The dangers of data: How the movement from pooled risk could create a class of uninsurablesRSS icon

Written by Mark Andrews on Thursday 25 January 2018

Data is the “new oil”; the world’s most valuable resource. Google has graphically charted that value in its meteoric rise from technology start-up to one of the largest companies in the world in just 20 years. Insurance professionals have historically been well versed in interpreting large data sets and making the most of this exponentially growing commodity; 90% of the world’s data was created in the last two years. In the insurance industry this has led to a sophisticated ability to pool data and offer a smoothed premium across a large group of risks, ranging from objects and hazards, to people and lives.

Over the last decade, however, we’ve seen the industry move away from historical risk events and towards predicting and pricing on customer behaviour. Instead of relying on pooled risk, we are seeing a shift towards individual customer pricing and the insurance of individual events (micro insurance). This means providing a far more personalised insurance policy for an increasingly focused customer segment. The numerous telematics policies that share driving data, and have the potential to save (safe) young drivers huge premiums, are a good example of this, but this still represents only a small part of the motor insurance market.

The insurance industry cannot wait to get its hands on even more data with which to segment risks and more accurately underwrite them, in a move towards personalised pricing and even personalised products. Insurers are already experimenting with this. We are witnessing the use of new data sources, such as last year, when Admiral proposed analysing the language from customers’ Facebook posts to better understand how safely, or dangerously, a person would drive. Facebook was met with a backlash and the experiment didn’t go ahead, but this is an indication of the direction data usage is heading.

Insurers, and more recently InsurTechs, are using this increased access to data to dismantle packaged products and move away from annual policies. On-demand and episodic insurance policies, along with item-specific and componentised insurance products could become the new norm, rather than a traditional annual package product. Eventually, we could even see real-time insurance, with pricing that fluctuates with the circumstances. Could we eventually see “invisible” insurance, where the relationship between insurer and customer is such that we trust the insurer to just charge us for the appropriate level of insurance?

The potential benefits have been widely touted, such as systems that accurately price risks without a ten page questionnaire; reduced premiums for low risk customers or healthy and active people who use fitness trackers; and drivers with telematics policies. But there are risks too with the possibility of some customers being priced out of insurance, leaving many people either underinsured or not insured at all. For example, those with lifelong health issues will struggle to get access to health insurance, and those on flood plains will struggle to get home and contents insurance.

The insurance industry has already responded positively in one case with the formation of Flood Re, funded by the introduction of a levy on insurers who sell home insurance products. But the emergence of all this new data does leave us with a dilemma for insurance – how far do we go with risk selection accuracy before we break the insurance model? Improving the accuracy of pricing to an individual level breaks the traditional insurance model of pooling and sharing different risks. The converse argument is – if I invest in a vehicle with automatic braking, lane control and driver alert features that make me a safer driver, I don’t want to be in the same risk pool as a 20-year-old banger!

Improved risk selection threatens to increase under-insurance (and non-insurance) and also non-disclosure, but the intelligent use of data enrichment sources improves customer acquisition – L&G’s new SmartQuote product won many awards in 2017 for its 5-question quote-and-buy journey.

Shareholders and investors insist on a decent return of equity so we do need to reduce fraud but not at the expense of removing higher risks completely. Insurance companies don’t make enormous profits but public perception is that they are ripping off customers if they do. Compare average profit for an insurance company to a tech company. How many Mutuals and Friendly Societies have disappeared in the last 30 years? Is shareholder profit now more important than insurance coverage for all?

The re-emergence of peer-to-peer insurance models over the last couple of years from the likes of Laka, Friendsurance and so-sure, has been pleasing to see. But the demise of Guevara, one of the early pioneers of P2P insurance, shows that perhaps shareholder return comes before customer protection. Is insurance getting too clever for its own good…?

Enjoy this article?

Why not share it...