Catastrophe Models Are More Accessible, Insightful and Prevalent Than Ever
Five years ago, catastrophe(CAT) modelling was relatively unknown. Today, CAT modelling for hurricanes and earthquakes is fast becoming the norm in property underwriting, for catastrophes that can obliterate otherwise stable businesses. Commercially viable CAT models started emerging only in the last quarter century. Earlier, rudimentary methods were employed to estimate catastrophic losses as historical loss data was scarce for low frequency, high severity events and standard actuarial techniques inadequate.
CAT modelling is the practice of using computing horsepower to mathematically represent physical characteristics of catastrophes. Dominant CAT models in use are AIR Worldwide(AIR), Risk Management Solutions(RMS), and EQECAT. These modelers develop probabilistic models that help organizations prepare for financial impacts of catastrophes. (Re)insurers, rating agencies, risk managers and brokers license models from these firms.
It was the unprecedented loss sizes experienced during Hurricane Andrew in 1992 that exposed deficiencies in the erstwhile actuarial approach to quantify cat losses. When the hurricane hit, AIR promptly issued a fax to its clients estimating model losses in excess of $13 billion. Months later, the Property Claims service reported an actual industry loss of $15.5 bn. Losses hit the market hard, resulting in insolvency of 11 insurers. Subsequently, adoption of catastrophe models grew briskly, turning into a more sophisticated and reliable basis to catastrophe risk assessment.
CAT models are designed to pinpoint locations where future events are likely to occur, intensity likelihood, estimated damage ranges and insured losses by future events. Factors that obviate use of traditional methods include: constantly morphing exposure landscapes, new properties in high hazard areas and changes in building materials and designs. Models combine historical disaster information with current demographic, building (age, type and usage), scientific and financial data to determine potential cost of catastrophes for specified geographic areas.
The process of developing CAT models is complex, drawing on expertise from a broad range of disciplines, including meteorologists, seismologists, geologists, engineers, mathematicians and actuaries. CAT models provide a wide range of outputs e.g. exceedance probability curves, real time loss estimates and loss tables.
Insurers use CAT modelling for underwriting and pricing. Models assess risk in an exposure portfolio, guiding underwriting strategy and reinsurance decisions. It helps reinsurers and brokers to price and structure contracts, while bond investors use it in pricing and structuring of catastrophe bonds. Some regulators allow insurers to use CAT modelling in rate filings for pricing.
The unprecedented severity of storms during the 2004-05 hurricane seasons led to CAT modelers facing criticism for underestimating losses. However, it is important to recognize that there is no one-size fits all approach and different approaches exist, each using different assumptions, data inputs and computational algorithms.
As in most of insurance, new technologies are making a dent in CAT modeling. Xceedance, a global provider of insurance consulting and services, offers On-Demand Catastrophe Modelling Services, using the open Oasis Loss Modelling Framework. It allows global and regional catastrophe modelling companies to implement models on the Oasis platform, while delivering modelling services on-demand to the insurance industry with no annual licensing requirements and the flexibility to choose peril models from its community of expert model providers.
The CAT modeling industry is also steadily moving towards greater use of AI, a step change from its focus on traditional statistical techniques. In data-assisted approaches, physical models simulate the underlying processes. Such usages are emerging in organizations such as Cytora and Reask. A number of partnerships have evolved tying carriers with insurtechs, example being global reinsurer Scor with insurdata and KatRisk.
Despite the widespread use of CAT models, as with financial models, it is not an exact science. But as the probability of extreme weather-induced catastrophes becomes acute, CAT models will grow as a vital component of risk management toolboxes for (re)insurers.
You get 3 free articles on Daily Fintech. After that you will need to become a member for just US$143 a year (= $0.39 per day) and get all our fresh content and our archives and participate in our forum.