Predicting future insurance markets: Simulation of technology adoption paths with a Bayesian Bass model

Abstract

Technological progress can give rise to new loss exposures or reduce existing risks. In either case, it strongly affects insurance companies' existing lines of business. As the magnitude of such effects is often closely connected to the innovation's market penetration, reliably predicting its adoption process is vital in enabling insurers to adapt their business strategy accordingly. To tackle this challenge, we propose a novel combination of contributions from the domain of innovation diffusion. Our methodology generates adoption scenarios that holistically account for different sources of uncertainty, offers a structured way to incorporate expert judgment and insights from pre-existing technologies, and updates its predictions as new information on the technology's diffusion becomes available. Its modular nature ensures that components can be easily replaced by suitable alternatives. Using publicly accessible data on privately-owned battery electric vehicles in Germany, the application of our approach for predicting changes to insurance losses and for supporting insurers' sustainability strategies is visualized in a motor insurance case study.