How to Measure Anything in Cybersecurity Risk, 2nd Edition

How Catastrophe Modeling Can Be Applied to Cyber Risk

Scott Stransky, assistant vice president and principal scientist at AIR Worldwide

Tomas Girnius, PhD, manager and principal scientist at AIR Worldwide

One may wonder how a company proficient in building models to estimate losses from hurricanes and other natural disasters can use their techniques to build a similar model for estimating losses from cyberattacks.

Hurricane Andrew spawned the catastrophe‐modeling industry. Although catastrophe models existed before that storm in 1992, they were not used by decision makers, nor were they used to their full potential. When the storm struck south Florida, AIR issued modeled loss estimates on the order of $13 billion, a figure that the insurance industry scoffed at for being far too high. As the claims for Andrew started to pile up, 11 insurers went out of business, and the rest of the industry began to see the value in running models. The “Hurricane Andrew of Cyber” has yet to strike the cyber insurance industry, and when it does, those companies using models will be far better off than those using so‐called underwriting judgment.

AIR is employing the same stochastic modeling framework (Figure B.15) that it has reliably used for its catastrophe model for nearly 30 years. This is best described by analogy to hurricane modeling. Hurricanes can be visualized and have been widely studied. We begin with historical data on hurricane events, publicly available from the National Hurricane ...