Pricing the Earthquake Exposure Using Modeling

Catastrophe hazard modeling has become an important tool for ratemaking in lines of business subject to low frequency, high severity type losses. Natural hazard events such as hurricanes, tornadoes, and earthquakes rarely occur, but their devastation can be overwhelming when they do. Few insurance companies have enough historical loss data to sufficiently price for these events. In our paper, we plan to demonstrate a methodology which details the use of a model's output in determining a statewide rate level indication for the earthquake line of business, as well as a methodology for determining more equitable territorial relativities within a state. Our paper will outline a practical and understandable methodology for dealing with some complex issues involved in pricing the earthquake insurance exposure. The emphasis of the methodology within our paper will be on practicality and potential regulatory acceptance. Another feature of our paper will be the inclusion of a section dealing with the reflection of the net cost of reinsurance in the proposed direct rates. A final consideration is the treatment of a model's output when it is believed the modeled results are less than fully credible. The CAS ratemaking principles address data considerations used in making rates. Catastrophe hazard modeling output is an important component of "other relevant data" that is referred to in the principles.[1]. A company's history of earthquake premiums and losses does not have sufficient predictive power for establishing adequate rates. Out paper will rely on the power of catastrophe hazard simulation of multiple possible events and the associated loss costs generated from these models.