Go up one level Go home

1 Factors affecting market penetration

There is a common view that the technology for vehicle automation is close to being operational. This is not the same as AVs being very close to being allowed on public roads. It is true that many car manufacturers have announced the imminent availability of their own version of CAVs. For example, Mashable compiled a set of announcements as follows:
Manufacturer
Automation
Available for delivery
Tesla
Level 4
2018
Audi
Level 3
2021
BMW
Level 4
2021
Ford
Level 4
2021
Honda
Level 3
2020
Nissan
Level 4
2020
Volvo
Level 4
2020

Despite this uncertainty, decision-makers and planners need estimates of when HAVs will be available and how quickly they will become a significant part of traffic. This has encouraged companies to offer these estimates. For example, McKinsey predicts (in May 2017) that "advanced levels of autonomous driving" (likely to correspond to highly automated vehicles) are 5-10 years away. Similarly, KPMG and SMMT project that Level 4 automated vehicles will be available in 2025.

The methodologies available to estimate the speed of adoption after they become available are limited and flawed: there is no market research tool that can actually put people in the real context to choose between conventional vehicles and HAVs at a given price. Instead, the common approach is to compare HAVs with other technologies; however, this is particularly difficult. Smartphones are often used as the comparable technology: like HAVs, smartphones had a premium price but were adopted very quickly due to their advantages were very significant. However, despite their price phones are much cheaper than cars, so the similarity is limited. Comparing HAVs with conventional car adoption is also flawed as adoption of the latter was significantly constrained by the rate of income growth.


Therefore, there is significant uncertainty in how quickly these vehicles will be adopted and become a relevant share of the market. The consensus at the Automated Vehicles Symposium 2014 was that widespread use of these technologies may only materialise some 15 years after first introduction. More recently, the Victoria Transport Policy Institute (VTPI)estimated (based on the adoption profile of comparable technologies such as automated gearshifting)that 50%penetration might not happen until 30 years after introduction. In contrast, KPMG and SMMT forecast that 80% of new cars will have Level 3 capabilities in 2025 (seven years after the projected initial deployment) and that Level 4 features will be in 25% of new cars by 2030. McKinsey goes as far as predicting that 50% of new cars in 2030 will be highly automated. Later in this section, we show a range of estimates from the Expert Panel.

Factors that will affect the speed of market penetration include:

  • The costs of AVs, compared to otherwise similar manually driven cars, as discussed below.
  • The ownership model for AVs. HAVs will save money when used as MaaS, as the driver will not be needed. However, HAVs may not save money - at least initially - for those opting for owning them, despite savings in insurance and parking costs. This is according to VTPI and this calculation ignores other advantages of HAVs that may justify the premium paid to purchase them, at least to high-income individuals.
  • Public attitude to AVs. As stated above, we can only gather imperfect perceptions of current preferences. A recent review commissioned by the UK Department for Transport found that stakeholder attitudes affect not just adoption of technology, but also affect the development phase of it. The same review highlighted that social and behavioural issues of AVs receive much less attention in published research than the technical aspects. This is consistent with the Automated Driving Roadmap published by the European Road Transport Research Advisory Council, which devotes no more than one paragraph to users' and societal acceptance.
  • Government support for AVs. Governments will not just introduce regulations, they can also opt to support the adoption of HAVs (on the grounds that they will save lives) and introduce measures such as changes in fuel tax to support electric HAVs. A review by Bryant Walker Smith highlights the various mechanisms by which governments can support AVs. These include:
    • Maintain road/sign infrastructure
    • Update vehicle registration to include automated capabilities
    • Leverage procurement to make fleets owned by or associated with government automated
    • Enforce driving laws (for example distraction and intoxication) more strictly
    • Standardise data management
    • Support deployment of wireless networks

A recent position paper from the Eno Center for Transportation supports the need for governments to maintain roads in good repair. However, it sees wireless networks as a distraction and recommends that investment in this is limited to testing its benefits. The discrepancy between these two, equally credible sources may be due to a difference in focus, between the immediate benefits of purely automated vehicles (that don’t need wireless networks) and full system benefits of CAVs (that do need wireless networks).

2 The profile of premium costs over time

The general expectation is that early models will have a relatively large premium cost. However, this will quickly decline as experience is gained and production rates are ramped-up. Most of the additional costs will be in sensors and the processing power to guide the vehicles. The cost of Light Detection and Ranging (LiDAR) was estimated by the Institute of Transportation & Development Policy and UC Davis to be near US$75,000 in 2014, and by early 2016, Velodyne began selling a form of LiDAR for US$500 per unit to Ford. The actuators for acceleration, braking and steering will already be in use for lower levels of autonomy and most are in use today.

We asked our Expert Panel to estimate this premium, over a conventional car, five years after HAVs become available, close to the time when they would be reaching 10% of the fleet.

The premium was not expected to be remarkably different in different parts of the world. The overall average was US$6,700 (~£5,200 or €5,950), however, Australasia expects it only to be US$5,600 and the US and Canada even less at $5,100 (~£4,000 or €4,500). Tesla lists “Enhanced Autopilot” for $6,000 and “Full Self-Driving Capability” for an additional $4,000 on its website when you configure any of their cars. While cars can be ordered with this functionality, Tesla has not enabled full self-driving capability on the software at this time.

Pic 10.png
Pic 10.png


This premium will be relevant for individual purchases but less so for MaaS fleets like Uber. The premium will be, by all accounts, less than a few months of a driver’s income in the developed world. The situation will be different in emerging countries where the cost advantage of CAVs for mobility companies will be less marked (and the infrastructure and signalisation will be, perhaps, of a lower quality).

3 Time profile of sales and share of fleet


There are several estimates of the rate of penetration of AVs in different regions of the world. As an example, a recent report published in the UK by the Transport Systems Catapult provides estimates for light and commercial vehicles under different assumptions or expectations of market development. The three scenarios considered are presented in the following table:
TSC_Market_Forecast_1.png

This is also shown graphically as:

TSC_Market_Forecast_2.png

Given the uncertainty about the availability of HAVs, the Expert Panel was consulted on this and asked for the time when they will be 10% and 20% of the car fleet rather than total sales. The mean expectation for the availability of HAVs was 2023, however, experts from the US and Canada expect them to be available by 2021. This is consistent with reports from manufacturers that they will be available by 2020-2022. European experts were less optimistic with an average year of 2025 and Latin Americans expect them even a year later than that.

The next table shows the average and standard deviations of the views of experts with respect to these three questions.


Pic AV available.png
Pic AV available.png



The following chart shows the cumulative distribution of answers to that question.

Note that the Expert Panel was consulted in late 2016; expectations tend to moderate as the target date gets closer (which may skew these results).
Pic 5.png
Pic 5.png


Experts from the US and Canada expect CAVs to reach 10% of the fleet on average seven years after they become available. Five years later they expect them to reach 20% of the fleet. Australia, New Zealand and Western Europe expect CAVs to reach 10% of the fleet eight years after they become available and 20% five or six years later. The rest of the world expects a faster rate of adoption, only four years to reach 10% and then six for 20%. It is noticeable that the range of values is wider for reaching these thresholds indicating different views about the speed of adoption.
10-20offleet.png
10-20offleet.png