Enacting Actuarial Fairness in Insurance: From Fair Discrimination to Behaviour-based Fairness

ABSTRACT

In line with developments in the personalisation of risk, the idea that insurance products should above all be ‘fair’ to the policyholders is increasingly voiced by commentators. The performativity thesis in Science and Technology Studies usually used to study economic markets can be used to investigate different enactments of ‘actuarial fairness’ in insurance practice. Actuarial fairness functions as a technical economic concept and was coined by the neoclassical micro-economist Kenneth Arrow (1921–2017). Faced with anti-discrimination legislation, the insurance industry has, since the 1980s, advanced the principle of actuarial fairness to legitimise their medico-actuarial technologies to discriminate between risk groups. In the absence of this actuarial fairness, it is assumed that dynamics of adverse selection—derived from neoclassical assumptions about economic actors— will result in the bankruptcy of insurance providers. The paradigmatic case of Fairzekering, a showcase of contemporary behaviour-based personalisation in car insurance, demonstrates an important shift in how actuarial fairness is enacted through behaviour-based calculative devices. Here, policyholders are enacted as being personally in control of their driving style while an interactive discount-infrastructure is set up to provide real-time feedback to incentivize policyholders towards ‘good behaviour.’ This enactment of behaviour-based fairness simultaneously implies a shift in the enactment of the economic actors involved, constitutive of the making of new economic ideas in behavioural economics.

KEYWORDS:

• Actuarial fairness
• insurance economics
• fair discrimination
• behaviour-based personalisation
• economic assumptions

Acknowledgements

This article has benefitted from the insightful comments of two anonymous reviewers and the editors of Science as Culture. Drafts of this article have been presented at the international workshops ‘Managing insecurities, creating uncertainties. An invitational workshop on the production and social effects of insurance rationality today’ (Sciences Po, Paris, 4–5 May 2017) and ‘Economisation of Lifestyle/Lifestylisation of Economy’ (KU Leuven, 9 May 2017). The authors thank the discussants and audiences of these workshops for their inspiring comments and suggestions, with a special mention of Tom Baker, Stephen Collier, Turo-Kimmo Lehtonen and Liz McFall. The authors would also like to thank Elisa Lievevrouw, Tijs Laenen, Sam Pless, Erik Schokkaert and Katrien Antonio for their productive comments on previous versions of the manuscript. Errors and omissions remain our responsibility alone. The research for this article is part of and supported by the Odysseus project ‘Postgenomic Solidarity. European Life Insurance in the era of Personalised Medicine’, funded by the Research Foundation Flanders (FWO) under grant number 3H140131.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Gert Meyers is PhD researcher at the Life Sciences and Society Lab (Centre for Sociological Research, KU Leuven). His research focuses on the adoption of behaviour-based personalisation in insurance. Gert has a broad interest in the sociology of economics, the sociology of markets, the politics of responsibility, approached from a pragmatist STS-perspective.

Ine Van Hoyweghen is Professor at the Centre for Sociological Research (CeSO) at KU Leuven. Her research interests are in sociology of biomedicine, Science and Technology Studies (STS) and sociology of insurance. She has been working on the politics, regulation and governance of genomic technologies, the relationship between health care economization and predictive medicine, and forms of accountability in the governance of emergent biomedical technologies.

Notes

1. To study this grown likelihood of using medical services, once people are insured and to estimate the price elasticity of demand for health care, two large randomised control trials have been set up in the United States of America, the RAND Health Insurance Experiment (Aron-Dine et al., 2013) conducted in the 1970s and the Oregon Health Insurance Experiment (Finkelstein, 2015).

2. The paper was part of a theme issue on health economics in the American Economic Review. The editor invited Kenneth Arrow, who had not previously written on health economics, as a ‘theorist’ in economics to write his reflections on the topic of health (Arrow, 2015). Arrow was, however, well informed on insurance practice as he followed courses to become an actuary while graduating:

  Much of the material to be learned was extremely detailed algebra, but I did encounter some interesting concepts such as moral hazard and adverse selection. They came to my mind again in the study of medical economics, and I began to see a more general pattern applicable elsewhere in economic behaviour. (1984, pp. 77–78)

3. Some handbooks do not even mention the name of the concept while explaining it (see, for instance, Mankiw and Taylor, 2014, pp. 542–544).

4. In a similar way, self-selection is being employed by insurers to attract women as a mechanism to avoid risk classification based on gender, which has been prohibited in the European Union since 2012 (Rebert and Van Hoyweghen, 2015). Insurance policies like Drive like a Girl in the UK and Zij Rijdt beter in the Netherlands, the latter using the same technology of Chipin, market towards women by adopting a ‘feminine’ lay-out. Men are also welcome in the policy: ‘To “drive like a girl” refers to a safer, smoother driving style. This is why we want everyone to be proud to “drive like a girl”’ (Drive Like A Girl, 2017).

5. A trope popular among insurers on the opportunities of telematics is that most of the drivers consider themselves as ‘better than average’ (e.g. Karp, 2012).

6. Although drivers are addressed as individuals who each pose a personalised risk by the driving behaviour they control, the results are clustered again and visualised in three broad scoring ranges, which makes one cannot speak of a completely ‘personalised’ differentiation. Therefore, it would not be entirely accurate to oppose fair discrimination, which takes place between risk groups, to behaviour-based personalisation, specifically as a distinction between group-based and individual-based fairness.

Additional information

Funding

This work was supported by Fonds Wetenschappelijk Onderzoek (Research Foundation Flanders) [3H140131].