Insurers Examine Big Data to Identify Risks

Insurance companies are increasingly turning to “big data” from satellites, social media and even cigarette sales at gas stations to help identify risks and build up customer profiles.

Insurers and reinsurers hope that real time analysis of data about personal behavior will enable them to project damage claims and fine-tune prices to fit the risk being covered, and also help them spot fraud.

Troves of data are being collected via the technology phenomenon known as the Internet of Things, where cheap, network-connected sensor devices are embedded in all manner of industrial equipment, transport vehicles, appliances in the home and even the health monitors and smartwatches that consumers have begun to wear on their wrists.

Hamilton Re, a new Bermuda-based reinsurer, hopes that heavy data-crunching technology will give it an advantage over rivals and boost its bottom line.

“If we do it successfully, we ought to be able to deliver our products at lower cost with an improved loss ratio,” said Bob Deutsch, chief strategy officer for the group.

“In underwriting, you have a better ability to plot whether you’ve got a concentration of risk in certain aspects of tornado alley,” Deutsch told Reuters on the sidelines of the annual conference of the reinsurance industry in Monte Carlo this week.

Insurers have long struggled with flawed information on policy and claims forms, according to Maurice Tulloch, CEO of UK and Ireland General Insurance at Aviva, which has about 500 professionals working on data analytics.

“Of the data we get back, a third of it is generally incorrect,” Tulloch told the conference, referring to traditional data collection.

Information from satellites, medical data from fitness devices, social media activity, construction plans, rainfall, storm drain systems, energy efficiency, and cameras monitoring road surfaces can all be put to use by insurance companies.

The growing mountains of data available for analysis could raise knotty privacy questions, although, for the most part, the data the insurance industry is looking to pore over is aggregate data about collective behaviors rather than information that can be linked back to individuals.

INSTANT INSIGHT

Insurers and reinsurers are investing increasingly in telematics, for example using data from smart phones to track the location and speed of vehicles, which could allow them to warn drivers of dangerous behavior or intersections, possibly cutting down on accidents.

Big data can give insurers improved understanding of risks but such predictive modeling can yield some unexpected results.

Swiss Re’s Chief Underwriting Officer Matthias Weber pointed out that smokers typically buy only one pack of cigarettes at a time, very often from petrol stations.

“If you know from your technology that somebody goes to the gas station once a day, including the weekends, it might be right that (that person) is a smoker,” he said.

Unless someone is driving 360 miles (579 kilometers) a day to their job, there is no reason to stop for petrol every day, so it might be a reason to enquire further, he said.

The growth of big data and other technological advancements is not without risks for the insurance industry.

Google, for example, possibly the biggest data miner of them all, could offer to do analysis for insurers or even become a competitor and offer insurance of its own one day, insurers said.

Automobile manufacturers’ progress toward developing driverless cars could crimp the insurance segment that represents 40 percent of property/casualty business in many countries and cause results in the rest of insurers’ portfolios to become more volatile.

“What happens when cars don’t hit each other?” asked Bryon Erhart, a senior executive at broker Aon Benfield.

(Additional reporting by Eric Auchard; Editing by Susan Fenton and Thomas Atkins)