Catastrophe Modeling for Corporate Risk Managers
O'Brien, Tom, Risk Management
Catastrophes, both natural and man-made, are unique in their ability to severely and immediately impact a company's bottom line. Time and again, history has shown how a single hurricane, earthquake, tornado or act of terrorism can destroy property, interrupt business, injure workers and otherwise throw a successful company into turmoil. To manage the infrequent but potentially devastating risk of catastrophes appropriately, risk managers must be fully informed about the likelihood of these events occurring and the potential impact they could have on their organizations.
Over the past 30 years, the frequency and severity of catastrophic losses has steadily increased, largely due to growing concentrations of property in high-risk areas, increased construction costs and, most recently, the magnitude of losses from terrorist attacks. For these reasons and more, sound risk management practice calls for sophisticated catastrophe risk management. Fortunately, new technology and software are available to help risk managers better understand and manage catastrophic risks.
Over the past two decades, the use of computer modeling to estimate future losses from catastrophes has become standard practice among insurers and reinsurers. Modeling is a critical component of insurers' operational success, helping them with underwriting, portfolio optimization, identification of market opportunities and other risk management processes. The earliest catastrophe models were developed specifically to evaluate loss potential at the ZIP code and county level. Developments over the past several years, however, now enable models to assess risk at the address level, offering the greatest insights to the corporate risk manager.
Where's the Risk?
Catastrophes impact every region of the United States, leaving all properties at risk from one peril or another. While risk managers are aware of the obvious risk of hurricanes in the Gulf Coast and along the Eastern seaboard, seismic events in the West, severe thunderstorms in the Midwest and terrorism in major urban centers, they must also take into account that within these regions there is wide variability in the actual loss potential.
Also deserving of recognition is the likelihood of extreme events in less obvious areas. In 1811, one of the largest earthquakes to bit the continental United States occurred in the New Madrid Seismic Zone, located in the Mississippi River Valley. If a similar magnitude 8.1 quake were to occur today, the workers' compensation losses alone could exceed $30 billion, with most of the losses being concentrated in the urban areas of St. Louis and Memphis, Tennessee. Of consider the unusual F4 tornado that tore across central Massachusetts in 1953. If that same tornado were to happen today, the resulting insured losses would likely exceed $1 billion.
Focusing on well-known perils at the expense of infrequent but potentially more catastrophic events is a common occurrence. A national real estate company with concentrations in the Midwest and Southern United States recently commissioned a catastrophe loss modeling analysis to prepare for its upcoming renewal. The company's existing policy had limits of $100 million for Florida windstorm and $50 million for earthquake coverage.
A casual look at the company's property values at a state level would have led one to believe that the coverage was appropriate. However, when the loss modeling analysis was performed, it became evident that the company was buying far more wind coverage than it needed and only half the earthquake coverage necessary to match its risk tolerance level.
Why was the coverage so far off? For wind, the distribution of the company's property within Florida made it nearly impossible for them to suffer losses in excess of $20 million from a single storm. But there was no way for this company to understand the potential impact on its portfolio until a catastrophe analysis that simulated the full range of possible storms was complete. …