Autonomous mobility – How demand and supply are moving closer to equilibrium
Autonomous vehicles: they are becoming real. We expect the adoption of autonomous vehicles (AVs) to gain serious momentum over the current decade as supply and demand move closer together.
Article information and share options
Autonomous vehicles (AVs) will have a transformational impact on the insurance industry. Insurers, in order to develop the best products, need to truly understand the complexity that characterizes autonomous vehicles and the new business models that may emerge.
This is the first of a series of articles on AVs. A second article will explore supply side AV production developments and their potential impact on the insurance industry. In the third article, we will look at the demand side and the different business models that may emerge in the market of AVs.
Shifting supply side factors include:
• Rapid technological advancements
• Greater infrastructure investment
• Increased collaboration between vehicle manufacturers and tech firms
• Increasingly supportive regulatory developments.
From the demand side – transportation of passengers and goods, and automation of specialized tasks – the benefits of AVs are becoming more visible, particularly as perceived safety improves. The emergence of new business models and ecosystems will further encourage the confluence of supply and demand, making the AV market increasingly attractive to investors, producers and consumers.
Levels of autonomy of AVs start at 0 (no automation) to 5 (fully autonomous in any conditions). The transition from levels 2 to 3 – from Advanced Driver-Assistance Systems (ADAS) to Automated Driving Systems1 – is used here to define AVs. In this article, when we refer to AVs, we mean vehicles with level 3 autonomy or above.
The supply side
1. Evolving technological developments
Two key technical elements are driving supply side development by various players2.
- a. Hardware breakthroughs: Most notably in sensors such as ultrasonic, cameras, LiDAR and RADAR3 equipment.
- b. Software and data innovations: These include: AV operating systems; faster data transmission and data processing solutions; and the use of artificial intelligence (AI) and machine learning (ML). These are key components in the evolution AV technology. The self-learning element of ML enables AVs to become safer and more robust over time.
A wide range of sensors are used in AVs; and sensor technology is rapidly improving. LiDAR sensors, for example, are now capable of measuring distances at rates greater than 150 kilohertz (150,000 pulses per second). Radar sensors embedded in tiny chips can now be easily used in smaller vehicles. Argo AI, a self-driving technology company backed by Ford and Volkswagen, recently announced that their LiDAR can accurately identify objects more than 400 metres away, almost 100 metres more than current sensors.4 AVs increasingly combine multiple sensors to achieve optimal outcomes across a wide range of conditions.
Level 1 to Level 2 autonomous features are becoming more common. These include: automatic braking; lane and cruise control; advanced driver assist (ADAS); some autopilot features; and traffic jam chauffeurs,5 6 Daimler’s Mercedes-Benz, for example, recently applied for global regulatory approval of its Level 3 conditional automated driving system, a further sign that this technology is maturing fast.7 Several Level 4-5 pilots are underway. These may be some years from commercial use, notably against key criteria such as the area of operations, and whether a safety driver or remote operator is needed.8 Some countries – including the US, China, Singapore, Sweden – are testing the commercial use of AVs for public transportation; despite some setbacks, progress has been considerable with large volumes of data being collected.9
Companies from a number of countries are leading the drive towards better technology for AVs. According to the KPMG Autonomous Vehicle Readiness Index, Israel leads in AV tech and innovation, hosting many AV company headquarters and large investments in AV-related firms (adjusted for population). Close behind is the US, with deep industry partnerships and enabling infrastructure such as cloud computing, AI and Internet of Things (IoT).10 However, the AV market with most potential is probably China, as a result of strong domestic demand, favourable regulations and heavy private sector investments.11
2. Intensified collaboration amongst players
AV development began in isolated pockets, with original equipment manufacturers (OEMs) or technology companies running independent R&D operations. This siloed approach is changing for the better. Tighter integration and growing collaboration have allowed multiple pilots to run in parallel. Daimler, for example, has partnered with Bosch; and Hyundai with Cisco. Further, BMW and Daimler have partnered (EUR 1 billion) to create a number of AV related services. Similarly, Ford and Volkswagen (VW) have partnered, with VW investing USD 2.6 billion in Ford’s Argo AI initiative.12 Partnerships are opening in areas beyond passenger vehicles. For example, Waymo and Daimler Trucks recently partnered to create an autonomous version of the Freightliner Cascadia truck. This is Waymo's first foray into the freight industry.13 Auto parts suppliers are joining the fray. Denso (auto part supplier) is working with Aurora (AV start-up) and Toyota on AV technology.
In addition to partnerships, OEMs are acquiring or investing in AV companies/start-ups or investments. General Motors acquired Cruise (an AV start-up) in 2016, which now forms a key part of GM's AV strategy. In 2019, Daimler Trucks acquired Torc Robotics to develop AV technology for trucks. Ford and Volkswagen invested USD 1 billion each in Agro AI, in 2017 and 2019 respectively.14 Big Tech is also active in the field. Apple acquired Drive.ai in 2019 and more recently Amazon acquired the AV start-up Zoox in 2020. All these point to an accelerated collaboration amongst players and a stronger likelihood of inter-operable standards.
3. Increasing investment in infrastructure
Public infrastructure is critical for a successful AV roll-out; but it has received relatively low attention with investors, inventors and policy makers. They have instead focused on technology and safety related to the AVs themselves.15 Key infrastructures that affect the adoption of AVs include:
- a. Physical infrastructure: Quality of roads, lane markings, machine readable sign boards, pavements, curbs, designated parking space etc.
- b. Communication networks: 4G/5G coverage
- c. Digital infrastructure: Sensors, transmitters, intelligent signals, high definition maps etc.
- d. Electric vehicle (EV) charging stations (assuming majority of AVs will be EVs).
High quality roads are a pre-requisite for AVs. Advanced technology can only go so far in compensating for bad roads. Only 41% of US roads are considered worthy of a “good ride,” by the International Roughness Index.16 Potholes, bad surfaces and worn-out markings create safety problems for AVs. By point of reference, Singapore has the best quality roads in the world, followed by the Netherlands and Japan; while Brazil, Russia and the Czech Republic are among those lagging.17
Connectivity and high-speed networks are critical for AV success. We are approaching critical mass, with the industry representative body Groupe Speciale Mobile Association (GSMA) stating that 5G was commercially available from 135 operators in 52 markets worldwide in 2021. It is expected that 5G will reach 1 billion users much sooner than the time taken by 4G and 3G networks to gain similar traction.
Assuming that the majority of AVs will be electric vehicles (EVs), adequate charging stations will be needed. Public stations manufactured by different suppliers currently lack standardized charging and payment systems.18 Close to 90% of charging units as of 2019 were private slow chargers for light vehicles in homes, multi-dwelling buildings and workplaces.19
Countries have different levels of infrastructure provision. Northern European countries have more EV charging stations per capita than Southern European counterparts. Alongside the best roads, Singapore has the best broadband provision, an accolade it shares with Australia.
4. Flexible policy and regulatory attitudes
Regulation is consistently seen as one of the most important issues for future AV deployment. Regulators and policy makers have to be convinced that AVs are safe and fit to be driven on public roads.
Motor vehicles are governed by strong regulations (vehicle registration, driver's license, laws around liability and insurance, emission norms, safety standards etc.). These regulations need modification to enable mass adoption of AVs. Key to such modification is the Vienna Convention on Road Traffic, ratified by more than 70 countries, which states “a human driver must always remain fully in control of, and responsible for the behaviour of their vehicle in traffic.” The key innovation of AVs is precisely to allow control to be reduced or taken away from the driver.
Europe: many markets proactively enacting pro-AV legislation
Many authorities are already working in a pro-AV direction. A range of countries have introduced regulations for AVs on public roads. The European Commission revised its General Safety Regulation (GSR) in September 2019, allowing some AV features - including driver drowsiness and attention warning systems - by May 2022 for new vehicle approvals and May 2024 for all new vehicle registrations.21 Germany formulated a "Strategy for Automated and Connected Driving" in 2015, with the aim of being a leading global supplier of AVs.22 The UK parliament approved the Automated and Electric Vehicles Act 2018, providing the regulatory framework allowing cars to use 'self-drive' automated lane keeping systems (ALKS) in slow-moving traffic at a maximum speed of 60 kmph.23
Asia-Pacific: different stages of AV legislation development
Singapore introduced AV rules in 2017 and was lauded for its relatively clear and comprehensive policies governing AV deployment. Singapore is currently consulting with industries and universities to develop a land transport vision and roadmap for AV development.24 China issued guidelines for road-testing of AVs in 2018;25 and 11 national ministries recently jointly promulgated the "Strategy on Developing Smart Vehicles".26 In Australia the National Transport commission has launched the Automated Vehicle Program to look into end-to-end regulation around AVs.27 Japan, South Korea and New Zealand have also introduced legislation around AVs in recent years.
North America: relatively hands-free approach
In the US, there is a limited overarching regulatory framework for AVs at the federal level. Regulation is instead legislated by individual states with varying results. Arizona, for example, has few restrictions; companies only have to notify authorities when testing AVs. California, by contrast, has a more involved regulatory process. A Canadian Senate Committee has provided 16 recommendations to Transport Canada and other stakeholders to build a coordinated national strategy on automated and connected vehicles. Provincial testing guidelines have been created for automated driving systems at levels 3, 4 and 5.28
Developing a coherent international approach
As of 2021, parts of the US, Canada, several European countries, China, South Korea, Japan, Australia, New Zealand, the UAE and Russia have all allowed some level of AV technology on their roads. The focus is on testing rather than deployment, and, given the variety of business and end-user factors, deployment locations may or may not align with testing locations.29 There have been some initiatives to harmonise international protocols. In June 2020 around 60 countries agreed to adopt a United Nations ALKS regulation for passenger cars. This was the first binding international regulation for Level-3 AVs, which was adopted by the United Nations Economic Commission for Europe (UNECE) World Forum for Harmonization of Vehicle Regulations. The regulation was initially drafted by the governments of Japan and Germany.30
Regulatory differences for AVs (level 3 and above) across major markets
Note: In the first three columns, green indicates regulations supportive of AV developments; and red means they are more restrictive. In the second set of three columns, green means there are specific AV regulations; red means such regulations do not exist.
The demand side
Most emerging technologies go through an adoption cycle based on consumer preferences, with early and late adopters. Customer attitudes to the same technologies can change as time passes, and better information becomes available.
The benefits of AVs are multi-dimensional. The obvious ones are lower costs and greater efficiency per individual journey; but value propositions can vary depending on the use case. Passenger AVs offer customers the ability to work while travelling; or more time for relaxation and entertainment (See Table 2). Logistics providers can take advantage of a wider range of delivery options. High-risk sectors, such as mining, can reduce the risk of human error by automating specialized vehicular tasks. These benefits may overlap in individual and commercial contexts.
Safety benefits will remain a critical success factor for consumer adoption
Safety concerns, both for passengers and other road users, are suggested as the key reason for low AV adoption. However, there is a growing realization that most conventional auto accidents are the result of driver error – the weak link that AVs will remove. Around 1.25 million people globally are killed and another 20 million are injured from motor vehicle accidents each year with human error accounting for around 90% of these accidents.31 Consumers appreciate that the journey to full autonomy is incremental and that vehicles will gradually evolve and increase safety. AV acceptance rates have been increasing continuously. Recent surveys show more than half of consumers willing to try full AVs, despite such vehicles still in nascent stages of development.32 Self- learning systems in AVs drive the improvement in safety over time as more and more data about the environment is collected. These improvements do not necessarily come with the release of a new version of the software but are gradually incorporated in the operating system.
The evolution of safety patterns in aircraft travel offers valuable lessons
Parallels can be drawn from the past (such as 1950s air travel), where advancements in technology boosted safety perception. Air accidents fell with more advanced and reliable aircraft (See Figure 2). Targeted investment in fault and accident analysis; more reliable safety technology; and improved design of support systems, including airports and communication systems, all played a key role in this improvement. Despite some residual risk, air travel has increased significantly over the decades; and safety perceptions are one dimension of this increase.
The acceptance and even popularity of AV technologies is becoming increasingly apparent. The American Automobile Association (AAA) suggests that 58% of its members now desire Level 1 and 2 features such as automatic emergency braking and lane assistance in their next car.33 Those who already have the technology want it to work better. Continuous development and feedback will act as a catalyst for increased safety and trust, enabling AVs to gain greater acceptance.
Note: Openness refer to the consumers' willingness to adopt AV technology
Figure 3 aggregates consumer preferences by country. Early adopters are markets where both customer openness and speed of adoption is high; late adopters are those where neither development is mature. Since AV consumer adoption is progressive, these trends are rapidly changing based on improvements made across a multitude of factors.
The market side
Cost dynamics could trigger innovative business models
With their technical requirements, AVs are more expensive than traditional cars. The average consumer is unlikely to be able to privately purchase a level 4/5 AV any time soon. Producers and intermediaries are therefore experimenting with service-based business models. These cover a variety of solutions, including ride-hailing; customized rental programs; and autonomous public transit. We will feature more on these variations in a future article.
Shared AV mobility for passenger transport offers compelling economic options
Figure 4 suggests that for some business models, such as taxi and ride-hailing services, the cost reduction brought by AV technology is significant because the largest component of operating cost - the cost of the driver – is removed. Figure 4 demonstrates that the cost per mile of an AV taxi is less than a third of the cost of a human driven (HD) taxi – hence the focus on service provision by many AV developers. If costs per journey reduce significantly, such private journeys could open-up new market segments currently dependent on other modes of transport, such as public transport users.
What is clear is that any operator with an AV fleet will need deep capital reserves both to invest in the expensive technology and to maintain their assets. Two conventional hail-ride operators, Lyft and Uber, cash flows both dented by the 2020 pandemic, scaled back their investments in AVs, selling operations to Toyota (2021) and Aurora (2020) respectively. There may be space for start-ups – if Lyft and Uber can be called such – in and around the AV ecosystem, possibly sub-contracting AVs; but only OEMs and tech giants are in a position to offer end-to-end services.
How AV journeys will be priced to ensure adequate investor returns – given the low marginal costs of one extra trip – will be a fascinating process, not least because AVs will at least initially be in competition with human hail-ride and taxi services with higher operating but much lower upfront costs. Journey pricing is an issue which could come before regulators or the courts with unfair competition claims.
This is one of the social factors that could stand as an impediment to AV adoption. Another is those who drive for a living. The Census Bureau estimates 3.5 million employed as truckers in the US in 201934. There are an estimated 1 million Uber drivers in the US, 3-4 million globally35. There will be political repercussions, at least in the short term, if these jobs are suddenly displaced by AV technology.
The economic rationale for AV logistics and good transportation could be substantial
Figure 5 shows how the adoption of AVs decreases the cost of delivering the parcel from the transportation hub to the final destination by around 40%. Also called last-mile delivery, this is the most critical and expensive part of the delivery process. This suggests cheaper prices for customers; and faster and more flexible delivery services.
AV providers will tailor products to retail, commercial and government customers. Each segment has its own requirements and service model: retail, providing individual rides for customers; commercial, having long-term vehicle leases; and governments with more specialist requirements.
Insurers need to prepare now
AVs could change the structure of vehicle ownership, and, thus, of vehicle insurance. Insurance will be marketed to owners of car fleets rather than individuals, while automated driving shifts the associated risks from human error to mechanical or computer malfunction. Liability will increasingly move to the manufacturer/designer of the equipment rather than the user. Insurers will benefit from: a much clearer and quicker understanding of what went wrong at an accident; large amounts of data on the risk profile of AVs; a swifter resolution of claims; and a (greatly) reduced accident rate.
Insurers will have to reinvent their auto coverage for AVs. Industry development suggests partnerships with major players in and around the AV ecosystem will be key. Insurers should identify and develop these partnerships early. Being involved in the early stages of AV development, insurance, as a key component of the regulatory environment, could function as an accelerator in the development of the market. AVs are being tested on roads now; insurers should do their best to get in at the bottom. We will feature more on AV insurance in later publications in this series.
In 1939, General Motors predicted that driverless cars would be available by the 1960s.36 In 2015, Silicon Valley stars were estimating 2020 as the point when fully AVs would be on the road.37 The reality is that AVs need a number of factors to align for the technology to get into gear. Five drivers will shape the future of AVs: (1) technological development of the vehicles themselves; (2) physical and digital infrastructure; (3) increased collaboration between OEMs and tech companies; (4) regulatory developments; and (5) costumers acceptance and adoption rates, of which safety is a key element. Each of these factors has progressed over recent years, more slowly than some would like, but steadily and inexorably. Despite the fits and starts in the past decades, we believe these predictions are now closer to reality than ever before, as recent advances coalesce to foster development of genuinely self-driving vehicles. Insurers should start preparing now for the important transformation that AVs will bring to the insurance industry.
1The jump between levels 2-3 moves from partial automation – with the human monitoring the driving environment – to conditional automation, where the driving is monitored by the automated system, but human override is still required. See the Society of Autonomous Engineers (https://www.sae.org/news/2019/01/sae-updates-j3016-automated-driving-graphic) for different stages of autonomous vehicles grouped into 6 stages
240+ Corporations Working On Autonomous Vehicles, CB Insights, 2020, https://www.cbinsights.com/research/autonomous-driverless-vehicles-corporations-list/
3RADAR is a detection system that uses radio waves distance, angle or velocity of objects. LiDAR is a method that uses laser to determine ranges (variable distances)
4Viknesh Vijayenthiran, "Self-driving technology company Argo AI makes lidar breakthrough", Motorauthority.com, 5 May 2021.https://www.motorauthority.com/news/1132154_self-driving-technology-company-argo-ai-makes-lidar-breakthrough
5Takafumi Uno, "Honda Launches World’s First Level-3 Self-driving Car", japan-forward.com, 9 March 2021.https://japan-forward.com/honda-launches-worlds-first-level-3-self-driving-car/
6The Society of Autonomous Engineers groups different stages of autonomous vehicles into 6 stages where level 0 has some basic driver assist/warning feature and level 5 is fully autonomous. SAE recently updated the autonomous driving levels. https://www.sae.org/news/2019/01/sae-updates-j3016-automated-driving-graphic
7Nick Carey, Paul Lienert, Tina Bellon "Britain's driverless car ambitions hit speed bump with insurers", reuters.com, 21 April 2021. https://www.reuters.com/article/us-autos-autonomous-insurance-insight-idUSKBN2C814K
8Brad Templeton, "Baidu Launches Robotaxi Service In Outer Beijing. How ‘Fully Driverless’ Is It Vs. AutoX?",https://www.forbes.com/sites/bradtempleton/2021/04/29/baidu-launches-robotaxi-service-outside-beijing--how-fully-driverless-is-it-vs-autox/?sh=41dccf5a6a05, forbes.com, 29 April 2021. According to recent updates, Waymo and Cruise have applied for permits to charge passengers for self-driving rides in the San Francisco Bay area. However, these will operate only in certain areas, under ideal conditions and speed and time limits
9For example, recently Singapore launched two autonomous bus services, https://www.traffictechnologytoday.com/news/autonomous-vehicles/two-on-demand-autonomous-bus-services-launched-in-sinapore.html. Baidu announced recently that it has got greenlight to roll out commercial self-driving bus programme in a metropolis in China, on a limited route in a specific area, https://www.barrons.com/articles/baidu-gets-green-light-for-commercial-self-driving-bus-service-in-china-51618390800
102020 Autonomous Vehicles Readiness Index, KPMG, 2020, https://assets.kpmg/content/dam/kpmg/es/pdf/2020/07/2020_KPMG_Autonomous_Vehicles_Readiness_Index.pdf
11In-depth: eMobility 2020: Statista Mobility Market Outlook, Statista, 2020
12Enabling the mass adoption of autonomous driving, PWC, 2020
13Strategic Partnership: Daimler Truck and Waymo, daimler.com, 27 October 2020, https://www.daimler.com/investors/reports-news/financial-news/20201027-dt-and-waymo.html
14Eugene Demaitre, "Volkswagen invests $2.6B in Argo AI for self-driving cars, expands pact with Ford", therobotreport.com, 12 July 2019, https://www.therobotreport.com/volkswagen-expands-ford-alliance-invests-argo-ai-self-driving-cars/
15Tyler Duvall, Eric Hannon, Jared Katseff, et. al. "A new look at autonomous-vehicle infrastructure", mckinsey.com, 22 March 2019, https://www.mckinsey.com/industries/travel-logistics-and-infrastructure/our-insights/a-new-look-at-autonomous-vehicle-infrastructure
17Autonomous Vehicle Readiness Index, KPMG, 2020
18In-depth: eMobility 2020: Statista Mobility Market Outlook, Statista, July 2020
19Global EV Outlook 2020, iea.org, 2020
20The future of autonomous vehicles, Future Agenda Ltd, 2020, https://www.iea.org/reports/global-ev-outlook-2020
22Strategy for automated and connected driving, Federal Ministry of Transport and Digital Infrastructure, Germany, 2015, https://www.bmvi.de/SharedDocs/EN/publications/strategy-for-automated-and-connected-driving.pdf?__blob=publicationFile
23Gwyn Topham "‘Self-driving’ cars could get green light for use on UK motorways this year", theguardian.com, 28 April 2021, https://www.theguardian.com/technology/2021/apr/28/self-driving-cars-could-get-green-light-for-use-on-uk-motorways-this-year
24Autonomous Vehicle Policy Framework: Selected National and Jurisdictional Policy Efforts to Guide Safe AV Development, World Economic Forum, 2020, http://www3.weforum.org/docs/WEF_C4IR_Israel_Autonomous_Vehicle_Policy_Framework_2020.pdf, https://opengovasia.com/china-issues-nationwide-guidelines-for-road-testing-of-autonomous-vehicles/
25Priyankar Bhunia, "China issues nationwide guidelines for road-testing of autonomous vehicles", opengovasia.com, 13 April 2018.
26"Smart vehicle strategy set to put China strides ahead", china.org.cn, 2 March 2021, http://www.china.org.cn/business/2020-03/02/content_75763236.htm
27Automated vehicle program, National Transport Commission, Australia, https://www.ntc.gov.au/transport-reform/automated-vehicle-program
28Global guide to Autonomous Vehicles, Dentos, 2021.
29Autonomous vehicles: The emerging landscape, Future Agenda Ltd, 2018.
30"UN Regulation on Automated Lane Keeping Systems is milestone for safe introduction of automated vehicles in traffic", unece.org, 24 June 2020, https://unece.org/transport/press/un-regulation-automated-lane-keeping-systems-milestone-safe-introduction-automated
31Julian De Freitas, Andrea Censi, Bryant Walker Smith et. al., "From driverless dilemmas to more practical commonsense tests for automated vehicles", PNAS, March 2021, https://unece.org/transport/press/un-regulation-automated-lane-keeping-systems-milestone-safe-introduction-automated
32Autonomous Vehicles Drive Interest Among Consumers, theharrispoll.com, March 2021, https://theharrispoll.com/self-driving-vehicles/
33Ellen Edmonds, "AAA: Today’s Vehicle Technology Must Walk So Self-Driving Cars Can Run", newsroom.aaa.com, 25 Feb 20201, https://newsroom.aaa.com/2021/02/aaa-todays-vehicle-technology-must-walk-so-self-driving-cars-can-run/
34Jennifer Cheeseman Day and Andrew W. Hait, "Number of Truckers at All-Time High", census.gov, 6 June 2019, https://www.census.gov/library/stories/2019/06/america-keeps-on-trucking.html
35Melissa Berry, "How Many Uber Drivers Are There", therideshareguy.com, 20 February 2021, https://therideshareguy.com/how-many-uber-drivers-are-there/
36At the 1939-1940 New York World’s Fair, General Motors’ interactive Futurama exhibit predicted high speed automated roadways in 20 years. For example, see “The Original Futurama”, Wired Magazine,
37Christina Rogers, "Google sees self-drive car on road within five years", marketwatch.com, 14 Jan 2015, https://www.marketwatch.com/story/google-sees-self-drive-car-on-road-within-five-years-2015-01-14