NAIC Catastrophe Risk Subgroup Reviews Proposal for Risk-Based Capital Charge for Results of Property Catastrophe Modeling

Apr 22, 2010


The Catastrophe Risk Subgroup (“Subgroup”) of the National Association of Insurance Commissioners (“NAIC”) met by teleconference on April 21, 2010 to continue its review of a 17-item draft proposal for a risk-based capital (“RBC”) charge for property catastrophe risk based on the results of catastrophe modeling.  A copy of the proposal with added comments is attached.

Subgroup Chairman Ronald Dahlquist, Senior Actuary with the California Department of Insurance, said that the Subgroup’s key tasks were to review model option choices and evaluate the issue of data quality accuracy and completeness.

Observing that the Subgroup previously had come to a general agreement on items One through Five of the draft proposal, Mr. Dahlquist invited discussion on items Six and Seven.

Item Six states, among other points, that a separate contingent credit risk charge will be calculated for hurricane and earthquake peril.  The charge for each will be 10 percent of the expected ceded loss generated by the model for a 1-in-100 year expected loss level.

Among the comments received in advance by the NAIC on this item was one referring to evidence that rating agencies apply credit risk charges that are much smaller than 10 percent of the reinsurance receivable.  Mr. Dahlquist asked if the time is right to refer the issues of whether there should be a contingent credit risk charge and how much to the Property and Casualty Risk-Based Capital Working Group (“Working Group”)-parent of the Subgroup.

Participants also raised questions about the reference in Item Six to receivables due to intercompany pooling or mandatory pools and associations.  Mr. Dahlquist agreed that further discussion was needed on these and related points.

The Subgroup then considered Item Seven, which states that insurers would be free to utilize whatever assumptions they deem appropriate in their modeling for risk-based capital purposes, since these are the same assumptions used for internal catastrophe risk management.  Under the provisions of the proposal, insurers also would be required to file a confidential report that includes commentary on their treatment of certain modeling options and assumptions.

Calling this item “the most substantive area of change” under consideration, Mr. Dahlquist noted that a number of regulators were uneasy about “giving companies a free hand to choose whatever modeling assumptions they wanted.”

“There is a degree of ambiguity and discretion that may be more than we’re bargaining for,” observed Subgroup member Howard Eagelfeld, an actuary with the Florida Office of Insurance Regulation.

Mr. Dahlquist said large insurers likely use a variety of models for different internal purposes, which could make it difficult for regulators to know which model should apply for RBC purposes.

One participant observed that Item Seven appeared to be in conflict with Item One, which would require insurers to base its catastrophe loss calculation on one of three commercially available catastrophe risk models.

Mr. Dahlquist invited regulators and interested parties to submit written comments on this item.  However, after discussion, the Subgroup decided to leave the proposed language of Item Seven as it stands and send it to the Working Group for consideration.

It was suggested that modeled losses for higher risk levels  (1-in-200 year and 1-in-500 year losses) should be included in the RBC Report. 

The next Subgroup meeting was set for May 19.  Among the topics of the next meeting will be data quality accuracy and completeness.

Meeting materials are attached, including the aforementioned comments on items considered during the April 21 meeting.