Earthquake risk and insurance

Abstract

Our concern is with protection against natural catastrophes through financial insurance. Development of appropriate premium rates requires consideration of a wide variety of variables entering into the occurrence of the disaster, and a broad range of scientific and statistical investigations results. Topics discussed include: premium computation, distribution of large earthquakes in time, ground motion at sites, attenuation of energy with distance, damage description and actual practice in various countries. Statistical considerations that arise include: description, stochastic modelling, conditioning, spatial processes, (marked) point processes, uncertainty estimation and robust/resistant procedures. Study of the insurance problem is scientifically enlightening because it requires one to focus on the whole context of the problem; geology, seismology, earthquake engineering, and damage.

Modelling peak accelerations from earthquakes

“…It is therefore essential to appropriately model earthquake response for risk management and insurance applications. A general discussion on seismic risk and insurance, and some of their statistical considerations, is found in Reference [1].…”

Automatic methods for generating seismic intensity maps

“…In Brillinger [2] it is shown how relations obtained by analyzing M M intensities may be used in computing premiums for earthquake insurance. In particular using assumed loss ratio values for buildings of some type of interest, one may estimate the expected loss for such a building situated a given distance from an earthquake source.…”

A bivariate compound Poisson model for the occurrence of foreshock and aftershock sequences in Turkey

“…Earthquake occurrence can very well be described by stochastic point process models. Stochastic point process models for earthquakes have previously been discussed by Brillinger (1993), Vere‐Jones (1995), Ogata (1988, 1998) and Zhuang et al . (2002, 2005).…”

“…However, it has always been of interest to try to enrich time models of earthquakes with covariate information, i.e., magnitude, intensity, and the numbers of foreshocks and aftershocks.. Earthquake occurrence can very well be described by stochastic point process models. Stochastic point process models for earthquakes have previously been discussed by Brillinger (1993), Vere-Jones (1995), Ogata (1988Ogata ( , 1998 and Zhuang et al (2002Zhuang et al ( , 2005. In particular, the Epidemic-Type-Aftershock-Sequence model of Ogata (1988) has proven to be useful in the description and modeling of earthquake occurrence times and locations.…”