... because some insurance pricing is still incredibly crude. Money Box, the BBC Radio 4 consumer finance show, this week inadvertently provided a very powerful and practical example of the big data opportunity in insurance. The show wasn’t actually about big data. It was about how a divorce can increase the cost of car insurance.. But it highlighted a number of questionable data metrics used by some insurers, and consequently how crude insurance pricing can be.
Dubious data metrics used by some insurers
The show presented reasonably compelling evidence that car insurance premiums jump significantly for divorcees. Their estimate was in the region of 20-30%.
Industry representatives who were interviewed attempted to provide some rationale for the price increases:
It was one of the listeners who contacted the show who flagged the problem with these explanations most succinctly. “Divorced people on average probably do more miles, extra journeys visiting children and so on”. In insurance jargon, that listener had identified the problem of insurers relying heavily on ‘proxy’ measures of risk, not ‘direct’ measures of risk.
Marital status is not a direct measure of car insurance risk. It is a proxy measure. Insurers are assuming that because a person is divorced (the proxy measure) some actual risk factors change e.g. they drive higher miles, drive faster or drive late at night. Insurers do not know this. They assume it from marital status.
Similarly, a credit score is not a direct measure of car insurance risk. It is a proxy measure. Insurers do not know if a person is maintaining their vehicle or not (a direct measure of risk). But they appear to be assuming this from a credit score downgrade. The Money Box host, Paul Lewis, was perfectly justified in his incredulity. “So insurers use your credit score even though they are not lending you any money?”
Technology advances will produce more relevant data
This proxy versus direct measure of risk is extremely relevant today as more and more direct measures of risk are becoming available:
Winning insurers will refine their pricing by replacing proxy measures with direct measures. Actual miles driven and driving behavior will influence premiums, not marital status. The actual maintained condition of a car will influence premiums, not credit score. These insurers will be able to target good risks more aggressively and avoid bad risks.
Economic battles will be fought over data
But capturing this opportunity is not solely in the hands of insurers. There will be battles fought over how to share the economic benefits of better data and more accurate insurance pricing.
Vehicle manufacturers will own the data gathered from the sensors in their cars. They will want compensation for handing over this valuable data. Some technology companies specialising in advanced analytics such as artificial intelligence will be better than insurers at processing and interpreting large amounts of data. They will want compensation based on the value they provide to insurers.
Some initial skirmishes have already taken place. In 2016, the insurance company, Admiral, was about to launch a scheme that used social media data from Facebook, with an individual’s consent, to refine it’s car insurance pricing. Facebook blocked the initiative citing privacy concerns. But was privacy really at the core of this spat?
Facebook uses it's data to generate revenue. It sells targeted advertising based on this data. Cynics may speculate that the Admiral issue had more to do with economics than privacy. Why would Facebook allow an insurance company to create value from Facebook data without recompense? This situation could be indicative of the battles to come over the sharing of the economic benefits from big data.
Using big data to refine pricing in insurance is a huge opportunity but not necessarily easy to capture. One of the Money Box listeners set the bar high for data producers, data analysts and insurers. “You would have thought car insurance premiums go down (following a divorce), no arguments at the wheel equals one less distraction.” How will companies measure and price that metric?
To enquire about how Canbry can assist your organisation to research, analyse and make better strategic decisions in insurance, insurtech or fintech, contact us here.
The views expressed in this article are those of the author at the date of publication. The contents of this article are not intended as investment advice and will not be updated after publication unless otherwise stated.