Alberto Cevolini (University of Bologna) and Elena Esposito (University of Bielefeld)
Abstract: A wave of digitalization is sweeping through the insurance industry and promises to change the whole value chain of insurance business. Investments in InsurTech also suggest that this digitalization is inevitable. In our contribution, we want to explore some social consequences of this digital automation of insurance decision-making processes.
In the first part of our contribution, we address one of the main novelties of this digitalization: the use of predictive algorithms to adapt the policy premium to policyholders behaviour. This use is the basis of usage-based-insurance (UBI). What changes when the future is observed predictively moving from individual behaviour, rather than probabilistically moving from structural variables? What is the difference between the statistics underlying classical actuarial calculations and algorithmic techniques that dig into behavioural data?
In the second part, we investigate behavioural rates in a specific insurance case, that of third party liability motor insurance where these rates are already widely applied and where the experimentation of algorithmic technologies is more advanced. We deal with two main questions: to what extent can the increasing segmentation that behavioural pricing allows jeopardize the risk pooling and spreading mechanism on which the insurance industry is based? How does behavioural pricing impact the function of insurance? We will answer these questions starting from the first results of an empirical research we are carrying out on Italian telematics-based motor insurance in the context of the ERC PREDICT research project.
* zoom link available upon request