Technology Transforms How Insurers Calculate Risk
By: Quentin Hardy (NY Times) April 2016
SAN FRANCISCO — The world is a dangerous, unpredictable place — not a bad thing, perhaps, if you’re in the business of selling insurance, making loans or figuring out ways to prevent mishaps.
But these days, technology is changing some of the calculations around risk, whether for car insurance, life insurance, flood insurance or even vacation-related accident insurance, and slowly but surely disrupting the financial core of the insurance business.
Take the old saw that younger drivers are always the riskiest to insure, for example. Is it always true?
“A 22-year-old driver who parties a lot and drives 10,000 miles a year, or a 40-year-old teetotaling mom who drives 40,000 miles? I might think the mom is riskier,” said Dan Preston, the chief executive of Metromile. “Speeding a lot or braking hard, unless you’re going 100 miles an hour, it doesn’t make much difference.”
Instead, all those extra miles the mother is safely and soberly driving mean she is more often where the accidents are happening. Or, as Mr. Preston put it, “Seventy percent of the variants around expected losses can be derived from mileage.”
Financial technology is not just changed by faster computers with more powerful software. Technology also changes things like mobile computing, data collection and what people can measure. Put those together creatively and you change any business involved in calculating risk — in other words, most of the financial world.
On a planet covered by location-sensing satellites and cell towers, it is not difficult to put a device in a car and insure people based largely on mileage. That is the basic idea behind Metromile, which began in Oregon in 2012, and now sells insurance in seven states.
Customers fill out the usual information, like age and location — the things traditionally used to guess the likelihood a car will be wrecked or stolen. That data determines a Metromile base rate for insurance, and premiums rise based on miles driven. People who stay below the national average of about 12,000 miles annually can save substantially, compared with traditional insurance.
People are, in effect, agreeing to be observed in exchange for the savings, which enables Metromile to add services. It has become a way for Uber drivers to stay insured while they are driving for money, something their traditional policies may not cover.
Metromile’s location-based knowledge lets it warn drivers to remove their cars from a street that will be swept soon, helping them avoid a ticket. The tracking device can hook into a car’s computers, warning the driver when the car might need maintenance.
While people may be uncertain about government monitoring, they seem to be willing to have companies looking over their shoulders — at least if it saves them money.
It is something insurers want badly, said James Whittle, chief claims counsel for the American Insurance Association. “Lots of insurers are drilling down to the precise granularity of risk,” he said.
Expect more. Cellphones, sensors and cheap computing make it possible to value all sorts of objects, including cars, houses, cameras and skis, while gaining insight into how they are likely to be used next. That changes the betting for the risk industry.
“Metromile is the start of a broader shift,” said Mark McLaughlin, global insurance director for IBM. Insurance companies “want to shift from paying when something bad happens to helping you prevent something bad happening to you,” he said.
He predicts a time when a car insurer might learn drivers’ commuting routes and warn them if there is a chance of a hailstorm ahead, even text them a Starbucks coupon to encourage them to wait out the risk. Last October, IBM bought the digital assets of the Weather Channel, perhaps in anticipation of such a service. IBM declined to discuss future products.
Many traditional insurers are already adopting this observational approach and using discounts to sway behavior. Some examples include Progressive Auto Insurance, which has a service similar to Metromile, called Snapshot.
The company Vitality has united with several insurers to offer people lower rates in exchange for logging their health data. It gives people things like Fitbits and Apple Watches to automatically track their activities. State Farm Insurance, in conjunction with the home security company ADT, offers discounts for people who get home monitoring systems.
Insurance, which is after all a business where people take money with a promise to maybe pay for something someday, is a highly regulated enterprise that historically favored slow-moving companies with lots of capital. In this case, however, Mr. Preston says he has an edge on established competition.
“The challenge for the incumbents is that the low-mileage customers are most of their profits,” he said at Metromile’s headquarters here in a converted parking garage. “They are subsidizing the high-mileage drivers,” he said. “Those are the people we attract.”
Metromile’s co-founder and board chairman is David Friedberg, an early employee at Google, who also founded what became the Climate Corporation. Climate, which used data on things like rainfall and soil quality to insure several types of crops, was bought by Monsanto for more than $900 million in 2013.
The new methods of measurement are affecting lending, too. ZestFinance looks at thousands of nonstandard variables, including technology use and how people complete forms, to assess a lender’s risk of things like fraud and loan default. Upstart is a lending service that targets younger borrowers, setting rates based on college major and job history, among other factors. Not surprisingly, ZestFinance and Upstart were also founded by former data specialists.
In March, Palantir, a privately held company that has used large-scale data analysis for a variety of tasks, including catching terrorists and determining how banks can efficiently sell subprime bankruptcies, formed a joint venture with Credit Suisse that will use large-scale data analysis to detect insider trading.
The next move may be not just microslicing houses and drivers’ behavior, but also individual objects in a house. Trov, a start-up in Danville, Calif., uses a smartphone app to photograph and value objects like skis and cameras. Customers can then buy insurance for a weekend of skiing. It might even sense when a policyholder is near the slopes and ask if the policyholder wants to activate the insurance.
“The next generation understands the atomization of everything,” said Scott Walchek, the founder and chief executive of Trov. Atomization means slicing things to their smallest components, in this case not insuring a whole houseful of goods but individual objects. “Insurers have to get ready for customers who expect to have their ownership of everything, how they use things, understood. On-demand everything is the zeitgeist.”
