Loss Predictive Power of Strong Motion Networks for Usage in Parametric Risk Transfer: Istanbul as a Case Study

Abstract

Parametric catastrophe (CAT) bonds are financial derivatives that provide a capital influx after a large natural disaster. More than US$10 billion in earthquake risk has been transacted since the late 1990s using mathematical relationships between monetary loss and ground accelerations recorded by seismological instruments. This paper investigates how the quality of these relationships varies depending on the functional forms used and the geographic distributions of both the strong motion sensor network and the exposure assets for which coverage is provided. We refer to this quality as the loss predictive power of the network. The analysis concludes that the recurrently used simplistic indices based on the nearest measurements to estimate loss at a particular location are advised only in the most optimistic circumstances in which the density of instruments is high where assets at risk are located. In other cases, more sophisticated ground motion estimation methods are recommended. The Istanbul area, where the Bosphorus I & II Ltd. CAT bond series was implemented in recent years, is treated as a case study.