All Over the Map: Catastrophe Modeling and the Impact of Residential Windstorm Deductibles
Oct 24, 2013
Editor's Note: The National Hurricane Center re-categorized Sandy from a "hurricane" to a "post-tropical cyclone" shortly before landfall one year ago this month. This decision dramatically highlights the issue of hurricane deductibles and the many ways that different states handle them. This AIR Current provides guidance for insurers in understanding deductibles in the context of catastrophe modeling.
After Hurricane Andrew impacted Florida in 1992, U.S. insurers began to use higher deductibles for catastrophic perils as a critical tool for controlling potential gross losses from extreme events.
Most policies in Atlantic and Gulf coastal state markets now feature storm deductibles set as a percentage of the insured value of the dominant coverage rather than as a traditional fixed dollar amount. This is the case both in the private sector and for "wind pools," state-backed alternative insurers designed to promote availability and (sometimes) affordability of coastal coverage.
Such deductibles shift some risk to consumers, but in return enable insurers to operate at the higher solvency levels with lower capitalization hurdles, which improves both the availability and quality of coverage in high-risk areas. In the case of wind pools, they reduce the portion of risk borne by statewide taxpayers via post-event debt or assessments.
Unfortunately, the evolution of windstorm deductibles has been haphazard and it has left insurers with a patchwork of laws, regulations, and practices that are, quite literally, "all over the map." Furthermore, immediately after large storms, individual states have recently made widely varying interpretations of such rules, mostly in the direction of unfavorable surprises to insurers.
Key Features of Storm Deductible Requirements
A few key considerations differentiate the deductibles requirements that affect each state:
- Is there a statewide (as opposed to insurer-specific) definition of the trigger for a storm deductible?
- Are storm deductibles restricted to hurricanes, named storms, or all wind losses? More generally, what is the meteorological and legal definition of the triggering event?
- How is the amount of the deductible calculated? Is it a percentage (and of what base), or a flat dollar amount? Is it applied on a per-event or a seasonal basis?
- In New Jersey, no NWS hurricane warning was in effect at landfall and the NHC had reclassified the storm as post-tropical.
- In New York, no sustained hurricane-force winds were measured at landfall.
- In Rhode Island, no NWS hurricane warning was in effect.
- In Connecticut , no NWS hurricane warning was in effect, and no hurricane-force winds impacted the state.
- In Delaware, no sustained hurricane-force winds were measured at time of landfall.
- In Maryland, no NWS hurricane warning was in effect.
- In Pennsylvania, deductibles were not prohibited but were discouraged, in keeping with the consideration that they did not apply in neighboring states.
Catastrophe models will never be able to predict the actions of a government agency. However, AIR's models provide detailed track information and highly granular intensity footprints for simulated events. This information can provide companies with guidance on when, and for how long, the simulated wind intensity over an affected area meets the likely thresholds for governmental declarations in order to assess the possible impact on portfolio losses.
Sensitivity Testing the Impact of "Stroke of the Pen" Risk
One of the most powerful uses of catastrophe models is the ability to test the sensitivity of risk to policy conditions. This capability is particularly useful in assessing the enterprise risk when deductibles are changed retroactively by political or legal actions. An exposure data set may be characterized by multiple deductible scenarios—percentage of value, flat dollar, seasonal—and subjected to loss analysis under expected conditions or under alternative conditions that may prevail if certain policy terms are in dispute. Each event will be represented by its meteorological and ordinal parameters and the financial module will determine the appropriate loss results under each scenario.
When policy condition scenarios are tested thoughtfully in advance, insurers and the trade associations in which they participate are equipped to accurately explain the impact of potential public policy decisions to government stakeholders, thus reducing the chances of hasty decisions that may be financially devastating to the insurers. In this fashion, catastrophe modeling can not only be used to assess the future, it can actually help to influence that future.
Conclusion
The use of models in catastrophe risk assessment helps insurers, consumers, and governments understand the decision points and consequences associated with extreme events. This is never truer than in the critical hours just before and after a storm, when quick and balanced decisions must be made about how millions of insurance contracts are to be interpreted.
The issue of storm deductibles has created uncertainty for U.S. insurers due to the wide variation among state regulations and insurer practices, and this uncertainty has been exacerbated by the sweeping decisions made by multiple government agencies in the immediate aftermath of the meteorologically unusual Hurricane Sandy. Models provide a number of options for parameter identification and the simulation of financial loss outcomes under various scenarios and contract interpretations—options that can be used to facilitate communication and reduce ambiguity in advance of the next large Atlantic storm.
Quick Reference Table of Storm Deductible Rules Along the Atlantic and Gulf Coasts*
State | Perils Allowed | Seasonal Deductible | Trigger Mandated | Trigger Types | Wind Pool Standard | Post-Sandy Prohibition |
Texas | Windstorm | No | No | NWS, Intensity | 1%–5% | No |
Louisiana | Hurricane, Named Storm, Windstorm | Yes | No | NWS, Intensity | Varies | No |
Mississippi | Hurricane, Named Storm (MWUA only) | No | No | NWS, Intensity | 2% | No |
Alabama | Hurricane | No | No | NWS | 5% | No |
Florida | Hurricane | Yes | Yes | NWS, Conditions, <72 hours | Varies | No |
Georgia | Hurricane | No | No | NWS, Intensity | 1% | No |
South Carolina | Hurricane | No | No | NWS, Intensity | 2%-3% | No |
North Carolina | Hurricane, Named Storm | No | No | NWS, Intensity | 1%–5% | No |
Virginia | Named Storm | No | No | NWS, Intensity | USD 1K–5K | No |
Maryland | Hurricane, Named Storm | No | Yes | NWS | 1%–10% | Yes |
Delaware | Hurricane | No | No | NWS, Intensity | USD 2K | Yes |
Pennsylvania | Hurricane, Named Storm | No | No | NWS, Intensity | Varies | Discouraged |
New Jersey | Hurricane (Some ZIPs) | No | Yes | Measured>74 mph | 5%–10% | Yes |
New York | Hurricane (Some counties) | No | Yes | Landfall of Cat >=1, Measured >74 mph | 2% | Yes |
Connecticut | Hurricane | No | Yes | Measured >74 mph | 5% | Yes |
Rhode Island | Hurricane | No | Yes | NWS + Measured >74 mph | Varies | Yes |
Massachusetts | Windstorm | No | No | NWS, Intensity | 1%–5% | No |
*AIR acknowledges the assistance of the Property Casualty Insurers Association of America (PCI) as a source for some of the information presented.