Driving Down Risk

Driverless cars could not only wipe out the majority of attritional motor losses, but could also force insurers to answer one of life’s most challenging philosophical problems, says Dan Ascher.

The problem goes something like this: a runaway trolley is hurtling down a track headed straight for a family of five who have no way of escaping the cart's deadly path.

But they're in luck, there's a set of points that allow a controller to divert the trolley onto a separate track where there stands a single 90-year-old man.

What should the controller do?

Intuitively, we want to say pull the lever. It's better to save five lives and kill one. But that's the dilemma. Is inaction in allowing five people to die worse than actively killing one?

And, as with all great philosophical problems, there is a counterexample that helps us answer that question.

A doctor is working overtime as five of her patients circle the drain. Each one is in need of a different organ: one a heart, another a liver, and so on.

A drunk stumbles into the hospital. He's healthy but intoxicated. Should the doctor carve him up and distribute his organs, thus killing one but saving five?

Our intuition says no.

And so we come to driverless cars. One estimate says that less than 25 years from now, three quarters of cars will be autonomous.

It's not hard to imagine a scenario where a driverless car is whizzing along a high street when a child runs into the road to retrieve a stray football, forcing the car to make a decision on what kind of accident to cause.

"In a multiple-accident scenario does the car decide to injure the 90-year-old man or the 10-year-old child?" asks James Woods, a partner at law firm Mayer Brown.

"We don't have answers for it yet," he continues. "As we move through the driverless age, we will encounter moral dilemmas as well as physical dilemmas."

There are very real problems that driverless cars haven't been programmed to solve, Woods says.

And, as is often the case, to find the person with the answers we have to follow the money, which takes us right to the underwriter's doorstep.

Join the club

The underwriter is always striving towards perfect information in a bid to accurately assess the risk. So when Tesco Bank wants to sell insurance it incentivises prospective insurance customers to take out a Clubcard. By giving the insurers access to information about their shopping habits, customers are, in effect, allowing underwriters to profile them based on their weekly shop.

For example, if you walk into a Tesco and buy a copy of Gardener's World magazine and a pack of Werther's Original and collect the points on your Clubcard, Tesco is likely to consider you a better risk than someone who picks up a copy of Top Gear magazine and a 24-pack of Fosters lager.

While this information does not enable underwriters to know exactly how an individual will react in any given situation, by looking at how shopping habits correlate with claims data carriers can establish how a driver is likely to act.

But that can only be a best guess. Driverless cars present a very different challenge.

With autonomous vehicles, the underwriter knows exactly how the car will react in any number of given scenarios.

However, what he or she does not know is how the world will respond and interact with these robots that are expected to dominate our streets less than a quarter of a century from now.

And while they can assess the risk of an individual driver based on their shopping preferences, when autonomous vehicles hit the road underwriters will be forced to profile the behaviour of everyone around those vehicles other than the driver, about whom they have no information.

Perfect driver

The reality is that driverless cars will be programmed to behave like the perfect driver.

They will indicate before turning, stop fully at T-junctions and never undertake. But perhaps most importantly, they will have the ability to brake more sharply than a car driven by a human, which will prevent them from rear-ending other vehicles - reducing attritional losses.

However, historically the insurance industry has been reactive rather than proactive and some have speculated that the advent of the autonomous vehicle could kill the motor market as personal lines policies become commercial liability programmes.

And while motor carriers sit back and wait for an external force to turn their business on its head, a handful of insurance metaphysicians are looking to shake the industry up from the inside out.

These revolutionaries want to strip the industry back to its basic principles.

"Anybody who's been in this space knows that the truth is it has been - almost by definition - inefficient for quite some time," says former ANV boss Matt Fairfield.

Fairfield is reported to be working on a start-up initiative that will include a managing general agency (MGA) business to write specialty risks on behalf of non-traditional sources of capital.

"People need to get used to the new reality," he warns.

The industry veteran says there is "no doubt" that casualty risk can be structured so that it will appeal to the alternative capital providers.

"We think the entire casualty market eventually will be able to be consumed by the alternative market," he says.

"I absolutely think alternative capital will eventually be structured to write all risks and write even the existing property risk a bit better."

He equates an insurance contract to a derivative, arguing that both are legal contracts derived from the value of something else.

"We have been able to monetise derivatives before, whether they be tied to financial instruments [or] commodities," he goes on. "There is a decent amount of data to do the same thing in insurance."

Crosstown traffic

"Right now, insurance companies are the largest arbiters between the clients who buy the policies and the capital markets which provide the capital," Fairfield says, arguing that pricing power is based on distribution.

"Many of the brokers are in the underwriter's space already," he points out, adding that the existing insurance food chain is inefficient.

At present, for a retail policy that chain could include a retail broker, a wholesale broker, an MGA, the insurance company, a reinsurance broker, a reinsurer, a retro broker and a retrocessionaire.

"You could have, from one broker, seven mouths being fed. Or one broker having seven bites at the cherry," Fairfield explains.

"An awful lot of frictional costs are being created in the transaction," he says, predicting that the market will become "more and more" efficient.

According to Fairfield, in a self-governing, post-crash market there will be less emphasis placed on financial brands and less reliance on bodies such as ratings agencies.

"Before 2008, in financial services people brought brands. People thought AIG would never go bankrupt - until it did," he says.

"It's not about alternative capital versus traditional capital anymore. The race is on for everyone in this marketplace to have the most efficient capital base as possible.

"It doesn't matter what you call yourself. You better have the most efficient access to capital you possibly can because when you give in to an efficient transparent marketplace [then] your clients have a lot more choice."

If life gives you lemons...

But it isn't only the reinsurance market that is viewed as ripe for an overhaul.

"Insurance started off as a fabulous social good," says Daniel Schreiber, CEO and co-founder of peer-to-peer insurer Lemonade.

Schreiber claims to have known nothing about insurance before moving into the sector, giving him a fresh perspective on the market.

"It started off as communities coming together with a covenant of 'one for all and all for one'," he says.

"And you saw that if any mishap befell any one member of a certain group, by pooling their risks the pool at large would take care of them."

But, he says, urbanisation and industrialisation have meddled with the structure underlying community-based insurance, which is based on the principle of mutuality.

"You deal with one big - oftentimes faceless - corporation, and much of the social fabric has dissipated or been replaced by the machinery of Wall Street or Madison Avenue.

"Today when I make a claim off my insurance company, I'm fighting my insurance company for the same coin at the end of the day."

"It's a zero-sum game," he continues. "Every dollar they pay me in my claim is a dollar off their bottom line - that's the way insurance works."

"Insurance may be unique - certainly it's unusual - in that it makes its money by disappointing its customers," he adds. "If it were to really delight them all, it would go broke."

"If you wanted to design a system to bring out the very worst in human beings it would look a lot like an insurance company: alienating, adversarial and anonymous."

But Schreiber is cagey about the detail of how Lemonade will actually work, saying he is "saving the final piece of the puzzle" for when the product is launched in a few months' time.

Responding to the suggestion that a rose by any other name would still smell as sweet - or, more precisely, that manure by any other name would still smell as bad - Schreiber says: "AirBnB doesn't smell like the Hilton, Uber doesn't smell like a yellow taxi and Wikipedia doesn't smell like [Encyclopaedia] Britannica."

Keep your peers close

Enrico Bertagna, senior vice president of business development at Allied World Europe, says peer-to-peer insurance is really a very simple concept.

"Insurance is based on two key principles," the executive explains. "Firstly, it is founded on mutuality and the second pillar is statistics.

"The only difference between the old mutual system and peer-to-peer insurance is that peer-to-peer insurance is supported by technology, and particularly social media, in order to identify and group the members of the mutual fund. Technology is very powerful in profiling the members in terms of their behaviour and specifically in preventing fraud.

"So, if you compare peer-to-peer with the old mutual system, the old mutual system was built on pure statistics, but statistics only allow underwriters to look backward," he adds.

"With social media you can actually profile your clients or the members of the fund based on their behaviour and in real time."

With peer-to-peer insurance, the sums that the members pay into the mutual fund are used to pay the frequency claims, while some is used to purchase an excess cover on top of that provided by the mutual fund.

"This model is at the same time a threat and an opportunity for the traditional insurance market," Bertagna argues.

He says that for the personal lines market the new system could pose a threat, but that it is also a huge opportunity for the traditional excess carriers, which may see business that they would not otherwise have been shown.

Sure enough, in February Lemonade announced that it had received reinsurance backing from Berkshire Hathaway's National Indemnity, Everest Re, Hiscox, Munich Re, Transatlantic Re and XL Catlin.

The transaction will see the industry's newbies pitch their underwriting philosophy against that of the stalwarts - generating perhaps yet another existential crisis for the insurance sector.