From small to big data: (mis)uses of insurance premium for the government of hazards

If insurance is a technology for governing hazards, the tools for doing so have changed over time and seem to concentrate, in the second half of the twentieth century, on the notion of risk-based premiums. The narrative in the insurance literature is that by adjusting the premium to the risk, the insurer is sending a signal to the policyholder about his risk, therefore encouraging prevention. The purpose of this article is to contextualize this precept, by highlighting that the notion of premium as a risk signal was born within the narrow discursive framework of rational choice theory, thus building heavily on the perception of the insured as homo oeconomicus. More recent findings in behavioral economics, which question the rationality of agents, should have rendered this narrative obsolete. However, the still in the making behavioral insurance presents a combination of these theoretical insights, with the computation of individualized risk scores thanks to big data, supposedly embodying the ultimate risk-based premium. Two pilot products, taken as exemplars of behavioral insurance, exemplify this shift, and elucidate the contradiction: in practice premiums are now impacted by reward systems in large part disconnected from scores, thereby displacing the government of uncertainty.