“We could protect you against biased algorithms”

The predictability boom is about to change the business models of the insurance industry. No need to panic, says Amelia Lorenzo, Head of Market Dynamics & Opportunities at Swiss Re – as long as there are risks in society, there’ll be need for cover.
27 May, 2019 by
«Wie wär’s mit einer Algorithmen-Versicherung?»
GDI Gottlieb Duttweiler Institute

GDI: Ms. Lorenzo, the better predictive methods become, the better the individual risks for special situations or activities can be calculated. This could be a huge chance for insurers. They could potentially insure everything everywhere for everybody – a big deal, isn’t it?
Amelia Lorenzo, Swiss Re: Conceptually it is a big deal. In dollar terms, however, it might not look that great. More/better models does not necessarily mean more insurance business. Individuals are not ready to pay for hedging every risk they face. There is a lot of discussion about new individualized insurance solutions but there are only few sizeable success stories. Take the example of a mobility micro-insurance policy that covers just a single car ride. The pay-as-you-go insurance model has been around for over 20 years so such a cover could theoretically be a blockbuster but, in practice, it is a niche solution offered by few insurers.

Why? The ride-sharing market has been growing rapidly and has a very good outlook.  
No doubt about it but it offers almost no data for insurers to assess the risk of single rides. Ride-sharing companies emerged only few years ago and made public the only piece of data that was relevant to their success: the price per ride. How can insurers assess the riskiness of a single ride without information about the driver and the car? Insurers need a minimum set of historical and contextual data to price risks. It is fair to say that there is still not enough reliable and comparable data on ride-sharing services to make single-ride insurance a mainstream cover.

Getting there is just a matter of waiting few years, right?
In terms of data availability: yes. In terms of business volume: not necessarily. From the customer point of view, insurance for single car rides is an appealing proposition. Now put yourself in the shoes of insurers. They would need to sell millions of these single low premium covers to get an attractive profit in absolute terms. Today established insurers and retail brokers have little incentives to embrace and advance the development of short-duration individualized insurance. They are more likely to leverage new data sets and better predictive models to write new corporate business ('large' commercial deals) than to boost short-term individualized covers (micro transactions with low profit potential). Insurance start-ups will of course tap into these spaces left unattended by established players.

And if the large deals cease to exist?
That is a big IF. In the future, large commercial deals will change but they are unlikely to disappear. Technological developments are changing the risk landscape in all business fields and often shifting the risk ownership from the user to the manufacturer. Take the example of motor insurance. Today individuals are liable for car accidents. In the future, it is possible that driving software companies or manufacturers of autonomous cars will be liable for car accidents. All companies have limited resources and need to make choices. For established insurers it makes more financial sense to design liability solutions for software companies and car manufacturers than to design single ride car insurance solutions. One deal with a large corporate can yield more absolute profit than ten million short-duration pay-as-you-go policies.

Even if you do not earn much with individualized new insurance covers, they could still be valuable door openers. They could be the trigger for a customer to download your app and offer you the chance to sell other products.
They could indeed prompt customers to download an app but cross selling in insurance is not as powerful as in the consumer goods space. Reinsurers and insurers are evolving from pushing products to protect against major scary events (e.g. earthquakes, floods) to selling solutions for what keeps customers awake at night. They progressed a lot in the last decade but there is still a mismatch between available insurance products and the risks that individuals and corporates want to hedge. But I am confident that we will get there soon. The necessary mindset change is happening.

Insurance is not only a private business but also part of the social system. There is a link between insurance products and the social fabric of society. If we get into a situation, where more risk is retained because I can calculate my own risks better than before, this can damage the solidarity within society.
I disagree. There will always be risks that connect people. However, the nature of these risks will change over time. Today you can argue that you prefer to spend money managing your own health than contributing to a solidarity scheme that insures people who have unhealthy nutritional and lifestyle habits. However, you might be willing to pay insurance for your business or your community to become more cyber-resilient. There will be new/more risks whose consequences will be hard to predict. People will be keen to buy into an insurance scheme to hedge against those risks.

Can you name some of these “new/more risks”?
One example is the proliferation of smart cities, which increases cyber risk exposure. Increasing longevity, massive migration and increasing food scarcity are additional examples. If governments, insurers and reinsurers sit at the same table with a can-do mentality, they might design solutions - built on the premise of solidarity - to manage these type of risks effectively.

So the survival of the solidarity principle does not depend on how much old insurance schemes go out of business, but on the existence of hard-to-predict risks affecting society?
Yes. Increasing public information about risks and better predictive models will prompt people to retain some of today's insured risks. However, there will always be hardly predictable emerging risks for which risk pooling will be a good and affordable option. Ten years down the road, individuals may retain some property risks but might want to protect themselves against biased algorithms, theft of digital identity, etc. I do not think that the solidarity principle is in danger. I do believe, however, that insurers will need to expand the boundaries of insurable risks to grow its business and underscore the central role of re/insurance in economic and societal progress.

This would mean a boom in product development.
Yes, a boom in product development, distribution approaches and public-private partnerships.  

Will incumbent insurance companies generate this product development boom or will start-ups fuel it?
Both incumbents and start-ups will make it happen. Plenty of innovative personalized covers will come from start-ups. They are agile and can fully leverage technological developments to design appealing customer journeys and inexpensive solutions. However, insurance statutory capital regulation tends to limit the geographical impact of start-ups to one country. Incumbents are trying hard – I am managing a team dedicated to it! – to leverage comprehensive risk knowledge, global reach and Big Data to tap into new risk pools and to service existing risk pools in a more efficient and customer centric way. I believe that the re/insurance industry will become much more dynamic and innovative over the next 5 years.

Do you want to know more about how AI will benefit your business? Visit the conference "The Power of Predictions" on 4 June 2019.

Share this post