25 years on: lessons learned from Hurricane Andrew

25 years on: lessons learned from Hurricane Andrew | Insurance Business

25 years on: lessons learned from Hurricane Andrew
The following is an opinion article written by Tom Sabbatelli, senior product manager for North Atlantic hurricane models at RMS.



Last year, at a Lloyd’s Market Association event, I was asked, “Have hurricane catastrophe models become too complicated?” My response, “No, they haven’t.” Let me tell you why.

In 1992, Hurricane Andrew shifted the focus of the young catastrophe modeling industry away from earthquakes and reminded us of the devastating potential of hurricanes. As the insurance industry’s first major hurricane catastrophe, Andrew demonstrated that even following a long stretch of quiet Atlantic hurricane activity, “it only takes one” to cause catastrophic effects to life and property. Following Andrew, the industry was overwhelmed from a lack of preparedness for such a severe loss that, compounded with one of the first instances of post-event loss amplification, it contributed to more than 10 companies becoming insolvent. The lessons learned in Andrew’s aftermath shaped the structure of present-day hurricane models.

Celebrate excellence in insurance. Nominate a worthy colleague for the Insurance Business Awards!

Andrew’s high winds and storm surge unexpectedly exposed significant inconsistencies in building code implementation and enforcement. The need for stronger building materials and practices culminated in the first of several updates to the Florida Building Code. Each update strengthened the previous edition. We now understand that the year of a building’s construction serves as a critical data entry to hurricane catastrophe models, indicating the standards by which each subject-at-risk has been built.

Soon after Andrew, the Florida Legislature mandated insurance discounts to homeowners that installed enhancements designed to mitigate future wind damage. A wide variety of these products have been released to market, including storm shutters and wind-resistant roof shingles. Model vendors commonly refer to these features as “secondary characteristics,” but, when captured within the hurricane model, they can reduce modeled loss uncertainty and promote more accurate underwriting and portfolio management practices.

In 1995, with the industry’s adoption of hurricane models following Andrew, the Florida Legislature established the Florida Commission on Hurricane Loss Projection Methodology (FCHLPM). The Commission’s aim is to protect homeowners and insurers by setting standards to rigorously evaluate model methodologies. FCHLPM updates its standards every two years, leading to a regular cycle of frequent hurricane model updates.

Updated FCHLPM standards often lead to new hurricane model features and increased model granularity. For model vendors, these standards are not mere suggestions; a model must comply with all standards to be used in Florida residential rate filings. The industry as a whole watches the FCHLPM process closely and often considers it to be a determinant of a model’s overall scientific credibility, which only increases the process’s significance.

Andrew served as the industry’s first opportunity to learn from a real-time hurricane and helped define present-day hurricane model framework. Discoveries from subsequent hurricanes have inspired expanded functionality: Each new feature represents a lesson learned by the catastrophe modeling community.

A series of storms over the last 15 years – Ivan, Katrina, Ike and Sandy – highlighted the importance of capturing loss associated with storm surge. Destructive storm surges revealed difficulties in distinguishing between local wind and surge damage. Storm surge loss “leaking” into wind policy payouts is inevitable as claims adjusters are required to decipher how much damage to attribute to wind or to surge. These events demonstrated the need to realistically represent how losses should be paid out from a combination of wind and storm surge.

Sandy in 2012, an event where storm surge heights greatly overshadowed wind speeds, reinforced a commitment to hydrodynamic storm surge modeling, in which water accumulates over a storm’s entire lifecycle. The inundation of lower Manhattan paralyzed critical infrastructure, leading to new insights on downtime and business interruption. As a result, models now offer greater functionality in modeling essential contents below ground.

Currently, we find ourselves in a “hurricane drought,” much like the tranquil period that preceded Andrew. Many in our industry have not experienced an event on the magnitude of Andrew, but catastrophe models have better prepared us for such events. Only one insured company failed during the devastating 2004 and 2005 hurricane seasons hurricane season, reminding us of the need for comprehensive risk models.

Hurricane models will continue to evolve over the next 25 years and beyond, with every landfalling Atlantic hurricane serving as an opportunity to continue to educate our industry about hurricane science and damage. We may see regulators push to increase standards for modeling that will require a more sound and granular view of risk; however, this will only serve to strengthen our confidence in providing an accurate view of hurricane frequency and severity. With every new release, each added feature is a critical piece to achieving a better understanding of risk.

The preceding was an opinion article written by Tom Sabbatelli, senior product manager for North Atlantic hurricane models at RMS. The views expressed within the article are not necessarily reflective of those of Insurance Business.


Related stories:

If Hurricane Andrew were to strike again today, the uninsured loss would be staggering
Wind and hail coverage important for “all property owners"