parking to a repurposed curb, ranging from parklets to transit lanes to
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
34
paying for access, will take time. Acting now will likely help to match the arrival of AVs
with these changed mindsets. Depending on its structure, curbside pricing could be
incorporated into a broader congestion pricing plan rather than being a standalone
entity (as is discussed further in Policy Action 5).
The Landscape
While pricing structures may vary from one municipality to another, the potential of
AVs, coupled with an already shifting transportation demand landscape, provides cit-
ies with the opportunity to dramatically re-conceptualize the streetscape. We focus on
two major areas for consideration: curbside management and street space allocation.
Curbside Management
In January 2018, the National Association of City Transportation Officials introduced
a program, SharedStreets, to foster mapping of curbs. As reported in Wired: “Shared-
Streets may be exactly what both sides need. First, it will establish data standards for
curbs, traffic speeds, and transit data, formats that can be shared between companies,
agencies, even across cities. (No more, My computer can’t open that file.) Now, there’s a
common language for curb data and maps, with agreed-upon locations for curb cuts and
intersections.”
52
Uber is sharing data to help populate the curb profiles and, nationwide,
cities have begun to institute designated zones for pick-up and/or drop-off for TNCs.
Washington, D.C.’s Approach
Washington, D.C., offers some of the most ambitious examples of rethinking the curb to
reflect new priorities and demand. In D.C.’s Dupont Circle, officials noticed major con-
gestion on weekend nights as TNC drivers blocked the travel lane to pick up passengers.
As such, the City undertook a trial, turning 45 parking spaces into a pick-up/drop-off
zone from Thursday to Saturday after 10:00 p.m. These zones are conceptually similar
to taxi stands, but also involve restrictions on pickups outside of the designated zones
within a specified area of the city (D.C.’s zones allow both TNCs and commercial deliver-
ies to use the zones).
Results were favorable, with local businesses reporting “a decrease in curb access
conflicts, and anecdotal evidence points to reduced dwell times for ride service vehi-
cles.”
53
In August 2019, the government also removed parking in 12 zones to create des-
ignated commercial pick-up/drop-off spaces in busy areas; while such loading zones
are not novel, what is new is the government’s partnership with the startup curbFlow,
which provides a smartphone application through which registered delivery drivers can
reserve a spot in the zone. In November 2019, Columbus, Ohio, followed suit with its
own similar test.
52 Marshall, Aarian. “Uber Makes Peace with Cities by Spilling Its Secrets.” Wired, April 16, 2018, https://www
.wired.com/story/uber-nacto-data-sharing/. Accessed February 3, 2020.
53 International Transport Forum. The Shared-Use City: Managing the Curb. OECD, 2018, p. 55, https://www
.itf-oecd.org/sites/default/files/docs/shared-use-city-managing-curb_3.pdf. Accessed February 3, 2020.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
35
Cities are also experimenting with one-time or event-specific restrictions on TNC
pickups and drop-offs (such as for sports events or parades). These zones have gen-
erally been successful, leveraging the TNCs’ technology platforms to direct both pas-
sengers and drivers to the designated zones. Fort Lauderdale, Florida, officials found
that such zones were highly utilized and served to improve traffic flow and reduce
delays.
54
As we’ll see below, other cities have also begun to look more closely at curb-
side management.
Boston’s Approach
Boston has ambitious goals to reduce the number of cars on its roads, reduce conges-
tion and boost alternative transit modes—including AVs. Boston began investing in curb
management in 2013 when it launched its Complete Streets initiative, a transportation
policy and design approach to creating safe and convenient transportation and access
for all visitors and residents across a myriad of travel options. Roads are designed for
multimodal use and designers assume that streets are shared by all—and not dominated
by cars. As such, Complete Streets incorporates features like street trees and green
design elements, intelligent signals, “smart” meters, wayfinding and provisions for elec-
tric vehicle sharing.
55
This initial effort was bolstered by Boston’s umbrella GoBoston
2030 long-term mobility plan. Its Vision and Action Plan, released in March 2017, noted
Boston’s aspirational goals are to, by 2030:
• Increase public transit use by 33 percent,
• Drive a fourfold increase in bicycle use, and
• Cut the number of single passenger vehicle rides in half.
56
Boston also aspires to double the number of vehicles providing shared transportation by
2030.
57
Curb management has emerged as an important component for achieving these
goals, as pilots are beginning to show.
In February 2018, Boston’s mayor announced the results of a related program: Per-
formance Parking. During the year-long pilot, the City raised parking meter rates to learn
if doing so would reduce congestion, increase roadway safety and make finding an
on-street parking spot easier for drivers. Raising meter rates was proven to open up more
parking spaces for residents and business customers, and reduce congestion caused
by illegal parking. The city’s Back Bay neighborhood realized an 11 percent increase in
available metered spaces and a 14 percent drop in double parking.
58
54 City of Fort Lauderdale. Las Olas Boulevard Six-Month Safety Improvements Demonstration Project. July 2018,
https://www.fortlauderdale.gov/home/showdocument?id=31059. Accessed October 8, 2019.
55 “Boston’s complete streets,” City of Boston, https://www.boston.gov/departments/transportation/bos-
ton-complete-streets, accessed September 18, 2019.
56 Go Boston 2030 Action Plan. “Goals and Targets.” City of Boston, https://www.boston.gov/sites/default/files
/document-file-03-2017/go_boston_2030_-_4_goals_and_targets_spreads.pdf. Accessed September 18, 2019.
57 Ibid.
58 Office of Mayor Martin J. Walsh. “Results of Performance Parking Pilot Announced.” City of Boston, February 21,
2018, https://www.boston.gov/news/results-performance-parking-pilot-announced. Accessed September 18,
2019.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
36
Seattle’s Approach
Seattle has been transitioning from traditional curb parking to flex-zones since 2016.
Curb usage is defined by type of street and local needs. Commercial streets are prioritiz-
ing freight delivery and passenger pick-up/drop-off over private car parking. Other uses
are dedicated transit and bicycle lanes. For example, the flow along Rainier Avenue was
changed from two lanes in each direction to one lane in each direction, one lane in the
center dedicated to left turns and one lane dedicated to bus transit. While passenger
vehicles incurred a one-minute increase in travel times, transit and bicycle travel times
were reduced substantially.
Street Space Allocation
As cities begin to rethink the curb, a natural next step is to reconsider how they are
allocating their street space. If AVs enable traffic flow efficiency gains, cities may be
able to create an environment more accommodating to the needs of non-drivers with-
out causing a significant impact on the travel times of those in vehicles. In fact, cities
can move right now to take greater control over their street space allocation such that
they will be more prepared to match the arrival of AVs with a new mindset.
New York City’s Approach
In New York City, a pilot program is underway that turns curbside parking spots into
neighborhood loading zones from 7:00 a.m. until 7:00 p.m. on weekdays, and trans-
portation officials “have expanded loading zones in commercial areas in recent years,
creating about 2,300 new zones around the city.”
59
In addition, New York has completely
banned cars from dozens of areas in order to prioritize a safe pedestrian experience.
Meanwhile, New York is also assembling a new freight master plan. While much of the
discussion around pick-ups and drop-offs has recently focused on TNC activity, pack-
age delivery is having an equally profound effect on North American road networks. In
Manhattan, surveys show that it can take delivery drivers up to an hour to find a parking
space. As a result, they often end up double-parking, or need to park for long periods of
time in certain areas. It can take up to an hour to complete deliveries to a single high-rise
office or apartment tower, for instance, which means the truck is on the street all that
time, taking up space.
Given these congestion concerns, and as research has shown that moving to nighttime
delivery more broadly could cut freight-related emissions by almost two thirds,
60
New York
has launched a program to shift more deliveries to the late-night hours, with 500 compa-
nies voluntarily committing to deliver their goods between 7:00 p.m. and 6:00 a.m.
61
It
might not be reasonable to expect human drivers to deliver at all hours, but bots can.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
37
COVID-19 Commentary
As Americans struggled to maintain six feet of separation from strangers, during the
pandemic it suddenly became all too obvious how little space most cities allocate to
pedestrians; many sidewalks are not even wide enough to accommodate two peo-
ple passing with six feet between them. To make physical distancing possible and
to provide transportation options for those who do not own a car but did not want to
take transit, cities closed streets to vehicles in favor of active transportation corridors.
With indoor restaurant dining restricted, officials also allowed restaurant operators to
expand their patios to sidewalks and parking spaces. Some cities have closed whole
streets or squares to cars to accommodate restaurateurs. Though not yet as common,
it is also realistic to think cities may open up street space to retailers, so customers
can browse their wares outside rather than in small shops. Lastly, many urban plan-
ners have pointed out that if health officials do not want people crowding in parks,
cities need to provide a lot more ordinary open space where residents can be outside
without bumping into one another—and that means turning abundant roadways over
to public uses instead of reserving them all for cars.
Longer term the current changes to curb use can provide an effective bridging
strategy for the arrival of AVs for passengers and/or delivery. Once citizens see the ben-
efits of less auto-focused curb usage, it will be an easier step for policymakers to create
more road space for dining and retail plazas or loading zones for delivery vehicles. More-
over, AVs will always obey the law, ensuring cars stay out of bus or bike lanes; do not
accidentally turn into a street where they are prohibited; and in any areas where vehicles
and pedestrians share street space, will stick to lower speed limits so walkers are safe.
Additionally, employing technology to match travelers with TNCs can potentially
reduce congestion on the sidewalks. Rather than having many people congregate in
one area (say, around the door of a restaurant), a system that manages AVs can direct
them to spread out along a series of blocks, with riders having to walk a short distance.
Getting Started
Rethinking the roles of curb and street can enhance the urban environment now and
prepare for the arrival of AVs tomorrow. The following are ways to tap the value of
repurposing the curb today while also positioning this valuable asset for generating
greater value in the future.
Step 1: Map the Curb
The starting point for enhancing the value of the curb is understanding how it is used
today. Very few municipalities have an inventory of how curb space is allocated. How
much is dedicated to parking, freight zones, transit and bicycle lanes, etc.? Fewer have
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
38
a quantitative understanding of how those intended spaces are utilized. When one
thinks about the miles of curbs and the multitude of uses over time, it is not surprising
that curb inventories are rare. Given the importance of understanding curb usage, a
number of organizations are enabling mapping the curb. Shared Streets is partnering
with cities to create an open data standard for the curb. Several start-ups are offering
mapping tools that track the locations of curb cuts, fire hydrants, bus stops, parking
meter and other curb infrastructure and usage. The maps can be enhanced with data
from AVs captured as they operate and “map” the environment.
Step 2: Establish a Prioritization Framework
Repurposing the curb is controversial. Constituents are used to parking in front of their
residences, finding parking close to retail establishments, and having door-to-door
TNC service. Commercial establishments want parking for their customers and access
for deliveries. Municipal governments count on parking revenues. Repurposing away
from traditional uses will likely be met with opposition. Establishing a prioritization
framework and a supporting rationale will provide a clear structure for decision-making
and hopefully reduce opposition to the change. The framework should support the
municipality’s long-term vision and prioritize the gamut of curb-use alternatives, from
transit and bicycles to parklets and traditional parking. Priority ranking should reflect
local geography and needs.
Equity is a critical consideration when establishing the framework and thinking
about where you would prohibit cars. Many cities are closing streets or limiting parking
on high streets or in more dense urban areas. This makes sense from a congestion and
efficiency perspective, but these areas are also often where lower-income residents
live—whereas many higher-income people live in less-congested suburbs. It is import-
ant when adding bike lanes or limiting parking that you address any impacts on citi-
zens who might struggle to, say, bring home their groceries if there is no place to park.
Additionally, limiting traffic and/or parking on high streets may lead to pass-through
or parking spillover to adjacent residential ones. If these are also in low-income areas,
officials need to consider—and likely mitigate—any impacts.
Finally, adding trees, bike lanes, and other public amenities often makes certain
areas more desirable. Poverty advocates have sometimes even opposed bike lanes
because they are the first step towards displacement of low-income and minority res-
idents. We do not believe the solution is to not make investments in transportation
infrastructure or policy, but instead need to use robust community consultation, clear
metrics to measure any negative impacts, and have a strategy in place to mitigate the
impacts as necessary.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
39
Step 3: Pilot Alternative Curb Uses and PUDO Zones
The best way to explore alternative uses for the curb is to design a pilot project or two.
The objectives of a pilot are (1) to test if a change in curb usage generates the desired
outcome and (2) to understand the political and operational issues of changing the
role of the curb. Ideal pilots are those where there is a direct opportunity to address
a pressing problem such as safety concerns, congestion or even double-parking. The
pilots should also be:
• Limited in scope,
• Limited in their funding requirements,
• Capable of providing measurable outcomes in the short-term, and
• Reversible.
Examples of pilot projects include substituting parklets for parking in Boston’s
Back Bay neighborhood and dedicated pick-up/drop-off sites in Washington’s Dupont
Circle neighborhood. For those cities where AVs are being tested, explore a pilot that
leverages learning from this new mobility option.
Much like Washington, city governments can undertake a plan to release parking
spaces for drop-offs in areas of high demand. Future trials can potentially go further
than those in the nation’s capital, though, in that passengers calling a car within a
certain radius of a PUDO zone will be actively directed, through their app, to walk to
the PUDO zone where the car will meet them. The purpose behind this more ambitious
measure is to reduce any risk of double-parking or other dangerous maneuvers in
traffic as drivers attempt to reach their customer. This concept has been successfully
tested and will likely be rolled out more broadly.
The city will need to determine the appropriate placement and spacing for PUDO
zones. The zones should have enough space for vehicles to pull in and out quickly and
with little back-up. They will also want curb cuts so passengers of all ages and abilities
can board and disembark safely. They will need to be clearly marked so passengers
and drivers do not become confused or lost as they search for them, and others do not
mistakenly park in them.
PUDO zones do not need to be permanent but can become parking or loading
zones when PUDO demand is low. As much as possible, cities should gather data on
curb and PUDO pilot projects to understand both successes and challenges. If pilots are,
on balance, helpful with reducing congestion, they can become a permanent measure.
Step 4: Establish a Curb-Use and Street Space Allocation Master Plan
The curb inventory, prioritization protocol, results of pilot projects and the experience
of curb reuse and street allocation efforts in other cities provides the raw material for
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
40
developing a curb use and street allocation master plan. The master plan provides
all stakeholders with an understanding of how the curb and street will be used in the
future, fostering better decisions ranging from residential and commercial investments
to car ownership. The starting point for the master plan is a set of guiding principles as
to how the plan will be constructed, which should be developed in collaboration with
stakeholders. Using these principles, the plan should detail street-by-street usage. As
with all plans, it will evolve over time, but having a clear starting point is a key first step.
Quick Notes
The Concept
• Traffic congestion, increased demand for curb use and the proliferation of delivery
vehicles and TNCs are making city planners rethink the use of the curb.
• Cities have begun to institute new uses for the curb, from designated zones for
pick-up and/or drop-off for TNCs to replacing parking with dedicated transit lanes
or parklets.
• To discourage the use of personal cars in the urban core, cities are increasing the
cost of parking and reducing the number of parking spaces, purposely making it
more difficult for drivers to park.
• Unlike drivers, AVs can be programmed to follow curb-use rules.
• If shared-ride AV fleets become an initial use case, many existing parking spaces
can be repurposed.
The Landscape
• Washington, D.C., is exploring dedicated pick-up/drop-off locations by time of day.
• Boston is experimenting with raising parking fees, removing parking spaces to
reduce incentives for personal vehicles, and substituting parklets for parking.
• Seattle is testing flexible curb usage based on local context and stakeholder needs.
• New York City is piloting curbside loading zones and late-night deliveries.
Getting Started
1. Map the curb.
2. Establish a prioritization framework.
3. Pilot alterative curb uses and PUDO zones.
4. Establish a curb use and street allocation master plan.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
41
Policy Action 3: Manage and Reduce Congestion
The Concept
In many growing cities across the United States,
congestion is on the rise. Transportation experts
typically present three main strategies for miti-
gating congestion: encouraging increased uptake
of shared rides such as those offered by TNCs;
promoting increased walking and usage of public
transportation options such as buses, subways,
and bikeshares; and/or implementing congestion
pricing programs.
With their potential for shared rides, TNCs
were initially pitched as a way to address gridlock,
but research has found putting ride-hailing cars on
the road is often just increasing traffic by drawing people off transit and filling the streets
with passenger-less vehicles. Coupled with contemporary urban growth patterns (i.e.,
sprawl) and natural population increases, American roads are, in many places, more
clogged than ever. As AV travel will be less expensive, both economically and tempo-
rally, there is the potential for commuters to shift to automated single-occupant AVs or
to live farther from their places of work. Another concern is the specter of cities clogged
with “zombie” cars: automated vehicles that drive through the city awaiting a passen-
ger. If the initial use case for AVs follows the TNC single passenger model, then the intro-
duction of AVs is likely to make congestion worse, not better.
Fortunately, this need not be the case. AVs offer cities the opportunity to devise
toll pricing that encourages ride sharing. For instance, cities can adjust the toll based
on the number of people traveling in an AV or even waive the fee entirely if there are
three or more passengers. Alternatively, cities can employ financial penalties for
passenger-less vehicles or those with only a single passenger.
Beyond incentivizing ridesharing, AVs might also provide a boost to public trans-
portation. Just as passengers who know that they can reliably call on a TNC through-
out the day when needed might be encouraged to utilize public transportation with the
security of a vehicular backup on demand, AVs may give rise to more multi-modal travel-
ers. For instance, a commuter that previously might have driven to all stops may instead
choose a bikeshare for their morning commute, take a subway to a store in the after-
noon, walk back to the office, and use an AV to get home in the rain after dark. As AVs
lower parking demand, some researchers argue that urban density can increase, which
makes walking, cycling, and transit more appealing and, in turn, lessens congestion.
AVS: POTENTIAL BENEFITS
TO CITIES
• Less traffic congestion via
toll pricing that encour-
ages ridesharing
• Better public transporta-
tion as multi-modal trips
become more enticing
• New revenue stream via
AV access fees with easy
tracking
• Increased traffic-flow
control via deployment of
congestion pricing
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
42
As we ponder scenarios in which congestion is either exacerbated or alleviated,
we must acknowledge that the key factor will be the degree to which thoughtful pol-
icymakers are willing and able to seize the opportunity of a shifting transportation
landscape to improve flow within cities. In the absence of effective public policy, we
will likely see little improvement in congestion. Policymakers do not need to wait for
the widespread adoption of AVs to prepare for the transportation future they want to
build for their communities, as policies can be developed for TNCs today that will be
applied to AVs in the future. The presence of a driver is irrelevant when considering the
primary policy lever for mitigating congestion: congestion pricing.
One avenue for congestion pricing, as discussed above, involves the potential
to incentivize shared rides. It’s important to note that, in contrast to TNCs, tolling
AVs will potentially be less expensive and easier to assess and enforce because AVs
have real-time, highly-accurate continuous location monitoring. For instance, if a city
decides to charge a fee each time a vehicle enters the city center, AVs will have the
ability to automatically incorporate these fees into their pricing structure.
AVs offer cities another advantage: because AVs collect data on every aspect of
their movement, cities will not need to pay for observational infrastructure such as
cameras to collect tolls. Instead, they can use a software-based query to identify all
instances of AVs entering the city center and simply invoice the AV operators.
Distance-based fees, which assess tolls based on how far a vehicle travels, could
also be used to incentivize short trips. Route-based fees can be used to price discrimi-
nate, offering lower fees for AVs that travel less-congested routes, for instance. In gen-
eral, with AVs, cities will have a greater set of tools to use to influence driving behavior
and congestion.
The Landscape
Cordon-based congestion pricing has already met with success in London and Stock-
holm. London, for instance, charges a daily fee of the equivalent of $14.35 to enter
the city center while Stockholm’s charge changes depending on the time of day,
capped at around $12.50. Both cities found trips into the city center decreased and
about half of those were redirected to public transit. In addition, a study of Stock-
holm’s air quality after the implementation of congestion pricing found a five to 15
percent decrease in pollution.
62
62 Hawkins, Andrew J. “Why Congestion Pricing Can Save Cities from Their Worst Possible Future.” The Verge,
March 29, 2019, https://www.theverge.com/2019/3/29/18286830/congestion-pricing-nyc-gridlock
-autonomous-vehicles-traffic. Accessed September 19, 2019.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
43
A key policy question is this: who is subject to the congestion fee? Some experts
support universal congestion pricing—levying a fee for all vehicles entering a certain
area of a city. Others suggest the fee should be focused on the TNCs, who are cur-
rently viewed as adding incremental congestion to urban streets. (Currently, TNCs are
exempt from the fee in London, which many policymakers now acknowledge is one
reason for increasing congestion in the city’s core.) TNCs, fearing that their services
will be disproportionately tolled, point out that making ridesharing more expensive
will encourage people to drive themselves, putting more cars on the road.63
Local traffic patterns and politics will likely shape the answer to the “who pays”
question, however equity will be a second key consideration as part of a conges-
tion pricing strategy. As it becomes more attractive to live closer to work thanks to
the higher cost of travel, some lower-income residents could be at a higher risk of
displacement. As with other policies above that can also lead to gentrification, it is
important to have a strategy in place to address this risk. Additionally, in many cities
lower-income residents have already been forced out to poorer suburbs as neighbor-
hoods closer to the core have become more attractive over the last few years; these
residents may therefore face a higher financial travel burden under congestion pricing.
Officials will need to mitigate this in some manner.
One approach would be to adopt a variation on the Canadian federal govern-
ment’s carbon tax model: Everyone pays the carbon tax on the energy products they
use, but enjoys a credit on their income tax. For cities looking at congestion pricing,
they could offer a similar rebate for qualifying low-income residents.
64
Done properly,
this would leave the disincentives to travel by car in place, while lessening the burden
on those least able to pay.
New York City’s Approach
In the United States, New York City is planning a congestion pricing program that uses
an electronic tolling system to charge drivers entering parts of Manhattan. The fees are
designed to both reduce personal auto use in Manhattan and to raise funds to improve
transit. When implemented in 2021, New York will be the first U.S. city to enact conges-
tion pricing but others, such as Portland, Seattle, and Los Angeles, are commissioning
studies and looking to New York to see if leaders have the political support to follow
through. On a state level, Oregon has a voluntary pricing plan, where drivers can opt in
to paying a small fee for each mile they travel.
63 George, Tyler. “Universal congestion pricing is the way to go.” CommonWealth, September 15, 2019, https://
commonwealthmagazine.org/transportation/universal-congestion-pricing-is-the-way-to-go/. Accessed Septem-
ber 19, 2019.
64 Macdonald, Neil. “Justin Trudeau’s carbon tax is revenue neutral... for now.” CBC News, April 24, 2019, https://
www.cbc.ca/news/opinion/carbon-tax-revenue-1.5107979. Accessed January 23, 2021.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
44
Washington, D.C.’s Approach
While these programs all employ static costs, the true potential of congestion pricing
is that it can change based on real-time demand. Washington, D.C.’s 66 Express Lanes
project shows what such an initiative might look like. Implemented in late 2017, the high-
way’s tolls cost an average of $12.59 for a two-way trip but can sometimes increase to
well over $40.
65
Tolls fluctuate every six minutes to maintain a minimum average speed
of 55 miles per hour.
The results of the program have been mixed. Carpooling (exempt from the toll)
increased by 15 percent in the first months of the program and travel times decreased
during rush hour, but off-peak hours and alternative routes have seen increased traffic.
66
Reports show that where transportation officials put in better bus service alongside the
toll, transit use increased. Where better transit options are not available, however, rid-
ership has declined, and one of the biggest criticisms of the initiative is that the gov-
ernment forced car commuters to pay without offering any reasonable alternatives to
driving.
67
The D.C. case study illustrates that to make congestion pricing work, cities must
offer alternatives to driving. In London, the transit system offers faster travel than driv-
ing. About half of the 100,000 drivers in Stockholm who “disappeared” after the city’s
pricing plan went into action shifted to transit.
68
In California, tolling on the I-15 near
San Diego successfully led to a nine percent increase in transit ridership.
69
Improving
transit is therefore a critical component of the congestion management recipe—and
should begin well before implementing road pricing.
As TNCs have reduced transit demand in some cities, some politicians suggest it
is time to rollback transit. This is problematic from a number of perspectives, includ-
ing the impact it could have on lower-income residents who might not be able to afford
ridehail services. Instead, cities should continue to fund and expand public transpor-
tation options. They can also provide low-income transit passes, and where they have
65 Smith, Max. “I-66 tolls turn 1 year old – Are they working?” WTOP, January 4, 2019, https://wtop.com/dc-transit
/2019/01/average-i-66-price-speeds-from-first-year-of-tolls-and-extended-hov-hours/. Accessed February 3,
2020.
66 Lazo, Luz and John D. Harden. “Year-old 66 Express Lanes have caused shifts in commuter behavior, but not
necessarily in ways officials hoped for.” The Washington Post, December 8, 2018, https://www.washingtonpost
.com/local/trafficandcommuting/year-old-66-express-lanes-have-caused-shifts-in-commuter-behavior-but
-not-necessarily-in-ways-officials-hoped/2018/12/08/6e78d944-e832-11e8-a939-9469f1166f9d_story.html.
Accessed February 3, 2020.
67 Smith, Max. “Are I-66 tolls promoting transit use?” WTOP, June 6, 2018, https://wtop.com/dc-transit/2018/06
/are-i-66-tolls-promoting-transit-use/. Accessed February 3, 2020.
68 Tools of Change. “Stockholm’s Congestion Pricing.” Tools of Change Highlights Series, October 2014, https://
www.toolsofchange.com/userfiles/Stockholm%20Congestion%20Pricing%20-%20FINAL%202014.pdf.
Accessed February 3, 2020.
69 U.S. Department of Transportation/Federal Highway Administration. “I-15 Congestion Pricing Project Monitoring
and Evaluation Services.” https://ops.fhwa.dot.gov/congestionpricing/value_pricing/pubs_reports
/projectreports/interst15_congestion.htm. Accessed April 10, 2020.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
45
a stake in the bikeshare program, offer subsidized memberships on those systems as
well. Hamilton, Ontario, for example, offers a three-month bikeshare pass for $3 to
qualifying residents.
Wrapped up in congestion pricing and better transit options is undoubtedly
also the topic of land use. Many cities have segmented uses, but if people can live
in the same neighborhoods as they work and play, it will cut transportation demand.
Done properly, mixing uses while permitting gentle density (that we define as build-
ings between three and six storeys) and implementing policies that encourage more
affordable housing are all key components to a holistic, inclusive travel demand man-
agement strategy.
Seattle’s Approach
Seattle is the golden example of transit performance in the United States. Beginning
in 2014, Seattle increased service by 270,000 hours annually, boosting the number of
households within a ten-minute walk of ten-minute or better service from 25 percent to
64 percent.
70
Seattle has focused on a network system to make travel easier, rerouting
bus routes as new rapid transit comes online, increasing frequency rather than simply
focusing on coverage.
71
From 2000 to 2016, downtown transit increased from 29 percent
to 47 percent of trips, while drive-alone trips declined from 50 to 30 percent
72
—showing
that transit planning can be an effective congestion management strategy in itself.
While Seattle increased transit funding, other cities, including Houston, Texas,
and Columbus, Ohio, famously redesigned their networks to do more with the same
amount of money. Houston, for example, now devotes 80 percent of the transit agency’s
resources to maximizing ridership (versus 20 percent to providing access to people liv-
ing in low-density, expensive-to-serve areas).
73
Such an approach is not limited to big
cities either, as Richmond, Virginia, revamped its own network to focus on frequency,
and enjoyed a ridership increase of 17 percent.
74
70 Bliss, Laura. “Why Seattle Is America’s Bus-Lovingest Town.” CityLab, May 11, 2018, https://www.citylab.com
/transportation/2018/05/seattle-the-city-that-respects-the-power-of-the-bus/559697/. Accessed February 3,
2020.
71 TransitCenter. “There’s a Reason Transit Ridership Is Rising in These 7 Cities.” TransitCenter, February 27, 2019,
https://transitcenter.org/theres-a-reason-transit-ridership-is-rising-in-these-7-cities/. Accessed February 3,
2020.
72 SDOT Blog. “A closer look at Seattle’s rising transit ridership.” Seattle Department of Transportation, January 3,
2018, https://sdotblog.seattle.gov/2018/01/03/a-closer-look-at-Seattle’s-rising-transit-ridership/. Accessed
February 3, 2020.
73 Zbikowski, Conrad. “How Houston Reimagined Its Transit Network and Increased Ridership.” Streets.mn, Febru-
ary 25, 2019, https://streets.mn/2019/02/25/how-houston-reimagined-its-transit-network-and-increased
-ridership/. Accessed February 3, 2020.
74 TransitCenter. “The Network Effect: Richmond’s Transit Evolution.” TransitCenter, December 4, 2019, https://
transitcenter.org/the-network-effect-richmonds-transit-evolution/. Accessed February 3, 2020.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
46
Alberta’s Approach
Edmonton, Alberta recently approved its own bus network redesign, which cut 240
routes to just 140. Edmonton’s approach stands out because the transit agency recom-
mends an app-based, on-demand van system to serve low-ridership areas. Edmonton
would not be the first place to deploy on-demand transit; Calgary, Alberta already ser-
vices suburban communities in this manner, and Cochrane, Alberta’s transit system is
composed solely of on-demand vans. The new Edmonton network, though, gives an
example of what the future might hold for cities that use transit to address conges-
tion management: A frequent fixed-route network serving areas where travel demand
is highest and therefore cars are least efficient, supplemented by on-demand service.
Elsewhere, on-demand service may be provided by a public agency, TNCs, or even com-
muters calling their own private cars to pick them up from a transit stop.
COVID-19 Commentary
With the rise of the pandemic, public transit use plummeted. Many systems saw rid-
ership drops of 80 to 90 percent. Urban planners and transportation experts argue
passionately that as economies reopen, shifting public transit riders to cars will have
an incredible impact on congestion and emissions. Nevertheless, pundits, employers,
and some politicians—not to mention car manufacturers!—are telling Americans they
should avoid buses and trains in favor of an automobile. Moreover, some planners
predict citizens will migrate to the suburbs or to towns outside the city in an attempt to
reduce their exposure to COVID-19. On the other hand, due to the coronavirus at least
some people may begin to work from home more frequently, reducing the number
of travelers on the road. The long-term effect of the virus may therefore be to either
exacerbate or alleviate congestion—and this will likely vary across and within cities.
Either way, this will be a new factor policymakers must consider when crafting policy
to address congestion.
AVs may make living in sprawling suburban communities and commuting by car
relatively more appealing, so they might exacerbate a trend to stay out of urban cores
and/or off transit out of fear of catching the disease. Should there be a major shift
towards cars or an increase in urban sprawl (whether influenced by AVs or not), the need
to address congestion and ensure motorists are paying the full and fair price of their
trips will become all the more critical. This requires cities to adopt congestion pricing
for all vehicles, whether owned by companies or citizens. AVs make congestion pricing
technically more feasible, and possibly even politically more palatable. AVs also make
multi-modal travel more appealing, so travelers may feel more comfortable opting for a
bike, a bus, or a walk when they leave the house, secure in the knowledge they can hail
an AV later, should they need it. As discussed above, automated transit makes public
transit a more attractive option, even given COVID-19 transmission concerns.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
47
Lastly, part of the goal of congestion pricing is to push some trips to either before
or after the peak travel periods. If this is successful, a secondary impact may be that
there are fewer people all arriving at work simultaneously, reducing congestion in the
lobbies of major office towers. While this effect may be small, it will help with some of
the policies around physical distancing and elevator use in public spaces.
Getting Started
We believe cities should take policy action on congestion reduction now, before the
introduction of AVs. Why now? Once AVs are traveling city streets, it will be much more
difficult to overcome industry objections to congestion pricing or similar policies.
Indeed, as TNCs become more entrenched
75
it is imperative that cities move to enact
policies soon. The arrival of AVs will only exacerbate the problem if cities haven’t yet
devised incentives for ride-sharing and other strategies for reducing congestion. The
following is a road map for getting started.
Step 1: Profile Existing Traffic Patterns and Congestion
The starting point for addressing congestion is understanding the specifics of your
traffic problem. Answering the following questions provides a starting point for devel-
oping congestion mitigation strategies:
Traffic Management Objectives Strategic Questions
• What outcomes should the program provide? Are the outcomes equitable?
• Is the goal a reduction in travel time or in the number of cars entering the city?
• Do you seek to reduce the number of residents with cars?
• Is the program intended to raise funds for public transit or green transport?
Step 2: Learn from Existing Programs and Establish Objectives
Your congestion profile and insights from existing programs are the inputs needed to
identify your congestion management objectives.
When establishing the objectives, ensure that the outcomes are measurable. You
need to make sure you know if you were successful. While you are acting to address
current congestion issues, think about future objectives in an AV environment.
Step 3: Define Your Guiding Principles
Along with the objectives, establish a set of core principles that will guide the design
of the program.
75 In a 2019 assessment (commissioned jointly by Uber and Lyft) of TNCs’ impact on vehicle miles traveled
(VMT)—a standard metric for overall vehicle activity—the study found that in Suffolk County, which encom-
passes Boston, personal and commercial vehicles accounted for 92 percent of VMT while Lyft and Uber contrib-
uted eight percent.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
48
Sample Traffic and Congestion Management Program Goals
• Provide equal treatment of all travelers.
• Create net public value (benefits greater than costs).
• Encourage transit and active mobility modes.
• Support shared ride options.
• Ensure financial and environmental sustainability.
• Minimize job loss and provide transition assistance to those negatively impacted.
• Test the program for a set period of time.
Step 4: Develop a Congestion Pricing Strategy and Communications Plan
Perhaps more than any other transportation policy, cities need a strong communica-
tions strategy to win over the public as they launch road-tolling measures. In 2019,
the Eno Center for Transportation undertook an overseas study tour to gather lessons
American leaders might use in pursuing congestion pricing. They propose a number of
key insights (available on their website), but the overarching takeaway is that success-
ful cities do not win over citizens by focusing on revenue. Instead, the Eno participants
note policymakers should focus on equity and offering greater access to residents.
New York officials positioned the plan as a way to address the woeful state of the
city’s transit system, while in Los Angeles it is portrayed as a way to tackle stifling
congestion. While these two pitches are more likely to succeed, Chicago’s mayor “sug-
gested congestion pricing as a way to plug a hole in her city’s budget.”
76
The Eno team
argues this narrative will likely not win the day among constituents. (And when Con-
necticut Governor Ned Lamont tried to sell a highway toll program to raise revenue, it
similarly failed.
77
)
Stockholm won over its public by launching congestion pricing
as a temporary initiative. When commuters saw the impact, they became
proponents of a permanent change.
The approach and narrative city officials use to promote congestion pricing will
likely make or break the plan, so there must be a strong process, a policy based on data,
community engagement, and a willingness to employ pilot projects and iterate based
on results. Stockholm, for instance, won over its public by launching the initiative only
temporarily and when commuters saw the impact, they became major proponents of a
76 Puentes, Robert and Brianne Eby. “Takeaways from a European Congestion Pricing Study Tour.” Eno Center for
Transportation, October 16, 2019, https://www.enotrans.org/article/takeaways-from-a-european-congestion
-pricing-study-tour/. Accessed February 3, 2020.
77 Short, Aaron. “Europe’s Congestion Pricing Lessons for America.” Streetsblog NYC, October 31, 2019, https://
nyc.streetsblog.org/2019/10/31/europes-congestion-pricing-lessons-for-america/. Accessed February 3, 2020.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
49
permanent change. Policymakers want to have a strong narrative articulating how the
plan will improve the transportation network for everyone through better travel times
and other benefits.
Step 5: Improve Transit Uptake and Performance
Studies show that the best time to change someone’s travel habits is when they
change jobs. Transit agencies can reach out to major employers, especially those who
are new to the city, to offer some kind of “Welcome Wagon” program—perhaps a free
two-month transit pass for all new employees. Many companies do not want to invest
in expensive parking for their staff, so they might be willing to extend the offer by sup-
plementing the employee’s transit pass in perpetuity.
On the infrastructure front, measures like all-door boarding or installing transit
priority signalization improve reliability and travel times without adding more buses.
Identifying major demand corridors for dedicated transit (like Toronto’s King Street or
Seattle’s Third Avenue) can also dramatically improve performance at no additional cost.
Finally, cities can use revenue raised through congestion pricing to invest in bet-
ter transit. While they can add routes and improve frequency everywhere, there is
a particular benefit to looking at equity concerns, ensuring strong transit service in
lower-income areas.
Step 6: Design a Pilot, Learn and Scale
As policymakers eye AVs playing a larger role in our communities, forward-thinking
leaders should consider a pilot program to learn how congestion can be managed
today and in the AV future. The pilot should enable learning at low risk, and could
be designed based on geography, modes, or time of day. Built with a coalition of the
willing, the design should reflect stakeholder interests. It should also have clear per-
formance measures and a transparent dashboard. Ideally, such a project would be
sufficiently staffed in preparation for bumps along the road.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
50
Quick Notes
The Concept
• Congestion is a problem in many growing U.S. cities, and AVs have the potential to
either alleviate or exacerbate that congestion.
• Vehicle congestion in cities will improve only via effective public policy.
• Policies can be developed for TNCs today and applied to AVs in the future.
• Unlike TNCs, tolling AVs may be less expensive and easier to assess and enforce.
• AVs offer cities more tools for influencing driving behavior and congestion.
The Landscape
• Outside the U.S., congestion pricing has already met with success in cities like Lon-
don and Stockholm.
• Cities have seen significant decreases in car trips to the city center; Stockholm also
reported improved air quality.
• In the U.S., New York City is taking the lead with a congestion pricing program due to
begin in 2021.
• Seattle reduced drive-alone trips from 50 to 30 percent from 2000 to 2016, illustrat-
ing that transit planning can reduce congestion.
Getting Started
1. Profile existing traffic patterns and congestion.
2. Learn from existing programs and establish objectives.
3. Define your guiding principles.
4. Develop a congestion pricing strategy and communications plan.
5. Improve transit uptake and performance.
6. Design a pilot, learn and scale.
Policy Action 4: Establish Data-Sharing Guidelines and Agreements
The Concept
AVs will collect a vast amount of data about all aspects of their operation. They will
be able to collect data on passengers as well as those outside of and around the
vehicle. Inside the vehicle, passenger-facing cameras can watch the rider, potentially
using data about their facial expressions and body language to select and display
advertisements. The vehicle’s outward-facing cameras will record, in real time, what is
happening on the street. At the same time, the AV records data about its own behavior
(speed, braking, proportion of time operating in autonomous versus manual mode,
etc.) as well as external road and environmental conditions such as the weather, traf-
fic congestion, location of potholes and the like.
AV operators are already resisting sharing their data, much as TNCs have. Indeed,
TNCs collect data on their road speed and other operational data that cities could bene-
fit from having, but they are resistant to sharing disaggregate data with municipalities.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
51
Their concerns are tied to issues of confidentiality and competitor access through free-
dom of information requirements.
Least controversial is the use of vehicle operating data. AV operators say that
data belongs to them and they see no reason to share it since it pertains to their own
business operation. Passenger-related data is more controversial. Some argue that
passengers give up their presumption of privacy—and hence data about them and
their travel—when they board an AV, much the same way that the traveling public
has acceded to security cameras in public venues in the name of providing increased
security, but it is not clear if the public will view the cabin of a vehicle in the same way.
AV Data Questions
• Who owns AV operations data?
• How can and should that data be shared?
• Who owns passenger-related data?
• What uses of passenger-related data are acceptable and unacceptable?
• Should AV operators be able to sell passenger data to advertisers?
• Should passengers be able to opt in or out of data-sharing?
• Who owns data about the behavior of the people outside the vehicle?
While these are important issues to grapple with, cities are even more interested
in the data AVs collect about their operating environment. Such data can allow mobil-
ity planners to better understand, in real time, congestion times and patterns. Cities
can use this information to better time lights, understand road usage and inform their
capital investments more effectively. Operators, however, plan to closely guard this
data, fearing that sharing even the most basic details may put them at a competitive
disadvantage. For instance, if there are only two operators in town, when that data
is shared and subject to the Freedom of Information Act, operators are legitimately
concerned that their competitor could calculate their data from the aggregate, thereby
gaining insight into their business. This could lead to significant tension between AV
operators and planners at the municipal level.
The last data tranche—information about the behavior of people outside of the
vehicle—is the most disconcerting from a privacy perspective. For instance, say an
external camera catches someone committing a crime. Most would support using
such data for law enforcement. But what if the camera records people doing their daily
activities and that footage was made available to advertisers? Though cameras are
located and operating atop buildings and in locations throughout cities, most citizens
expect that the footage will only be used for law enforcement. Should AV operators be
able to monetize this data? How else might this information be used?
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
52
Municipalities also have concerns about sharing. Some are concerned that
accepting the data will be costly. There are direct costs for data storage and protection.
There are also potential liabilities and indirect costs of data breaches and responding
to Freedom of Information Act requests. Some cities and privacy advocates also worry
that cities do not have the internal expertise to effectively protect and analyze the
massive amount of data that AVs will generate.
Data-sharing is a thorny issue because it goes to the heart of privacy and secu-
rity issues around personal data. Even the notion of providing non-personal data is
controversial, because operators view it as proprietary, competitive information. We
believe the solution for cities is to partner with AV operators to solve problems. Cit-
ies can pose questions to AV operators and let them use their own analytics team to
provide the answers. If cities want to know the congestion windows on a particular
street or where traffic speeds exceed 25 miles per hour, AV operators can provide
these answers without having to turn over the data they so dearly wish to keep con-
fidential. However, in some cases, it may be necessary to obtain AV operator data to
ensure they are meeting governmental expectations.
The Landscape
As municipalities grapple with mobility data issues, the following topics are
being addressed:
• Guiding principles for data-sharing
• Data handling, sharing and liability
• FOIA and public record requests
• Law enforcement requests
We discuss each of these in brief below.
Guiding Principles
Los Angeles’ data-sharing principles address de-identifying data, limiting the receipt of
raw data, providing data to law enforcement as needed, blocking data transmissions to
third parties that seek to monetize the information, and developing security protocols
to ensure the safeguarding of data received by the City. Seattle’s principles stress con-
stituent privacy and using data to enhance equity and inclusion. Portland’s principles
add the idea of lifecycle stewardship of the data and non-discriminatory use of the data.
Barcelona includes guidance for maintaining the public trust. The idea of data “from
everyone and for everyone” is part of Amsterdam’s data-sharing philosophy.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
53
Sharing of TNC Data
After resisting calls for data-sharing, Uber has begun providing information to munic-
ipalities and the public though its Uber Movement application. Movement provides
data and tools that enable city planners to enhance mobility. The data enable plan-
ners to better understand traffic flows, congestion, and safety risks. The information
is available for more than a dozen cities in North America.
Ford, Uber and Lyft are also sharing data in conjunction with the SharedStreets
initiative. This data, including pick-up and drop-off counts and road speeds, will allow
urban transportation departments to improve mobility.
Sharing of AV Data
Sharing of AV data has been part of the testing protocol in many geographies. In Cal-
ifornia, AV testers are required to report disengagements to the state. This data is
published annually and the results are widely covered in the press. In Boston, where
AV testing has been taking place for several years, testers are required to submit quar-
terly performance updates containing mileage driven, reports of accidents, and les-
sons learned. These are public documents, but no detailed data are shared.
Beyond the statutory reporting, some AV operators have begun sharing this data
with other industry participants and researchers. In August 2019, Waymo became
the highest-profile AV operator to agree to release its data to researchers, following
Aptiva’s lead from a few months earlier.
78
Industry experts claim Waymo and others
are beginning to share data with researchers because solving some of the technical
challenges of AVs is proving more difficult to solve than operators had anticipated; by
sharing data with researchers, they hope the entire industry can benefit.
79
Getting Started
We believe cities should be thinking about how they can use the data that TNCs are
gathering today—and that AVs will gather in the future—to make their cities more
mobile, environmentally friendly, enjoyable, and livable places while protecting their
citizens’ privacy. If cities start now, with TNCs and micro-mobility providers, they’ll
be well-positioned to expand to AVs when they arrive. Those cities embarking on AV
pilots should incorporate a data-sharing protocol into their testing regulations. The
following steps can help municipalities address the data challenge today.
78 Hawkins, Andrew J. “Waymo is making some of its self-driving car data available for free to researchers.” The
Verge, August 21, 2019, https://www.theverge.com/2019/8/21/20822755/waymo-self-driving-car-data-set
-free-research. Accessed September 25, 2019.
79 Stewart, Jack. “Why Waymo is sharing some of its self-driving car data.” Marketplace, August 21, 2019, https://
www.marketplace.org/2019/08/21/waymo-sharing-some-self-driving-car-data/. Accessed September 25, 2019.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
54
Step 1: Determine Data Needs and Wants
The starting point for establishing a data-sharing protocol is identifying the informa-
tion needs (real requirements) and wants (“nice to have” information) of the munic-
ipality. What questions can be effectively answered through mobility/AV data? The
queries should span from long-term infrastructure planning to short-term pothole
repair. The data can also substantially inform congestion management programs. The
questions list should be developed by engaging every city department, not just the
transportation department. Departments dealing with elder affairs, health, education
and revenue all have an interest in learning from mobility data. The list of questions
should also reflect the needs of regional and/or state planners. The city should also
specify the process for asking questions and receiving answers, including the time-
frame and level of detail to be provided.
Step 2: Draft Data-Sharing Guidelines
Beyond leveraging data to enhance the delivery of public services, the municipality
should also define the rules of the game for data-sharing with the city and with other
organizations. Based on domestic and international experience, the following are
foundational principles for the guidelines:
80
Data-Sharing Guidelines: Key Principles
• Transparency: Clarify and publicize the kind of data being collected and for what
purposes. Aim for data minimization rather than maximization.
• Responsible Stewardship: Act in the best interest of citizens. Safeguard confidential
information. Make data serve people, not vice versa.
• Inclusion: Use data to promote equity, target economic and social disparities, and
encourage affordability.
• Public Value: Use insights gleaned from data to drive inclusive economic growth,
improve security, and address climate change-related challenges. Data must bring
utility to the city’s residents.
• Portability: Easily move non-confidential data. Prevent information from becoming
“siloed” and “sequestered.”
The resulting guidelines should be documented and accessible to mobility pro-
viders and the public. The guidelines should also have built-in flexibility so that revi-
sions can be incorporated as both the city and the public’s thinking evolve.
Step 3: Begin Implementing with Existing Mobility Providers
The development of a data-sharing program can begin immediately. While AVs will
gather extensive amounts of information, current and near-term mobility providers
80 Blomfield, Tim, Brian Feldman, Monica Pellerano, and Philip Essienyi. “AV Data Policy and Governance.” 2019,
prepared for the City of Boston by Harvard Kennedy School students.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
55
can also be sources of data for decision-making. TNCs and taxis are existing sources
that can be tapped for mobility insights. While incumbents might balk at providing
information since it was not a term of their operating agreement, it is important to ask.
If it is not possible now, it can be at the time of an agreement renewal. Also, it is likely
that the TNCs will be some of the first AV operators in the urban core. The AV operating
agreement might be tied to sharing all of the operator’s data, not just the AV portion.
Micro-mobility providers present a low risk test bed for the design and operation
of a data-sharing protocol. As bikeshare and electric scooter companies seek local
operating agreements, cities have an opportunity to pilot data-sharing agreements.
Step 4: Evaluate the Value of Information Sharing and Refine Guidelines
Data-sharing arrangements should be evaluated periodically to confirm desired
outcomes are being achieved. Are the city’s questions being answered? Are there
new queries? Are the data captured about users and the general public being pro-
tected as designed? Are there changes needed to meet an important public purpose?
The answers to these questions will inform the process of refining and reissuing
data-sharing guidelines.
Quick Notes
The Concept
• AVs collect a vast amount of data about all aspects of their operation.
• AV operators are already resisting sharing their data, much as TNCs have done.
• The open question is: who owns this data? How can and should it be used?
• Issues of privacy, security and competitiveness may position cities against AV operators.
• Cities should establish a data-sharing protocol for current and future mobility options.
• Partnering with mobility providers to answer key planning questions is an effective
alternative to cities obtaining and analyzing the data themselves.
The Landscape
• Cities around the globe have established data-sharing principles that stress privacy,
equitable use and transparency.
• TNCs have begun sharing data that sets a precedent for how information flows can
be established.
• AV operators have begun sharing their own mapping data with others in the indus-
try and researchers with an eye toward collectively solving technical AV operating
challenges.
Getting Started
1. Determine data needs and wants.
2. Draft data-sharing guidelines.
3. Begin implementing with existing mobility providers.
4. Evaluate the value of information sharing and refine guidelines.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
56
Policy Action 5: Reposition Revenues
The Concept
Municipal and state governments rely on revenue from private vehicles, largely derived
through their state’s excise tax on gasoline. As of July 2020, such taxes ranged from
$0.15 per gallon in Alaska to $0.61 in California. Some states levied additional state
sales or local taxes. States rely on this revenue to fund road maintenance, transporta-
tion infrastructure, and transit subsidies. Additional, significant sources of municipal
revenue include parking fees, parking tickets, and moving violation fines. For a city the
size of Boston, these can amount to as much as $88 million annually—the equivalent
of the entire cost of snow removal for a year. The magnitude of the potential revenue
loss could be more than $100 million in other major cities.
These revenues have been declining nationwide. Increasing auto fuel efficiency
is reducing fuel consumption and therefore gas tax revenues are declining. In some
cities, parking revenues are declining as municipalities disincentivize people from
driving into town by reducing or eliminating parking spots.
The market entry of AVs will speed up the revenue decline. Autonomous vehicles
will commit far fewer, if any, moving violations, drying up this revenue stream. If the
shared ride AV model takes hold—or if car commuters send their vehicles home during
the workday—demand for parking will also decline, taking with it parking meter fees,
municipal lot parking fees, and fines for meter violations.
Because we believe AV fleets will be comprised of electric vehicles, gas tax rev-
enues will also decline, providing an additional catalyst for cities to take action to
replace these revenues.
We see two major areas where cities can look to make up for lost funds: transpor-
tation charges and land use/property taxes
Transportation Charges
While eroding traditional revenue streams in transportation (through gas taxes, park-
ing fees, and tickets), AVs also provide a new opportunity for cities to view the trans-
portation network more like a utility, with travelers paying for what they use. In order
to operate, the expectation is that AVs will keep a record of where they travel, mean-
ing cities can use that data to charge cars for the road space they take up. When in
motion, this is not dissimilar to the congestion pricing outlined in Policy Action 3, but
it goes further by charging everyone for the miles they have traveled, not just in con-
gested downtown cores. This pricing can be first applied to TNC operators (and many
states and cities already charge a fee to operate), but in the future even private drivers
should pay a per-mile fee for the road space they use.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
57
AVs provide a new opportunity for cities to view their transportation networks
more like a utility, with travelers paying for what they use.
When a vehicle is stopped, they can “rent” space along the curb to drop off or
pick up passengers. Cities are already familiar with the concept of charging for curb
space for parked vehicles, but this is typically limited to longer stays (at least 15 min-
utes or more) and is generally confined to very popular locations. In the future, cities
can use GPS data to charge for all passenger (and freight) loading and unloading,
regardless of the duration the vehicle is stopped or the area of the city it is in. Prices
can vary throughout the day depending on demand, but as with per-mile travel pricing,
the goal would be to fairly charge users for the space they take up.
Land Use and Property Taxes
Cars today require a huge amount of space in cities. Researchers estimate there are
anywhere from three to eight parking spaces per car in America,
81
which means all
that land cannot be used for more productive—and more tax revenue-generating—
activities. If AVs reduce parking demand, local governments can use policy levers
to encourage residential and office development on this land, with more property
tax revenues helping to offset lost revenue. Moreover, if this development leads to
higher overall city density, it could reduce the per-capita amount the government
must spend to provide the same level of service, easing the strain on local revenues.
A study conducted in Calgary, Alberta, for example, showed that if the city could
consume a quarter less land through a denser growth pattern, it could save $11 bil-
lion in capital costs alone.
82
Governments can begin to eliminate parking minimums in their jurisdictions,
leaving it up to developers (through community consultation) to determine how much
space they should allocate to parking. They can also institute parking maximums, and
encourage (or even require) parking garage owners to design any new garages such
that they can be converted to residential or office use in the future, as some develop-
ers are already doing. More boldly, cities could begin to tax existing parking lots at a
higher rate, to spur non-parking development.
81 Goldstein, David. B. “Does Every Car Need 8 Parking Spaces? Ride-Sharing Can Save Emissions by Reducing
Parking, Too.” National Resources Defense Council, March 9, 2015, https://www.nrdc.org/experts/david-b
-goldstein/does-every-car-need-8-parking-spaces-ride-sharing-can-save-emissions. Accessed February 3, 2020.
82 Thompson, David. Suburban Sprawl: Exposing Hidden Costs, Identifying Innovations. Sustainable Prosperity,
October 2013, p. 6, https://institute.smartprosperity.ca/sites/default/files/sp_suburbansprawl_oct2013_opt
.pdf. Accessed February 3, 2020.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
58
Most ambitiously of all, officials can consider moving towards split-rate prop-
erty taxation, acknowledging the role their investments play in increasing the value of
land. In this scenario, cities tax land at a much higher rate than they tax the building,
to recognize the fact that land-value increases are oftentimes the result of new invest-
ments made by the government and surrounding landowners, rather than the direct
owner. In Pittsburgh, for instance, the City taxes land at a rate five times that of the
tax on buildings.
83
This encourages owners to develop their land for more productive
uses, and also shares the wealth with government as land values rise, with the taxes
raised being reinvested back into city services and infrastructure—that oftentimes
leads to higher land value still.
Achieving revenue neutrality will likely come from a combination of AV pick-up
and drop-off (PUDO) charges, vehicle miles traveled (VMT) fees, and increased prop-
erty taxes coming from repurposed parking lots and spaces. In Toronto, for example,
city officials, seeking a model to achieve revenue neutrality after the introduction
of AVs, found that a combination of PUDO fees and repurposing parking garages
into living spaces was sufficient to reach that neutrality. On top of these initiatives,
split-rate property taxation attempts to encourage building upgrades and densifi-
cation while more fairly distributing the financial benefits of new surrounding infra-
structure—such as transit stops, parks, bike lanes, and other amenities—that make
a community more desirable.
While city officials may feel trepidation at the prospect of lost transportation
revenue in the years to come, the potential benefit of AVs is that local governments
can become less dependent on revenue from detrimental activities (such as speed-
ing or improper parking) and instead find new funding from more vibrant, active
residential and office uses, and by charging people for the road space and services
they actually use.
The Landscape
In many cities, passengers are already familiar with VMT charges thanks to taxi pric-
ing. Taxis often impose an initial fee, then add an incremental charge based on dis-
tance and waiting time. It is therefore not a big leap to adopt a similar approach when
determining a VMT fee that a TNC would need to pay to the government. Currently,
cities that do charge TNCs—including Chicago, New York City, and Boston—require a
flat per-trip fee (for example, $2.75 for a trip in Manhattan), but in states including
Wyoming, Nevada, and Rhode Island, TNCs pay a percentage fee per ride. While this is
83 Ibid, p. 15.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
59
not completely synonymous with a per-mile fee, it does vary the amount of revenue
the government receives based on distance and, because TNCs employ “surge” pric-
ing when demand is high, it also acts as a form of a congestion fee. It could therefore
form the initial foundation of a more advanced VMT pricing plan.
While a very small number of cities
have a form of congestion pricing (as
noted in Policy Action 3), in 2015 Ore-
gon launched a voluntary project that
charges participating drivers 1.5 cents
per mile. Participants attach a device
to their vehicle and are refunded the
state gas taxes they pay at the pump.
84
Utah offers low-emissions vehicle owners the opportunity to pay a similar charge, and
Washington state is also considering a per-mile program.
85
Though these programs
are voluntary, they again show the potential of a VMT charge to offset lost revenues
at the pump.
Most cities are already familiar with charging a fee for curb access—but as park-
ing, rather than for a drop-off. Charging for parking has traditionally been a very static
affair, but a few cities are becoming more advanced in their approach. Los Angeles
offers varying rates throughout the day in Hollywood, reflecting temporal demand.
San Francisco’s SFpark program similarly varies pricing charges by time of day and day
of the week, with pricing updated once a month to moderate demand.
Thus far, no cities charge for curbside drop-offs or pick-ups lasting just a few
seconds, but some have begun to rethink how they manage their curb space. Seat-
tle’s latest comprehensive plan introduces the concept of “flex zones” in the parking
lane, with uses changing throughout the day as demand shifts.
86
During rush hour, the
lane can be designated a commuter pick-up/drop-off zone, then used for deliveries
in the morning, and parking during the workday, for example. The government has
prioritized activities based on the needs of the street and the activity it serves. Paris,
84 Miller, Stephen. “Oregon’s Pay-Per-Mile Driving Fees: Ready for Prime Time, But Waiting for Approval.” Streets-
blog USA, June 26, 2017, https://usa.streetsblog.org/2017/06/26/oregons-pay-per-mile-driving-fees-ready
-for-prime-time-but-waiting-for-approval/. Accessed February 3, 2020.
85 Groover, Heidi. “Long road ahead as Washington transportation officials recommend per-mile charge to replace
the state’s gas tax.” The Seattle Times, December 17, 2019, https://www.seattletimes.com/seattle-news
/transportation/long-road-ahead-as-washington-transportation-officials-recommend-per-mile-charge-to
-replace-the-states-gas-tax/. Accessed February 3, 2020.
86 City of Seattle. Seattle 2035 Comprehensive Plan. December 2018 edition, p. 76, https://www.seattle.gov
/Documents/Departments/OPCD/OngoingInitiatives/SeattlesComprehensivePlan/CouncilAdopted2019.pdf.
Accessed February 3, 2020.
AVS AND MUNICIPAL REVENUE STREAMS
• New AV pick-up and drop-off (PUDO)
fees
• New vehicle miles traveled (VMT) and/
or access fees
• Increased property taxes from repur-
posed spaces
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
60
meanwhile, has removed 43 percent of its personal parking spaces from the street
since 2001, while reserving some slots for deliveries on major streets.
87
These and other initiatives are paving the way for further measures down the
road. They provide cities with better data as to demand in different areas, and help
drivers and passengers become more comfortable with using online platforms as part
of their pick-up/drop-off activities. Moreover, these programs are making city streets
more efficient today; data show Washington’s three-month pilot, for example, cut
double-parking by 64 percent in the study area.
88
The next step would be to develop
and implement a pricing scheme for curbside activity, and while no city has formally
established one yet, a major International Transport Forum study determined that
were Lisbon to bring in such a measure, it could completely replace lost parking reve-
nue by charging just 0.10–0.28 Euro ($0.11–$0.31) per pick-up or drop-off.
89
As local governments seek to revitalize their downtowns, reduce congestion, and
promote density, across the country they are lifting minimum parking requirements
that have forced developers to put in a set number of parking spaces for different
commercial and residential activities, whether there was a demand for those spots or
not. Buffalo, New York was the first city in the country to completely remove minimum
parking, but hundreds of local governments have eliminated these requirements in
certain geographies.
90
This makes it less expensive to build new development, and
also helps promote density and alternative transportation options.
Other cities, such as Montreal, Quebec, have adopted an even more aggressive
approach. The Canadian city currently taxes parking lots at a rate of $29.70 per square
meter. The government has also reformed its zoning to permit taller buildings. The effect
of these two measures together has been to prompt many commercial parking owners to
sell their lots to developers.
91
These taxes can provide much-needed revenue; Montreal
brings in about $23 million annually, while a KPMG study found that if Toronto, Ontario
were to implement a similar measure, it would net between $171 and $535 million per
87 Grabar, Henry. “Give the Curb Your Enthusiasm.” Slate, July 19, 2018, https://slate.com/business/2018/07
/curb-space-is-way-too-valuable-for-cities-to-waste-on-parked-cars.html. Accessed February 3, 2020.
88 Pyzyk, Katie. “CurbFlow pilot reduced double parking in DC by 64%.” Smart Cities Dive, November 14, 2019,
https://www.smartcitiesdive.com/news/curbflow-pilot-reduced-double-parking-in-dc-by-64/567268/.
Accessed February 3, 2020.
89 International Transport Forum, The Shared-Use City: Managing the Curb, p. 80.
90 Poon, Linda. “Buffalo Becomes First City to Bid Minimum Parking Goodbye.” CityLab, January 9, 2017, https://
www.citylab.com/equity/2017/01/buffalo-is-first-to-remove-minimum-parking-requirements-citywide/512177/.
Accessed February 3, 2020.
91 Lampert, Allison. “Taxes on Montreal parking lots have doubled.” The Gazette/Global News, January 29, 2013,
https://globalnews.ca/news/385451/taxes-on-montreal-parking-lots-have-doubled/. Accessed February 3,
2020.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
61
year. The incentive to sell parking lots leads to new development down the road, with
the commensurate increase in tax revenues from higher-use projects.
92
The Calgary Parking Authority is currently constructing a 250,000-square foot,
510-car parking garage that will eventually be converted into a residential develop-
ment. Acknowledging the likely trend of lower parking demand in years to come, the
Authority intends to convert their garage into multi-family housing once it has outlived
its usefulness as a garage, with only minimal structural alteration required.
93
As cities encourage development, it will be important to ensure a significant
amount of the new construction is of affordable—often rental—housing. Permitting
only luxury condos or townhouses will raise revenue, but will be less helpful to many
members of the community who cannot afford such expensive homes. Cities can also
encourage more mixed-use development and gentle density (that we define as build-
ings between three and six storeys), which will help reduce travel demand and sprawl.
Reducing travel demand and sprawl help with the wear and tear on infrastructure and
cut the need to invest in more suburban infrastructure. Moreover, density makes public
transit more cost effective. When managed properly, these measures will make housing
and transportation more affordable for all residents and cut pressure on city finances.
COVID-19 Commentary
When the pandemic first hit, some critics were quick to assume dense urban areas
exacerbated the spread of the coronavirus. Researchers are still in the data-gathering
phase, but initial evidence shows it was not density that made the crisis worse, but
crowding. There is an important difference between the two concepts—density is the
number of people who live or work in a particular place, while crowding is related to
direct contact with other humans. One measure to reduce crowding is increasing the
supply of affordable housing. As AVs enable repurposing parking facilities into hous-
ing and reduced parking requirements for new developments lower housing costs
de-crowding can take place.
Another benefit of repurposing parking into housing, office and retail space is an
increase in tax revenues from these higher value land uses. This is especially attrac-
tive, as state and local government have seen their revenues plummet with COVID-19.
92 Forman, Gideon. “Why Toronto Should Issue a Commercial Parking Levy.” Torontoist, September 16, 2016,
https://torontoist.com/2016/09/commercial-parking-levy/. Accessed February 3, 2020.
93 Israelson, David. “Future-proofing parking lots.” The Globe and Mail, July 30, 2019 (updated August 26, 2019),
https://www.theglobeandmail.com/business/industry-news/property-report/article-future-proofing-parking
-lots/. Accessed February 3, 2020.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
62
Getting Started
Already, conventional TNCs are reducing parking demand in some communities, and city
governments are looking to encourage greater densification and urban development for
social, fiscal, and environmental reasons. We therefore believe it makes sense to move
towards new revenue models today, rather than waiting for the arrival of AVs—as by
then, cities will already be feeling the financial pinch of this new technology. While no
single silver bullet will mitigate the decline by itself, using a range of tools will likely be
more politically feasible and will lead to changes in urban form over time.
Step 1: Understand Your Revenues and the Policy Levers You Can Pull
In order to ensure that your city government will recover revenues lost during the tran-
sition to AVs, it is important to take stock of the revenue that may dwindle with their
arrival. Different cities have the existing ability to levy different fees (due to state-given
authority), so coming up with a comprehensive total will provide a hard number that
officials can use when determining future fee rates—to cover funding losses without
being excessive.
Simultaneously, officials should determine what authority they have to imple-
ment new rates on companies, passengers, and developers. While most cities across
the country share a number of powers—such as the ability to charge for parking—
some cities have been granted additional abilities through state legislation. Knowing
whether your state has expressly permitted or prohibited certain activities, like the
right to institute a charge on TNCs or to place a tax on parking lot space, will show
what tools you have at your disposal as you look to recuperate lost revenue.
Where you do have authority, consider the very clear objective of imposing a new
fee. A per-trip fee was an appropriate first response to this new technology on city
streets, for instance, but would likely be inadequate for compensating a city for lost
parking and ticket revenue, let alone trying to mitigate congestion. More advanced
mechanisms can accomplish governmental goals more effectively, so understanding
the tools at your disposal, and the resources necessary to enforce each one, is an
important first step to ultimately solving urban challenges.
Citizens are generally not favorable to parking fees, and TNCs may try to foster
opposition to future curb access fees as well. As such, officials will want to have a
communications strategy in place to win support for these measures. Parking guru
Donald Shoup recommends that parking charges should be set to ensure there are
always one or two spaces available and that revenues raised should be invested back
into the community; similarly, cities will want to make clear that their goal is to keep
one or two spaces available at all times for drop-offs, and consider committing to
investing at least a portion of the funds raised back into the surrounding community.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
63
Step 2: Identify Existing Curb Demand and Areas for Flexible Curb Space
This step aligns closely with Policy Action 2, with officials better understanding the dif-
ferent demands of their curb space. Where possible, local government can require TNCs
to provide data on pick-up and drop-off locations (batched into streets and times of
day to ensure customer anonymity). Alternately, cities can use historical parking data
to track downward or upward trends in parking demand, or more traditional measures
such as surveys to monitor where TNC demand is greatest. Like Washington, they will
then be able to determine advantageous sites for flexible curb usage pilot projects, with
an eye toward eventually implementing a pilot project in the busiest areas.
If resources allow, cities can also move towards a dynamic paid parking plan,
like SFpark, to help citizens become more familiar with flexible pricing that helps shift
demand.
Step 3: Ensure All Loading Zone Signage Is Specific, Visible and Managed
To help citizens become more comfortable with flexible curb use, signage should be
very clear in any areas of change. Some cities offer videos explaining changes, and
as much as possible, cities should launch information campaigns to inform citizens
about new expectations—especially during any pilot projects. Some cities use fun
informational campaigns to engender support for paid parking (for example, a parking
meter posting notices such as, “Your parking fee today paid to mow 130,000 blades
of grass at Huntington Park” nearby). A similar strategy can be used when informing
citizens about curbside charges.
Step 4: Confirm Current Tax Code Doesn’t Incentivize Parking Lots
Some municipalities vary tax rates across property classes and across neighborhoods.
The tax rate on higher-density apartment buildings is greater than on single-family
dwellings, for example. It is common for cities to charge a lower tax rate on parking
lots than built structures, and some cities do not include parking garages when calcu-
lating the floor-area ratio of a building. These measures have made it relatively more
attractive to build parking lots or garages than more productive developments.
It is important to understand the existing tax code, and then propose areas for
reform. Oftentimes tax structures run counter to the stated objectives in a city’s mas-
ter plan, such as densification, the promotion of active public transportation, and fos-
tering more complete communities. As much as possible, cities should modify the tax
structure such that it acts as both carrot and stick to promote densification, especially
as parking demand dwindles.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
64
Step 5: Determine Appropriate Areas for Eliminating Parking Minimums
While reforming the zoning code to eliminate parking minimums often lies with city
council, officials can prepare a recommendation as to where removal is most appro-
priate. Most cities have begun in the downtown area, but other areas might be equally
suitable. Likely candidates are dense places with good access to transit that have a
strong mix of uses. San Diego, for instance, eliminated requirements for parking on
sites within a half-mile of a transit stop.
94
Quick Notes
The Concept
• Municipal and state governments rely on revenue from private vehicles through gas
tax, parking fees and driving or parking violations.
• These revenues have been declining; market entry of AVs will speed up the decline.
• Transportation charges and land use/property taxes can make up for the shortfall.
The Landscape
• Washington and Oregon are experimenting with vehicle miles traveled (VMT) fees.
• San Francisco and Los Angeles adjust parking fees based on demand.
• Washington, D.C.’s pick-up/drop-off zone pilot cut double parking by 64 percent in
the study area.
• Buffalo has eliminated parking space requirements on new developments.
• Montreal charges a very high property tax on parking facilities to stimulate
higher-value land use.
Getting Started
1. Understand your revenues and the policy levers you can pull.
2. Identify existing curb demand and areas for flexible curb space.
3. Ensure all loading zone signage is specific, visible and managed.
4. Confirm current tax code does not incentivize parking lots.
5. Determine appropriate areas for eliminating parking minimums.
94 Pacheco, Antonio. “San Diego eliminates parking requirements for transit-adjacent projects.” The Architect’s
Newspaper, March 6, 2019, https://archpaper.com/2019/03/san-diego-eliminates-parking-requirements
-for-transit-adjacent-projects/. Accessed February 3, 2020.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
65
vi. pilot project guidelines
Municipalities around the globe are running AV experiments that explore how the
policy issues described above can be addressed. Each pilot reflects the unique con-
text and objectives of that community. Based on what we have observed, we present
below the key guidelines for municipalities to consider in order to undertake a suc-
cessful pilot project.
1. Determine jurisdiction. As creatures of their respective states, not all cities have
the same authority when it comes to different AV interventions. Louisville, for
example, controls its own transit network and can therefore unilaterally make
choices about transit deployment, but in Boston, public transit is controlled by a
state-level entity. Once you have determined an area for intervention, you must
consider whether you can act alone or build a coalition of willing partners. In the
latter case, you must think about how to propose the initiative to other govern-
ment agencies in a way that resonates with them and craft a funding model that
can satisfy all parties. If this is not possible, you might need to look at a pilot
project in which you have full control.
2. Determine what “success” looks like. Mobility expert David Zipper has written a
useful policy brief on pilot projects. His main takeaway is that a successful pilot
project is not one where the specific service works, but one that “produces data
that informs a public decision to either accept or reject its hypotheses about the
new mobility technology.” It is important to understand, specifically, what you
want to know, what metrics you will use to assess that knowledge, and whether
you have the capacity to gather and evaluate the metrics. Be sure to speak with
officials outside of your transportation department, as they might have insights
or ideas that are important but beyond the lens of transportation planning alone.
As part of this process, also do not be afraid to reach out to cities and states that
have pilots underway. How did they determine what success looked like and what
to measure? What were they able to measure, and where did struggles arise?
Finding out this information allows you to build on what has already been done,
and if you offer to share data with your contacts, it can help develop a reciprocal
network of data-gathering local governments. Finally, you must do thorough com-
munity consultation, in order to understand the needs of your stakeholders and
what concerns and opportunities they foresee.
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
66
For policymakers, act now and proactively plan to make your community’s AV future
happen, rather than passively waiting for that future to “happen to you.”
3. Reach out to appropriate partners. Once you have determined your objectives,
you will need to issue a Request for Proposals. We encourage you to go beyond
merely putting out a call and waiting for responses. Instead, determine which
companies might have a product that can most appropriately meet your needs.
Similar to firms using a headhunter to find the best candidates, cities need not
be afraid to alert appropriate companies to an opportunity. If you have previously
determined the specific criteria for and measurements of success, they will know
what to expect from the relationship (around data-sharing, level of service, and
other criteria), which should help encourage only serious candidates to apply.
4. Craft a robust communications strategy. We believe the potential benefits of AVs
merit a controlled pilot. It is important to communicate the goals and expected
advantages of the initiative (for instance, less congestion, more equitable access
and environmental gains) to the public while also addressing any concerns (for
instance, fears about AV safety or employment impact). Research at MIT has
found many politicians view AVs as more of an economic development tool than
a solution to transportation problems. If officials support the project under this
pretense, they may be disappointed with future results. It is best to clarify your
project’s limited objectives early, show the direct benefit to citizens within the
target area, and then use effective communication tools (including social media,
traditional media and several others).
Above all, we encourage you to recognize that any pilot project will be about more
than just moving people from place to place. It will be about introducing your citizens
to automated technology, so you should think carefully about how you would like to
do that: will it be in a car, or will you inject some excitement into your transit system by
having people first experience automation on a bus? Whatever you ultimately decide,
once AV companies put their vehicles on your roads and begin to build a coalition of
supportive customers, it will be more difficult for you to modify deployment if it is not
meeting local objectives around sustainability, mode shift, or equity. Therefore, we
encourage you to act now and proactively plan to make your community’s AV future
happen, rather than passively waiting for that future to “happen to you.”
autonomous vehicles are coming Five Policy Actions Cities Can Take Now to Be Ready
67
vii. acknowledgements and contact information
We are grateful to AARP for their collaboration and support of our AV policy scrums.
And thank you to the public servants from the states, cities, and organizations, as well
as the Harvard Kennedy School students, who participated in these important and
innovative policy engagements for the public good.
For more information about AV policy or hosting a scrum, visit the AVPI website.
We welcome your questions or comments. Mark Fagan, Lecturer in Public Policy and
Director of the AVPI, can be reached at mark_fag[email protected]vard.edu.
A publication of the
Taubman Center for State and Local Government
Harvard Kennedy School
79 John F. Kennedy Street
Cambridge, MA 02138
