© Reuters

“It was kind of a space age experience,” was how one early user of Google’s new driverless cars described the encounter. It is easy to see why — the idea of a fleet of cars driving around with no need for any human control seems like science fiction.

Yet according to Fehr & Peers, a consultancy, half of US traffic will be autonomous at some point between 2040 and 2050 as major manufacturers refine their technology and make it affordable for the mass market.

For the insurance industry, “it is going to have a major impact”, says John Cusano, senior managing director of global insurance at Accenture.

Motor cover is one of the mainstays of the insurance industry. According to Autonomous Research, an investment research company, it accounts for 42 per cent of global insurance premiums. The biggest cause of accidents — and hence the need for cover — is driver error. This accounts for about 90 per cent of accidents. Cut out the humans and their errors, so the theory goes, and the need for insurance will fall.

Researchers at Autonomous have analysed the potential costs to the industry. Claims frequency — the percentage of those who make a claim in a year — will fall from 9 per cent now to 2.4 per cent by 2060, they estimate, offsetting a slight rise in claim costs. The result, they add, is that the global average motor premium will fall from $541 now to $334 by 2060.

In some developed markets, where the adoption of driverless technology may be quicker than in developing markets, the change could be even more stark. Autonomous expects UK premiums, for example, to fall by 63 per cent by 2060, causing profits for the motor insurance industry to fall by 81 per cent.

The decline in premiums is expected to start well before then. Although fully driverless vehicles are still in the test stage, a new range of gadgets are already making cars much safer to drive. Over the past decade, manufacturers have switched focus from making accidents less harmful when they happen — by installing airbags, for example — to ensuring that accidents do not happen in the first place.

Tools such as blind-spot detection, parking assistance and cruise control are already relatively common in some parts of the world. Swiss Re reckons that such systems will reduce accidents on motorways by 16 per cent by 2020, and on other roads by 12 per cent.

The next generation of gadgets will include systems that enable cars to gather information from other road users and from infrastructure such as traffic lights to help spot and evade hazards. Swiss Re expects these more advanced systems will lead to a fall of 45 per cent in accidents on motorways and of 28 per cent on other roads.

“Increasing sophistication and safety capabilities are being put in,” says Mr Cusano. “So the amount and severity of accidents, and the insurance required, will decline over time.”

From these advanced systems, it is just a small step to fully driverless technology that will enable drivers to take their hands off the steering wheel and, eventually, enable the manufacturers to remove steering wheels altogether.

Plenty of companies are already working on prototypes. Google and Tesla are among the most high-profile developers of driverless cars, but traditional manufacturers such as Volvo, Toyota, Ford and GM are also working on prototypes.

However, even if the technological challenges can be overcome easily, there are a host of other problems that must resolved before science fiction can become a reality. They range from the way that driverless vehicles interact with those driven by humans, to the legal framework and the way that these new vehicles are insured.

There are also public attitudes to consider. According to a survey by the UK’s Institution of Mechanical Engineers published in May, 55 per cent of people said they would be unlikely to want to be a passenger in a driverless car.

55%Proportion of UK public who say they are unlikely to want to be a passenger in a driverless car

Initiatives are under way to work through these obstacles. The UK government has set up several different working groups to assess the technological, practical, legal and insurance implications of driverless cars. Axa, the insurer, is part of three of those groups. The Venturer consortium, based in Bristol, is investigating a range of technological, legal and insurance issues. The UK Autodrive consortium is planning to set up a system of driverless cars that can be used on the town’s roads. Meanwhile, another consortium, Flourish, is looking at user needs.

“We believe that driverless cars will be an essential part of the future, and we don’t want the insurance industry to be blocking it. We want to be part of it,” says David Williams, technical director at Axa, who is involved in the projects. “By being involved early we’re getting an early view of the technology.”

One of the big questions for all of the consortiums to consider is how driverless cars will be insured. Motor insurance tends to be a personal liability product for drivers. But if the cars are driving themselves, vehicle manufacturers may need to insure themselves against the risk that one of their vehicles malfunctions.

Researchers at Autonomous think this will radically change the way motor insurance policies are sold. Today, 87 per cent of motor insurance policies are personal, rather than commercial. By 2060, says Autonomous, just 40 per cent of policies will be personal with the rest will be a mixture of commercial policies and product liability policies.

For some insurers the move to driverless cars is just the start. XL Catlin, an insurer, has just teamed up with Oxbotica, an Oxford university spinout that focuses on robots, to explore the potential for autonomous control in a range of vehicles and systems, not just cars.

“It is going to challenge our existing concepts of liability,” says Mike McGavick, XL Catlin’s chief executive. “These are going to shift from being individual risks to systems risks. It is a whole new insurance pool that didn’t exist before.”

Copyright The Financial Times Limited 2024. All rights reserved.
Reuse this content (opens in new window) CommentsJump to comments section

Follow the topics in this article

Comments