The aim of this article is to assess the impact of Big Data technologies for insurance ratemaking, with a special focus on motor products.The first part shows how statistics and insurance mechanisms adopted the same aggregate viewpoint. It made visible regularities that were invisible at the individual level, further supporting the classificatory approach of insurance and the assumption that all members of a class are identical risks. The second part focuses on the reversal of perspective currently occurring in data analysis (...) with predictive analytics, and how this conceptually contradicts the collective basis of insurance. The tremendous volume of data and the personalization promise through accurate individual prediction indeed deeply shakes the homogeneity hypothesis behind pooling. The third part attempts to assess the extent of this shift in motor insurance. Onboard devices that collect continuous driving behavioural data could import this new paradigm into these products. An examination of the current state of research on models with telematics data shows however that the epistemological leap, for now, has not happened. (shrink)

The recent proliferation of new data and technologies enables increasingly finer personalization of products and prices in every domain. In insurance, this revives and enlarges old debates around fairness that have never been completely settled. We will argue that the commonly accepted “actuarial fairness” as based on the “individual cost of risk” derives in fact from a conflation: while it indicates the average cost for a group of insureds from the perspective of an insurance company—and is therefore sound from a (...) business profitability viewpoint—it is arguable whether it represents the “fair price” for the individual insured. We first show in a historical perspective the intertwinement of conceptions of fairness with knowledge, in order to point to the alternative to actuarial fairness for insurance. We then describe the intrinsic difference between the insured and the insurer when underwriting an insurance contract. Finally, we build on this distinction to discuss the meaning of fairness in insurance prices. We thus hope to re-center the debate around insurance fairness on its underlying solidarity mechanisms rather than technical and actuarial considerations. (shrink)

Despite their common core in statistics, insurance and epidemiology propel two different forms of solidarity. In insurance, the collective is a source of protection, thanks to the pooling of risks; in epidemics by contrast, the group remains the source of danger for the individual. The aim of this paper is to highlight the conceptions of community and solidarity at play in epidemics in contradistinction to insurance, with a focus on the shift introduced by big data and algorithms. Paradoxically, while the (...) new technologies and epidemiology share a common view on the relation between the individual and the collective, tracing apps were not widely adopted in the Covid-19 crisis. This reluctance to use current technologies for the sake of epidemic containments highlights, beyond legitimate interrogations, a confusion between two imaginaries of the social: insurance solidarity where the interdependence is a source of rights, and epidemic solidarity that imposes duties. (shrink)

While it is often claimed that the emerging digital governmentality functions as a new apparatus of surveillance, the aim of this paper is to characterise this regime in relation to Foucault’s disc...

With their intensive use of data to classify and price risk, insurers have often been confronted with data-related issues of fairness and discrimination. This paper provides a comparative review of discrimination issues raised by traditional statistics versus machine learning in the context of insurance. We first examine historical contestations of insurance classification, showing that it was organized along three types of bias: pure stereotypes, non-causal correlations, or causal effects that a society chooses to protect against, are thus the main sources (...) of dispute. The lens of this typology then allows us to look anew at the potential biases in insurance pricing implied by big data and machine learning, showing that despite utopic claims, social stereotypes continue to plague data, thus threaten to unconsciously reproduce these discriminations in insurance. To counter these effects, algorithmic fairness attempts to define mathematical indicators of non-bias. We argue that this may prove insufficient, since as it assumes the existence of specific protected groups, which could only be made visible through public debate and contestation. These are less likely if the right to explanation is realized through personalized algorithms, which could reinforce the individualized perception of the social that blocks rather than encourages collective mobilization. (shrink)