July 18, 2018 by Sara Tatelman
In November 2017, an autonomous shuttle bus was driving its pre-determined route through downtown Las Vegas for the first time. When it noticed a delivery truck backing out of a laneway, the shuttle stopped, but the truck driver didn’t and the vehicles collided.
Though nobody was hurt, the shuttle was taken out of commission for the rest of the day. “Had the truck had the same sensing equipment that the shuttle has, the accident would have been avoided,” the City of Las Vegas noted on its official blog.
Just as it took years for horse-drawn carriages to give way to Model-T Fords, it will likely take years for human-operated cars to give way to autonomous ones. And that transition period—when two types of drivers have to share the road—will be the trickiest for insurance.
“For the last several decades, it has always been clear that when there’s a collision one of the drivers is at fault, so let’s figure out how to decide who is at fault,” Paul Kovacs, researcher at the Insurance Institute of Canada, said at the October 2017 Standing Senate Committee on Transport and Communications.
“In 40 or 50 years, if all the cars are fully automated, always the manufacturer will be at fault. In the period in between, which will be with us for a while, there will now be more disputes and uncertainty,” he added. “We will have to figure out, in each particular case, how much was the driver’s fault and how much was [the] technology’s fault, and that will not be easy to do, especially if there is not data from a black box or something to help clarify that situation.”
“We’re going to see more of a commercial liability product exposure because they are truly autonomous [vehicles] working in an autonomous world.”
Currently, existing auto policies are sufficient to cover autonomous vehicles in testing phases, Maureen Brown, vice-president and underwriting and operations lead at Munich Reinsurance America in Princeton, N.J., told Canadian Insurance Top Broker in an email. But when vehicles without steering wheels and brake pedals end up on the road, coverage will have to shift from auto liability to product liability.
In 2013, 87% of insured auto losses in the U.S. fell under personal lines policies, KPMG noted in a 2015 report. By 2040, it predicts that number will fall to 54%, with the remaining insured losses coming under commercial auto policies as well as product liability coverage.
Part of that will be due to a shift in vehicle ownership patterns, at least in major cities. Autonomous vehicles will likely be run on “a subscription model” similar to Uber, says Sami Shaker, a physicist and client relations manager at Kodsi Forensic Engineering in Mississauga, Ont. “In the near future, you will use your phone to summon your vehicle, which will be a pod. It will just pick you up and take you wherever you want. So millennials are no longer going to have to pay for expensive insurance.” Instead, insurers will have to cater to companies that own fleets of these autonomous pods.
“We’re going to see more of a commercial liability product exposure because they are truly autonomous [vehicles] working in an autonomous world,” says Rick Orr, president of Orr Insurance and Investment in Stratford, Ont. But while fleets will likely be popular in traffic-heavy areas like Toronto, he predicts smaller communities will see more of a mix between autonomous and traditionally driven cars.
“When I’m just zipping around Stratford, I want to drive,” Orr says. “So how does an insurer deal with that? That’s where you’ll see innovation come into the market.”
Despite this rapidly changing ecosystem, the industry should keep coverage options simple for consumers. “The aim would be to make it look as much like the current form of car insurance as possible,” Ben Howarth, senior policy advisor on motor and liability at the Association of British Insurers in London, says of new U.K. legislation around insuring autonomous vehicles. Passed in February 2017, the Vehicle Technology and Aviation Bill allows drivers to keep buying one auto policy from one insurer; injured third parties would continue to claim through that insurer as well.
Behind the scenes, however, policies would become more complex. Under the legislation, insurers can recover costs from manufacturers if the vehicle technology, not the human driver, caused the accident. “Any other person liable to the injured party in respect of the accident is under the same liability to the insurer or vehicle owner,” the bill reads.
Having the insurer and manufacturer sort out liability without involving the consumer makes sense, Ryan Stein, director of policy at the Insurance Bureau of Canada, said at the Senate Standing Committee in October. “The idea is to speed up the process for the injured individual, and there is a lot of merit to that.”
Autonomous car manufacturers such as Volvo, Google and Mercedes have promised to take full liability should one of their cars get into an accident while in driverless mode. However, it’s unclear if the manufacturers will self-insure, partner with an insurance company or simply reimburse claims. In any case, Howarth doesn’t think this will change the way British drivers buy insurance. “They want to be able to shop around every year and get the best possible deal, rather than simply taking an off-the-shelf product from a single manufacturer,” he says, adding some companies already offer policies linked directly to their vehicles. But those are “very rare” and “not particularly competitive” in terms of price, so he doesn’t anticipate significant uptake.
Insurers also have to grapple with how they can develop policies when they have little to no data to work with. One solution is to partner with various pilot projects. In the U.K., for instance, RSA is working with driverless pods as part of the Greenwich Automated Transport Environment project, and AXA is part of a consortium examining handover — switching the driverless mode on and off — through the Venturer project. While they’re responsible for compensating anyone hurt during testing, “the main reason the insurers are in those trials is for the experience and to learn about the technology and guide [manufacturers] about how they think about the risk going forward,” Howarth says.
Recent developments around ride-hailing insurance can also offer lessons for autonomous vehicle policies. For example, in July 2016, Intact partnered with Uber to cover their Ontario and Alberta drivers while on the job. “Rather than relying on a strictly personal lines solution, where licensing bodies and regulators would have to monitor and check if every car is properly insured, we created a structure where every Uber ride would be automatically insured,” Obaid Rahman, vice-president of corporate specialty solutions at Intact, told Top Broker in an email.
Just as an Uber driver needs a personal policy one moment and a commercial policy the next, an autonomous vehicle owner will need different coverage from moment to moment, depending on whether or not they are operating the car. So, insurers could work with specific manufacturers to make sure all users are covered when the car is in autonomous mode.
But what if the car is always in autonomous mode? According to the U.S. National Highway Traffic Safety Administration, 94% of traffic accidents are due to human error, so should the amount of traditional driving a policyholder does affect their premiums?
“I think it’d be very interesting for insurers to say okay, here’s the premium we’re going to charge— based on how often your car is in autonomous mode we’ll apply X discount, versus how often it’s in nonautonomous mode,” Orr says.
“The aim would be to make it look as much like the current form of car insurance as possible.”
He also points out that telematics tracking can help determine how friendly a particular neighbourhood is to driverless vehicles. If a car in autonomous mode keeps having to brake suddenly, for example, “I would think you’re in an area that might not be as discount-friendly as an area where there’s so many autonomous vehicles that they’re all working together better.”
While autonomous vehicles will lead to fewer accidents, the cost per accident will likely skyrocket. For example, in the 2015 report, KPMG predicted annual accident frequency per vehicle rates will drop nearly 80% by 2040. But the cost of each accident will jump from US$14,000 to US$35,000 in the same time period, since the technology becomes more expensive to repair.
“The question isn’t just about cost” of the actual repair, Howarth says. “It’s about making sure we’ve got the skills in the repair network to repair the cars efficiently. Can we repair the car locally like we do today, or is there going to be one big central place?” That could mean having to transport cars across the country, and having to supply replacement vehicles for longer periods of time.
However, the cost of personal injury is significantly more than the cost of repairs, he adds. So “if we can reduce those, A—it’s good for society because fewer people would be injured; and B—it would have a more significant effect on the overall cost of insurance. […] But it’s too early to say for certain.”
In terms of developing coverage for autonomous cars, Orr predicts insurers will have to do battle to stay in business. “They are going to have to come up with products, with discounts, whatever it is, whichever direction it goes, but they will certainly be fighting to stay relevant in the future.”
As for brokers, he thinks their role will be as important as ever. “Do you want an advocate that helps you? You’re sitting on the side of the road and you call your broker: ‘I’ve just been hit by a car and there’s nobody in it,’” he says. “Consumer education is still going to be required, and brokers are in every small town across the country, and they’re still in the best position to deliver that.”
Copyright © 2018 Transcontinental Media G.P. This article first appeared in the May edition of Canadian Insurance Top Broker magazine
This story was originally published by Canadian Insurance Top Broker.