Analysis of Extremes for Hurricane Wind Speeds and Residential Losses

People who live in the state of Florida recognize the impact hurricanes have on their lives. With
the hurricane season taking up six months of the year, Floridians feel that their lives are divided into two
seasons; “the hurricane season” and the “calm season.” Investigating hurricane risk is critical for the state of
Florida; both from a socio-anthropological viewpoint and from an actuarial viewpoint. Using theory from
record-breaking data, here we consider the prediction the intensity of the next “big” storm and compute the
probabilities of the waiting time for the future record. In addition to the prediction of future records, we also
consider the estimation of the extreme quantiles of the distribution of the intensities. From the actuarial
perspective, exposure is increasing at a breathtaking pace in Florida, so insurance companies are extremely
interested in protecting themselves against the “once in a hundred-year event” or probable maximum loss
(“PML”.) One model that estimates insured hurricane loss is the Florida Public Hurricane Loss Model. We
use extreme value theory to investigate the distribution of extreme quantiles of losses from the Florida Public
Hurricane Loss Model.