New tech & big data: are they good for insurance?
Richard Keating investigates how big data and new technologies are changing the face of the insurance industry, and examines the resulting threats and opportunities
New technology and data sets are profoundly affecting the insurance industry, by allowing insurers to better assess the risks they are exposed to and interact with their consumers in new ways. They are also drawing new entrants into the market, with ‘Insurtech’ startups entering the field, as well as established technology companies.
What are the most promising of these new technologies, and how will they change the industry? Will established insurers be able to retain their dominance of the market? And how will consumers and insurers react to the potential new market dynamics?
The selfie generation
One of the most striking new technologies to emerge is being developed by Lapetus, a US-based startup whose CHRONOS underwriting process involves the potential customer uploading a selfie of themselves, along with some basic personal details including age, sex and physical activity level. From the selfie, Lapetus uses machine learning techniques to validate self-reported body mass index, gender, evidence of smoking and an estimated age.
It then combines this information with that gleaned from the selfie, in order to produce a term assurance quote based on the desired level of coverage. The whole process is marketed as “prospect to policyholder in under 10 minutes” and claims to dramatically cut underwriting time.
Black boxes and fitness trackers
One of the most widely known current uses of technology within the insurance industry is when telematics boxes are sold as part of motor insurance (widely known as ‘black boxes’). Ptolemus Consulting Group predicts that, by 2020, nearly 100 million vehicles globally will be insured by such policies. The savings available to consumers can be large. Progressive, a US insurer, offers savings of up to 30% from the generic base premium according to driving style and usage.
These black-box-based policies are becoming increasingly popular in the UK, with data from the British Insurance Brokers’ Association showing 750,000 live policies at the start of 2016, an increase of nearly 25% on the 2015 figure.
Another area where technology is being increasingly used for insurance purposes is in the healthcare sector. Vitality, a UK-based health and life insurer, offers ‘Vitality points’ for consumers who are willing to track and share their daily activities, including walking, running, cycling, swimming or going to the gym. This can be done in a variety of ways, including fitness trackers and health check-ups. The points can be then exchanged for a variety of rewards, including cinema tickets and discounts on fitness products.
Poisoning the pool?
One of the key issues with the use of technology to assess and price risk is whether it will undermine the risk pooling that has been the basis of all modern insurance. Leveraging data and new analytic techniques may bring significantly more accurate risk models, and a much higher degree of personalisation to insurance quotes, but what of the potential problems this could entail?
One issue is that those deemed to be high risk could be effectively unable to get insurance at an affordable price. Whereas, previously, these people would have been cross-subsidised by lower risk individuals within the same risk pool, that may no longer be the case in a future of more tailored insurance quotes. An example of this may be someone identified as having a genetic disorder, where no preventative measures could have been taken, being given a large life insurance quote to reflect the uncertainty around the treatment and risk factors. This could also have particularly adverse outcomes for those consumers who are legally required to buy insurance, such as for cars in the UK.
On the other hand, if people have access to their own data, they may be able to accurately predict if they are low risk and come to the conclusion that they no longer need coverage at all. This could reduce profitability for insurers and further undermine the risk pool, pushing up premiums for the higher- risk consumers that remained.
Another concern is that new data will allow insurers to discriminate on price in other ways that could be deemed unfair. As Cathy O’Neil details in her 2016 book Weapons of Math Destruction, insurers may offer better prices to those consumers who have been assessed as more likely to shop around (and who generally tend to be richer and better informed). O’Neil writes that the result could be that “poor drivers who can least afford outrageous premiums are squeezed for every penny they have”.
A further issue is how customers will react to these new technologies, and how privacy concerns and data security will be dealt with. Currently, telematic sensors are generally marketed as ‘savings for safer drivers’, which is usually received well by consumers. The data collected from these sensors is typically used in a way such that premiums can only be reduced from a base level, but how will consumers react if this is not always the case? People may feel differently when presented with a black box that might raise their premiums. In the future, those consumers who choose to not have black boxes installed may have to pay much higher premiums as a result, potentially putting a real cost on privacy in this context.
A lot depends on how regulation reacts to these new advances. For example, there could be a requirement to provide some level of coverage to all individuals, despite the risks. One particularly controversial area is around the use of genetic data, which may be used by consumers and insurers as it becomes cheaper and more readily available. People with the ‘wrong genes’ might then find themselves subject to higher premiums, owing to factors over which they have no control and are unable to ever change.
Will companies be required to disclose the source code of their algorithms to ensure that the pricing they are offering complies with non-discrimination legislation? Consumers may feel they have a right to see how they are deemed to be a high risk. On the other hand, insurers might argue that this information is proprietary, and that releasing it could undermine their competitive position. In the future, regulators could have a role as mediators in such disputes.
Reshaping the future
The new data available to insurers, and algorithms being designed to utilise it, have the potential to totally reshape the industry. More personalised products might allow insurers to better assess the risks they are exposed to and design their products accordingly. Individuals may find products more tailored to their individual circumstances, thereby perhaps paying a premium more aligned to their assessed risk profile.
Data will become key to insurers, and those who are most able to use it may reap outsized rewards within the sector. This expertise in data might bring in new entrants to the markets, such as technology companies, and require incumbents to increase their capabilities in this space.
How consumers will react to these changes is still unclear. Younger people tend to be more comfortable sharing personal data, so may be willing to do so in exchange for better pricing. New regulation might be needed to protect vulnerable consumers and to decide which data can be used for pricing purposes.
Richard Keating is an actuary in PwC’s pensions team