The Use of Historic Loss Data for Insurance and Total Loss Modeling

Insurance catastrophe modeling of natural risks requires three key components of information: (1) a catalog of all possible events that can occur within the time period considered, (2) an exposure database consisting of the built parameters of insured structures or the population exposed to the hazard (usually in the form of a portfolio), and (3) the susceptibility of the structures or population at risk to the hazard. This chapter focuses on each of these components herewith highlighting recent trends in losses. To do this, CATDAT (Daniell et al., 2011) data are used for describing both economic as well as insured markets. Results indicate that normalized losses are in general decreasing most likely due to increasing hazard mitigation and decreasing vulnerability. Statistics as to the last 100 years (from 1900 to 2015) suggest that flood has been dominating historical losses. The statistical results have implications for future events. This article also considers how historic data are used in forming vulnerability functions along with how data might best be adjusted for changes in values, among others. Concepts such as replacement cost, hazard metrics, vulnerability functions, and spatial disaggregation of the average annual loss are then highlighted using an Australian earthquake model created by the authors.