Work continues to create a framework for establishing an agent to gather mortality data from companies that would be used both to provide experience that can be used to analyze trends as well as check individual valuation data against industry data, if a regulator felt it was necessary.

Regulators of a statistical agent subgroup of the National Association of Insurance Commissioners’ Life & Health Actuarial Task Force shared ideas and had questions about the development of a statistical agent.

The issue is being looked at in terms of companies using the 2001 CSO preferred class structure tables and is also being considered for possible use if a principles-based reserving system becomes a reality. The compiled data could be used in experience table development for issues ranging from annuity persistency to expense factors and mortality experience, according to discussions.

During a discussion on October 23, a number of questions came up over how the development of a statistical agent should proceed. For instance, one question raised was whether a statistical agent would simply gather data, or both gather and scrub/prepare data for use.

John Rink, a Nebraska regulator, said his department does not have the manpower to deal with processing data submissions from companies and that any work on this project should keep costs to both insurance departments and companies in mind. Utah regulator Tomas Serbinowski agreed, noting that his department would not want to handle data submissions.

Additionally, there was also a discussion about whether there would be one statistical agent that would be used nationally or several stat agents companies could use.

Companies should have an option of 2-3 statistical agents rather than a large number, said John Purple, a Connecticut regulator.

Another question raised was “the purpose for collecting the data.” John Bruins, a life actuary with the American Council of Life Insurers, Washington, requested clarification on whether data would be used to provide experience studies and ranges that could be used in valuation or whether it would be used to compare aggregate data against individual company filings.

Bruins noted that if the purpose is to conduct national studies then it is not necessary to collect information from all companies but rather from the top 50-75 companies in order to have enough data to accurately reflect the market.

It would be useful if insurers knew what is anticipated for the end product, said Armand dePalo, chief actuary at Guardian Life Insurance Company of America, New York. The reason, he explained, is that it becomes expensive to reengineer data calls.

Bruins added that it is important that the issue of who will pay for a statistical agent be addressed.

And, it is very important that data be afforded confidentiality protections, said Tom Rhodes, actuarial director with MIB Group, Westwood, Mass.

Randall Stevenson, a life actuary with the NAIC, responded that NAIC legal research indicated that it would depend on how the framework was structured.