Fair risk assessment in the era of big data
Article information and share options
Technological transformation promises to make re/insurance services more tailor-made and pricing more usage-based. This is likely to change the re/insurance value chain in numerous ways1.
Increasingly, re/insurers will have more contact with consumers, be able to predict more accurately the types of risks they're facing and help customers mitigate those risks more effectively. It's a very exciting and dynamic time for our industry. And it's good news for consumers who can expect to benefit significantly from the innovative spirit that characterizes the industry today. As re/insurers harness new technologies for the benefit of consumers and to fulfill their role in society, access to and use of data will be of fundamental importance.
Re/insurance is an information-based business that relies on access to data and the ability to process it. Being able to use data has always been essential to a re/insurer's ability to fairly assess, price and manage risk. Throughout our industry's history, the availability of data has grown steadily. Mechanisms to process that information have become more sophisticated and more insurable as a result. Data availability and the capacity to process it are key components of a functioning re/insurance ecosystem. This ecosystem, in turn, enables risk taking activity by the real economy. Specifically, a functioning re/insurance market allows individuals, businesses and the public sector to concentrate on value-generating activities without having to worry about potential risks stemming from these activities. Big data2 and tech-based innovation can potentially strengthen the economic and social functions of re/insurance.
Big data isn't the enemy of insurance as we know it
An increasingly common misperception is that big data actually means an end to pooling of risk. But this is not the case. The sharing of losses is the basic premise of insurance. Some losses turn out to be less than expected, making it possible to pay for those that turn out to be greater than expected. If the expected losses differ from the actual losses – therefore affecting the price of re/insurance for individual risks – this does not change the fact that losses are shared. This is especially true for reinsurers, who offer their clients (primary insurers) access to global risk diversification. By improving primary insurers' diversification, reinsurers essentially lower the price of insurance for consumers in local markets.
Re/insurers have always charged premiums based on estimated expected losses. However, as they have only partially been able to access the data needed to make such estimates, some consumers have not been charged premiums in proportion to the risk they incur. Take motor insurance, for example, where traditionally, premiums are only loosely tied to the true risk factors like driving behavior. In response to increasing consumer demand for fairer motor insurer premiums, Swiss Re has partnered with UK insurer Collingwood and start-up Cuvva to offer a new type of insurance policy for car owners in the UK. Using the Cuvva mobile app, which tracks driving behavior, consumers pay a flat monthly fee to cover the basic protection and top up their cover by the hour when they drive. In this case, policyholders provide more data than in a standard policy and in return pay lower premiums when they drive less than the average policyholder.
Fairness in insurance – a question of values
The Cuvva mobile app is just one example of how technological innovation and the resulting availability of data allows insurers to charge more granular premiums reflecting the true expected costs of risk. While this leads to reduced premiums for some people, prices for other risks increase, sometimes even beyond the point of affordability. This raises legitimate questions about fairness in insurance – a discourse that is necessary and unavoidable. At its core, it's not a technical discussion. It's about values. The insurance industry can help to frame this discussion and, in doing so, we must remain fully transparent about our practices.
We can start by considering the factors an individual can influence – for example, behavior and personal choices, such as taking part in high-risk sporting activities. In most cases, using such information for risk pricing is perceived as fair. And doing so gives individuals a clear incentive to take more responsible decisions. The issue is more nuanced when it comes to factors beyond an individual's immediate control, where charging a more differentiated premium may not seem fair or justifiable from a societal perspective. There is a case for this however, as the re/insurance market is exposed to an information asymmetry between the re/insurer and the insured. That is, individuals have a higher incentive to seek insurance for higher risks. When insurers cannot charge premiums based on the underlying risks in a setting where the insurance is not compulsory, the resulting adverse selection increases the price of insurance for all. Innovation can add to the problem, as consumers gain more access to information, which allows them to assess their own risk more accurately and reliably than re/insurers. The resulting imbalances can only be addressed if insurers have access to this information as well.
Society should have a say
In theory, unrestricted use of data for risk pricing is a prerequisite for eliminating information asymmetries and the resulting adverse selection. However, for some people, this comes at the cost of higher premiums or even potential exclusion from insurance. Understandably, insurance regulators are currently closely monitoring how re/insurers use information. Some regulators may be considering putting in place restrictions on the way re/insurers use information for setting premiums. This would be problematic. When deciding whether, and to what extent to restrict re/insurers' use of data, regulators should also bear in mind the clear trade off involved here – that is, more restrictions on the use of data for setting premiums increase adverse selection and raise premiums for all. There will be clear winners and losers irrespective of the chosen solution. When regulators make such decisions, they in effect become technocratic decision makers, picking those winners and losers. They make value-based decisions that have a profound impact on society. Insurance regulators should avoid taking such decisions on behalf of society. Instead, those decisions should be made via the political process – society can and should have the say.
That being said, throughout this process, we acknowledge that the use of data requires a robust governance and risk management framework. Compliance with legal requirements, however, is just one part of the exercise and is often not enough on its own. Responsible use of personal data is a major priority for re/insurers, and data protection and privacy requirements are taken extremely seriously. It's important that companies explain how they use data and help consumers understand the benefits. For instance, in collaboration with Zurich University, Swiss Re is helping identify gaps in the legal system, by comparing legal requirements with public expectations. Our aim is to assess whether a code on the ethical use of data is needed. While a legal framework for data protection helps ensure that certain standards are met by all players in the market and that a level playing field is maintained, some companies may choose to do more to show their customers how they manage their data responsibly.
Collaboration is essential to mitigate risks effectively
There is no question that technological innovation and big data will lead to a prolonged period of rapid evolution of the re/insurance market. It is to be expected that during this time regulators will try to understand how re/insurers use data and examine the potential risks this poses to consumer protection and financial stability. We, in the insurance industry, must collaborate in this process and show how we use data transparently and responsibly. Without this collaboration, we risk a misguided regulatory response that cuts re/insurers off from crucial datasets – ultimately, this would be to the detriment of consumers and a resilient society as a whole.
- For more information on how technology is transforming the (re)insurance value chain, please refer to the Swiss Re Institute report on "Technology and insurance: themes and challenges"
- There are multiple benefits associated with the use of Big Data, AI and technology-linked innovation. Several papers have been written on this matter, for more information, please refer to the CRO Forum's paper on Big Data & Privacy: unlocking value for consumers