Seismic risk management of insurance losses using extreme value theory and copula - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 777

risk managers of insurance and reinsurance portfolios. Furthermore, it is
desirable to be equipped with techniques for portfolio aggregation, where
the consideration of adequate dependence among different parts of a port-
folio is of high importance. To deal with such problems, a statistical approach
based on the marginal generalised Pareto models together with a versatile
copula-based dependence structure was investigated. Using a numerical
example, effectiveness of the proposed method was demonstrated in jointly
approximating very large seismic losses from two different building portfo-
lios. In particular, the proposed approach is useful for assessing risk expo-
sure of a reinsurance policy with a high retention level. The fast computation
with reasonable accuracy is an attractive feature of the method.
In the future, applications of multivariate modelling based on the copula
technique will be more popular, because of its capability/fl exibility by
dealing with marginal models and dependence model separately. Such
applications should not be limited to earthquake engineering and should
be transferred to any fi elds/problems that involve with multivariate random
data. In line with the present example (i.e. modelling of seismic loss data
from different insurance portfolios), extension of the current approach to
more than two portfolios is an important step. Another technical challenge
is the applications to numerous buildings and infrastructure facilities/com-
ponents (e.g. more than 50 000 structures simultaneously) - this is not easy
because the simulation of seismic intensities at a large number of locations
requires a huge covariance matrix and takes signifi cant time and computer
memory. Some approximation framework/procedure needs to be adopted
(e.g. hierarchical structure of buildings and system components and ade-
quate aggregation method). Lastly, a signifi cant challenge in seismic loss
estimation methodology is the accurate assessment of indirect seismic
damage costs, such as demand surge effects on building repairs/reconstruc-
tion, prolonged and inter-related business interruption within a supply
chain, and stagnant local economy, slowing down recovery from disasters.
These will affect the probabilistic characteristics of aggregate seismic loss
for rare cases - in such cases, capability of capturing accurate joint right tail
behaviour would be even more important.
28.6 References
Adams J, Halchuk S (2003), Fourth generation seismic hazard maps of Canada:
values for over 650 Canadian localities for the 2005 National Building Code of
Canada (Open File 4459), Ottawa, Geological Survey of Canada.
Atkinson G M, Goda K (2011), 'Effects of seismicity models and new ground motion
prediction equations on seismic hazard assessment for four Canadian cities', Bull
Seism Soc Am , 101 , 176-189.
Baker J W (2011), 'The conditional mean spectrom: a tool for ground motion selec-
tion', J Struct Eng , 137 , 322-331.