Florida Commission on Hurricane Loss Projection Methodology Reviews Committee Reports

 

The Florida Commission on Hurricane Loss Projection Methodology (“FCHLPM”) met yesterday, August 14, 2013 for the second day of a three-day meeting.  To view the three-day agenda, click here.

Yesterday’s focus was on FCHLPM committee meetings.  A report on each follows:

 

General Standards Committee

Chaired by Kristin Piltzecker, the FCHLPM General Standards Committee includes Barry Gilway, Bob Lee, Jai Navlakha and Dr. Jack Nicholson.  A recommendation was made to include under Standard G-1 “Scope of the Computer Model and Its Implementation” audit items a requirement that modeler-specific publications cited must be made available in hard copy or by a Web link. 

Changes also were made to Disclosure 5(B), which was amended to reflect the fact that 2007 data from the Florida Hurricane Catastrophe Fund (“FHCF”) will now be used to determine the percentage difference in average annual zero deductible statewide loss cost.

There was also a suggestion that the average annual zero deductible statewide loss cost should be required for each individual model component change, which is defined as a change affecting a single Standard.

Modelers present at the meeting expressed concern over potential intellectual property issues with isolating individual changes affecting single Standards. There was discussion to move the isolation of individual changes to the Audit section.

Modelers said they feel more comfortable with the current process that allows them to explain model changes and the interactions between changes during the on-site portion of the model approval process.  This requirement, they said, would substantially increase the amount of time for model approval.  Meanwhile, the regression testing provides the FCHLPM with the ability to audit all changes to models.

Further, analyzing changes in isolation may provide false results, they explained, due to the fact that multiple changes are often made in order to produce a single result.  Analyzing those changes in isolation will provide an erroneous result, modelers said.  The FCHLPM agreed to change Disclosure 5(B)(2) to require the average annual zero deductible statewide loss cost for each individual model component and subcomponent change.

It was suggested that Standard G-3, “Risk Location,” should include a requirement that when ZIP Codes used in a model are updated, all ZIP Code-based databases must be updated.  Further, the Committee suggested that all transitions among street addresses, geocode locations (latitude-longitude) and ZIP Codes shall be consistent and appropriate.

The FCHLPM settled on including a requirement that all geocoding methodology shall be consistent and justifiable.

 

Meteorological Standards Committee

The FCHLPM’s Meteorological Standards Committee is chaired by Hugh Willoughby.   Committee members are Randy Dumm, Bryan Koon, Dr. Nicholson and Floyd Yager.

The Committee suggested adding a paragraph to the purpose of Standard M-1, “Base Hurricane Storm Set,” stating that any variations between modeler hurricane characteristics and the new HURDAT2 fields must be consistent and justified.  A suggestion was made to include a statement in the Audit section of M-1 that changes to a modeling organization’s base hurricane storm set from previously accepted submissions will be reviewed.

Concerns with the use of HURDAT2 were expressed by modelers present at the meeting who do not believe it provides an accurate representation of operational wind radii.  It was stated that the National Hurricane Center will not use HURDAT2 for verification of wind radii forecasts.

A suggestion was made to change Standard M-4(B) to reflect that use and land cover databases with a publication date prior to 2010 will not be acceptable. Also, a suggestion was made to change the audit section of M-4, “Hurricane Windfield Structure,” to require a modeler to provide the previous and current storm parameters used in calculating loss costs, as well as provide the resulting spatial distribution wind structures.  Modelers expressed concern over having a hard publication date set for databases.  Their fear is that a set publication date will inevitably need to be updated and potentially lead to the need to update the databases on a more frequent period.  Ultimately the Committee agreed to change M-4(B) to:  “The land use and land cover databases shall be consistent with MRLC2012. Use of alternate databases shall be justified.”

Suggestions for changes to disclosures under Standard M-5, “Landfall and Over-Land Weakening Methodologies” were presented to the FCHLPM.  It was suggested that Disclosure 7 should be added to require a modeler to describe and justify underlying formulation of the storm surge calculation, including whether it includes wave action.  Modelers expressed concern that, by including this data in the disclosure section, they will potentially expose proprietary data.  The Committee agreed to keep this requirement in the audit section and therefore non-public. 

The Committee also suggested adding disclosure 14, which would require modelers to provide a comparison of storm surge calculated in the model worst-case scenario for the same five locations and compare with the National Oceanic and Atmospheric Administration’s (NOAA) Maximum Envelope of Water (MEOW) for each location.  Modelers again expressed concern that this addition may expose proprietary data.  Committee members asked if adding this requirement to the audit section would alleviate concerns, since the data would be removed from the public domain.  Modelers then expressed concern that this item should not be included in the audit section because there is no existing Standard for storm surge and therefore no Standard for an audit.

 

Statistical Standards Committee

The Statistical Standards Committee is chaired by Dr. Lori Medders.  The committee members are Bryon Koon, Florida Insurance Consumer Advocate Robin Westcott, Hugh Willoughby and Floyd Yager.

The major changes being suggested for the Statistical Standards are to move their section behind the Meteorological Standards section and to split Form S-2 into Forms S-2A and S-B.

The Committee stated that moving the Statistical section behind the Meteorological section would provide efficiency in evaluating models. 

Form S-2A will seek examples of loss exceeding estimates using 2007 FHCF exposure data.  Form S-2 B will seek examples of loss exceeding estimates using 2012 FHCF exposure data.  Form S-5 was broken into Part A incorporating 2007 FHCF exposure data and Part B incorporating 2012 FHCF exposure data.

 

Vulnerability Standards Committee

The Vulnerability Standards Committee is chaired by Randy Dumm.  The members are Bob Lee, Jai Navlakha, Robin Westcott and Floyd Yager.

The Committee suggested incorporating the term “Building” into Standard V-1,” Derivation of Building Vulnerability Functions,” for which several additional disclosures were suggested.  Modelers will be asked to describe the assumptions, data, methods and processes used to develop building vulnerability functions when some primary characteristics are unknown.  They also will be asked to describe how a model addresses wind-borne missile impacts and identify whether it has the capability to determine storm surge losses.  How building vulnerability functions for storm surges were developed will also be queried.

A new audit item was suggested for Standard V-1 that would audit modifications to the building vulnerability component, including reasons for the modification, its scope, and the impacts of the modification on the building vulnerability component.

Significant changes were suggested to Standard V-2, “Derivation of Contents and Time Element Vulnerability Functions.”  The suggested addition would require development of the contents and time element vulnerability functions to be based on historical data, tests, rational statistical analysis and site inspections.

New disclosures were recommended for V-2.  Modelers will be asked to describe modification to the contents and time element vulnerability component, and provide a flow chart documenting the process by which the contents vulnerability functions are derived and implemented.   They also will be asked to describe the number of contents vulnerability functions and whether vulnerability relationships are used for personal residential, commercial residential, mobile home, condominium insurance, etc.  Descriptions of the assumptions, data, methods, and process used to develop and validate contents and time element vulnerability functions concerning insurance company claims also will be required.

Additional audit items have been suggested for V-2.  The FCHLPM proposes to audit modifications to the contents and time vulnerability component in the models.

The Committee suggested amending Purpose Item 6 under Standard V-3 “Mitigation Measures-Mean Damage Ratio (Trade Secret)” to specifically include entry doors, garage doors and sliding glass doors as mitigation measures to be included in models.  Additional disclosure and audit items under V-3 were suggested to require modelers to describe modifications to mitigation measures in the model.

The FCHLPM also addressed the definitions noted below:

Bathymetry:  Spatial variation of ocean depth relative to mean sea level.

Conversion Factor:  Either the ratio of the one-minute 10-meter wind to a reference wind (e.g. another level, gradient wind or boundary layer depth average) or a constant used to convert one unit of measure to another (as in 1 knot = 1.15 mph).

Geographic location data:  Information related to the geocoding process within the model software.

Gust Factor:  Ratio of the strongest wind speed within a specified interval of time (such as three-second or one-second) to the mean wind speed.

Material Change to a Model:  A change to a model that may result in, but is not limited to, one of the following examples –1. Change in the loss cost of $0.001 or greater per $1,000 in any ZIP Code.  2. Change in the Probable Maximum Loss in Form A-8 greater than one percent of the previous value.

(Note that either of these conditions implies a material change to the model, but is not necessarily the only change.)  The FCHLPM must determine in its judgment whether a change is material.

 

Should you have any comments or questions, please contact Colodny Fass& Webb.

 

 

Click here to follow Colodny Fass& Webb on Twitter (@CFTLAWcom)

 

 

To unsubscribe from this newsletter, please send an email to Brooke Ellis at bellis@cftlaw.com.