Abstract

Insurers traditionally use risk-specific characteristics of insureds to classify them according to risk. In this article, the practical relevance of information about multiple risks is demonstrated for the case of content insurance of a Swiss company. Two types of such information prove important: information about "spillover moral hazard" caused by mandated prevention affecting preventive effort in an unregulated line, and information about "common impulses" reflected in the loss experience of related lines. Both contribute to an improved prediction of loss probability.