Florida Commission on Hurricane Loss Projection Methodology Meeting Report: June 23

Jun 25, 2008

On Monday, June 23, 2008, the Florida Commission on Hurricane Loss Projection Methodology (“Commission”) held a meeting to review the EQECAT, Inc. USWIND® 5.13/WORLDCATenterprise™ 3.11 model and Florida Public Hurricane Loss Model  Version 3.0 (“Public Model”) for acceptability.

The meeting was called to order by Chairman Randy Dumm, Insurance Finance Expert, Florida State University.  The following Commission members also were in attendance:

  • Kristen Bessette, Actuary, Property and Casualty Industry
  • Terry Butler, Consumer Advocate, Florida Department of Financial Services
  • Howard Eagelfeld, Actuary, Florida Office of Insurance Regulation
  • Jainendra Navlakha, Computer System Design Expert, Florida International University
  • Jack Nicholson, Senior FHCF Officer, Florida Hurricane Catastrophe Fund (“FHCF”)
  • Scott Wallace, Executive Director, Citizens Property Insurance Corporation
  • Hugh Willoughby, Meteorology Expert, Florida International University

Also in attendance were members of the Commission Professional Team and representatives from EQECAT and the Public Model. 

The EQECAT representative explained the USWIND® 5.13/WORLDCATenterprise™ 3.11 loss model by detailing its measurements, execution and what changes have been made since its last submission.  In the previous submission, three standards dealing with trade-secret information could not be verified.  During the model’s second trade-secret audit, however, those standards were verified to evidence substantial changes from the previous submission.  To view information on EQECAT modeling and previous model submissions, click here.

A Public Model representative provided an update from the Version 2.6 Public Model submitted in 2007.   This year’s submission, Version 3.0, incorporates modifications suggested by the Professional Team and newer data.  The Public Model, which is publicly-available (except for the source code used to write the actual program), currently is based at the Laboratory for Insurance, Economic and Financial Research located in the International Hurricane Research Center at Florida International University (“FIU”).  The Public Model is operated by a team of professors and graduate students from the departments of computer science, finance, and statistics at FIU, with the help of meteorology and engineering experts from other universities and institutions.  To view information about the Public Model, click here.

Both models were reviewed by the Commission for the following standards:

General Standards

  1. Scope of the Computer Model and Its Implementation
  2. Qualifications of Modeler Personnel and Consultants
  3. Risk Location
  4. Independence of Model Components
  5. Editorial Compliance

Meteorological Standards

  1. Base Hurricane Storm Set
  2. Hurricane Parameters and Characteristics
  3. Hurricane Probabilities
  4. Hurricane Wind Field Structure
  5. Landfall and Over-Land Weakening Methodologies
  6. Logical Relationships of Hurricane Characteristics

Vulnerability Standards

  1. Derivation of Vulnerability Functions
  2. Mitigation Measures

Actuarial Standards

  1. Modeled Loss Costs
  2. Underwriting Assumptions
  3. Loss Cost Projections
  4. Demand Surge
  5. User Inputs
  6. Logical Relationship to Risk
  7. Deductibles and Policy Limits
  8. Contents Vulnerability
  9. Additional Living Expense
  10. Output Ranges

Statistical Standards

  1. Modeled Results and Goodness-of-Fit
  2. Sensitivity Analysis for Model Output
  3. Uncertainty Analysis for Model Output
  4. County-Level Aggregation
  5. Replication of Known Hurricane Losses
  6. Comparison of Projected Hurricane Loss Costs

Computer Standards

  1. Documentation
  2. Requirements
  3. Model Architecture and Component Design
  4. Implementation
  5. Verification
  6. Model Maintenance and Revision
  7. Security

During the standard verification process, Commission members asked EQECAT representatives questions about hurricane windfall structures, land-cover validation methods, historical storm record statistics, mitigation measures, underwriting assumptions, and output ranges. 

Regarding the Public Model, Commission members inquired about the substantial changes from Version 2.6 to 3.0.  Commission members also cautioned that verification standards for the new data need to be increased and typographical errors need to be corrected. 

Commission members expressed interest in the Public Model Version 3.0’s use of oceanic temperature statistics in adherence to Standard G1.  Public Model representatives stated that temperature increases throughout the depth of the ocean could lead to increased intensification of hurricanes.  Of the models submitted to the Commission, only one other captures this type of information.  A Public Model representative stated that, although the Public Model Version 3.0 uses historical data only, the oceanic temperature information could possibly be used to make predictions.

After lengthy discussion, both models were accepted. 

Should you have any questions or comments, please do not hesitate to contact Colodny Fass.



To unsubscribe from this newsletter, please send an e-mail to ccochran@cftlaw.com