Four hours after Hurricane Andrew made landfall near Homestead, Fla. in 1992, catastrophe modeler AIR issued a statement that the insured losses could exceed $13 billion in Florida. Actual losses equaled $15 billion.
Executive Summary
Traditional catastrophe models lack the precision needed to accurately help insurers estimate losses and prepare for the influx of post-disaster claims, says catastrophe risk expert Karen Clark. Instead, she advocates the use of newer open loss modeling platforms, which empower insurers with advanced scientific processes for estimating claims and losses for live events in real time.Five days after Hurricane Irma made landfall near Cudjoe Key, Fla., AIR estimated the losses in the U.S. would be $25-$35 billion (revised from an earlier range of $20-$40 billion). PCS has recently pegged the U.S. loss much lower (exact figure available directly from PCS), and as of this writing in early October, the Florida insurance department has tallied just $4.6 billion in estimated insured losses.
Hurricanes Harvey and Irma clearly demonstrated that 25 years of cat modeling and numerous model updates have not improved the accuracy of traditional hurricane models. Recent events indicate a trend in the opposite direction with wider ranges and less precision in the loss estimates.
