Big Data means insurers can find out more about us than ever before. So how could the growth in available information affect workplace insurance? John Greenwood investigates
Smartphone GPS signals and pedometer motion readings, electronic purchase transaction records, digital photos and web-browsing history – every day we are all creating lots of data in which insurance companies would be very interested.
The amount of data in existence is growing exponentially – IBM reckons that 90 per cent of the data in the world has been created in the past two years. So massive is the growth in information generated about individuals that the term Big Data barely seems to do it justice.
For group life insurance and group PMI the implications are clear: the more an insurer knows about an individual or group of individuals, the more accurately it can price risk. But it can also avoid risks it does not want to underwrite, creating complex issues for delivery of insurance services across the population.
Aware of these potential new complications, last November the FCA launched a call for inputs on how Big Data was affecting consumer outcomes and competition in the retail general insurance sector. It is also seeking views on whether its regulatory framework affects developments in Big Data or unduly constrains innovation in the interest of consumers.
The response to this consultation is imminent and, while it is formally restricted to the GI sector, it is likely to have implications for the policing of behaviour in the life sector too.
So while the potential exists for an increasingly selective approach from insurers, how much of a threat is it? Moneysupermarket.com chief data scientist Orlando Machado says: “The amount of data already available is huge and growing at a fast rate. The techniques for analysing it are evolving quickly. But the culture within insurance companies is not evolving at the same pace.”
Machado believes part of the reason for this is that insurers are uncomfortable about handing to machines too much of the decision making about the value judgements that influence underwriting. The fear is that scenarios could be created where the algorithms put insurers in breach of their obligation to treat customers fairly.
“Part of the concern is the concept of the black boxes being used,” says Machado. “It is possible to come up with more accurate pricing models but that involves sacrificing the interpretability of models.
“Insurance companies interpret regulation as meaning they have to be able to justify decisions. If they are using a black box, it is hard to do that. If all they have done is put in a black box and the numbers come out, they can’t justify that. So that leaves insurers to deal with models that err on the side of transparency.”
The gender underwriting directive is an area where lawmakers have decided that selection is not permitted. Few think this EU-made rule will be dropped any time soon, despite the Brexit vote.
Machado adds that insurers do not have to interpret all, and in some cases any, of the massive piles of data that new technologies generate in order to reap a benefit from it.
Take telematics for car insurance: the black box in the car effectively imposes a curfew on young drivers. “This does not require analytics,” says Machado. “It is a nudge towards different behaviour. You can check the data afterwards but you do not have to go to the lengths of analysing it.”
In the retail space, predictive data already helps shops to understand how to manage their inventory, reducing waste, maximising shelf space, allowing dynamic pricing, facilitating personalised marketing and enabling detailed customer analysis to be carried out.
It can also spot surprise correlations between the types of things people buy. So are there uses for insurers beyond filleting out the bad risks?
Association of British Insurers head of strategy Matt Cullen says: “There is a huge proliferation of data out there, much of which can be useful to insurers. The most emotive areas are in pricing and underwriting. But these are only some of the many ways the massive amount of data out there is helping insurance.”
Cullen highlights three ways in which data is helping insurance companies: technology solutions that help with insurance claims handling, understanding where there are going to be floods, and sensors in the home.
“The pricing side, through underwriting, will always be controversial because anything that changes pricing creates winners and losers. People often say the move towards data is destroying the principle of insurance because it reduces the extent to which you are sharing risk and pooling. I disagree. It has always been appropriate to price on the risk that a situation presents. All Big Data is doing is changing the level of granularity with which this is done.
“But we have to be sure that it works from a treating customers fairly perspective and it is important to be on top of any moral issues. The forthcoming paper from the FCA will look at this.”
So if you are somebody who buys a bottle of whisky for your grandfather every other day, could the data generated by your purchases one day make your life insurance more expensive?
“You always get people who slip through the net because the proxy being used is not appropriate to them. But that can be minimised by different approaches. It is a risk that needs to be managed,” says Cullen.
Some in the group life industry have questioned whether, in five years’ time, workplace cover could be contingent on the disclosure of data. This may sound like a Big Brother approach but younger workers are generally more accepting of the idea of sharing their information. A less confrontational approach by insurers could be to offer individuals more than a base level of cover if they agree to share their data. Consent will clearly be an issue in the spread of Big Data in the insurance sector.
“If you get Facebook data but use it only for marketing purposes, that is not so severe. When it comes to pricing, though, that’s a different question,” says Cullen.
So could we see people either required to hand over their motion data from their iPhone or other smartphone or be offered better rates if they agree to? While those who use wearable technology such as a Fitbit or Jawbone know they are actively monitoring themselves, many people are unaware that their smartphone is measuring how many steps they take every hour of the day.
“What we may see as a key trend is the shift from financial risk transfer to lifestyle consultant, giving people helpful advice on managing their actual risk coming with multiple touchpoints through the day,” says Cullen.
If companies believe they can create a better proposition by utilising customers’ Big Data, they will consider how to best access that. And one way will be by offering incentives.
“All of this is wrapped up in the ownership of data,” adds Cullen. “We could get to a position in the future where we have a state-backed data hub, held on the cloud, and individuals barter slices of their data to share with providers at their discretion.”