She spoke to NU a week before the May 20 meeting of the Casualty Actuarial Society in San Diego, and a day after a CATF held a conference call concerning the latest draft of recommended changes to the instructions. On the call, CATF decided to schedule a vote on the draft for next month's NAIC meeting in Philadelphia, allowing comments until June 3.

“We know why regulators are doing this [so quickly],” said Stephen Broadie, assistant vice president of the National Association of Independent Insurers in Des Plaines, Ill., noting that July 1 is the deadline for getting items to the Blanks Task Force. Then, changes dont get made until the 2004 statements are put together.

“We understand this, but there have been five drafts since March,” he said. “On a number of conference calls, we were discussing prior drafts while the Working Group reworked new changes. In some cases, language is being drafted on the fly,” he added, suggesting that a more deliberative process is needed since loss reserves–a key subject of the opinions–are the biggest item on insurer balance sheets.

John Purple, who chairs the CATFs Actuarial Opinion Instructions Working Group (AOIWG) and is chief actuary of Connecticut's insurance department, believes the projects fast pace has “kept people really focused,” although he noted that the group had actually been working on the project long before the first draft was released for comment.

Roughly 10 or 11 months ago, regulators concluded that it was time to step back to see if the opinion instructions, first put in place 10 years ago, were still meeting their needs, he said.

For example, if a company sets its reserves at the low end of a range of estimates, “we want to know that. We dont see that in the opinion,” he said, explaining that the opinion letters currently only tell regulators whether the reserves fall within a reasonable range.

While the dollar ranges that actuaries come up with when theyre analyzing reasonableness might show up in an actuarial report that supports an opinion, “we dont see that report until some time after the blank is filed,” he said.

Opinions are filed with insurer annual statements in March. The deadline for preparing supporting reports, which must be presented to a companys board of directors and are available for regulatory examination, is May 1.

Both NAII and COPLFR voiced early concerns about initial AOIWG drafts that would have put actuarial ranges and best estimates in the opinion letters. The concerns arose because statements of opinion are public documents.

“We are aware of no other instance where there are two sets of numbers in the public domain for items on an entitys balance sheet,” Ms. Sweeny wrote in an April 8 letter to Mr. Purple, referring to reserves actually booked by an insurer and a second estimate (or range of estimates) developed by an actuary.

Having two sets of numbers in the public domain has significant tax and Securities and Exchange Commission disclosure implications, she said, suggesting that a model law would need to be developed to protect the confidentiality of the estimates, which she said could even get into the hands of competitors in some states.

Mr. Purple understands the concern.

“The biggest argument [they had] was that to put out another estimate creates problems for the company and the actuary,” he said, noting that it puts the actuary in a position of opposing their client. The latest draft of the instructions reacts to this concern, requiring that a best estimate or range appear only in the actuarial report, not in the opinion itself.

But that compromise does not go far enough, according to NAII.

“We dont think its appropriate to require it in either place. Instead, its a matter for actuaries to individually determine,” Mr. Broadie said.

“We certainly think a report should include material to allow regulators to see how actuaries arrived at their conclusions,” he said. However, making the disclosure of estimates mandatory bestows an “artificial precision” that elevates their status beyond mere tests of reasonableness, he added.

Practically, however, both Ms. Sweeny and Mr. Purple said that all the actuarial reports theyd seen (and authored) included best estimates or ranges or both.

Mr. Purple noted that by giving in on the issue of disclosing estimates in the opinion, regulators have put back a timeliness issue. Regulators still wont know whether companies book within an actuarys range until reports become available.

As a possible fix, CATF is now considering a proposal raised by the Texas insurance department that would allow for a summary schedule of estimates to be made available to regulators prior to the release of the report. While reports are confidential in most states, a model law might be needed to keep the new schedules confidential, he said.

A second concern for actuaries relates to a proposed instruction that would require them “to state whether or not there is a significant risk of adverse material deviation” in the items they opine on, according to Ms. Sweeny.

“As you sit and write, There is no risk of adverse material deviation, we think that comes close to an assurance that nothing bad will happen,” she said, suggesting that it was unrealistic to think that actuaries could provide such a guarantee–or that they would provide a guarantee without making their opinions much more difficult to read.

“Actuaries will believe it behooves them to add more explanatory language” about the limitations of that kind of guarantee, she said. “That just adds to the verbiage and makes it harder for the reader to find and understand whats important.”

“Youre going to get boilerplate language on the 80 percent of companies that dont have problems,” according to Mr. Broadie, agreeing that actuaries will be “loathe to state that there is no risk of material deviation” for those companies–a situation that obscures the need to focus on the 20 percent of insurers that do have problems.

“We may get more boilerplate,” Mr. Purple conceded. “But were of the opinion, wed like to try and see what we get,” he added, noting that the intent is to get a better understanding of the companies that “are in trouble or trending that way.”

A final proposed instruction that concerns actuaries and industry representatives is a data certification requirement.

“The requirement asks for a lot of new language that would be a significant waste of time and effort,” Mr. Broadie said. It asks the officer of the insurance company who prepared the data that an actuary relied on for his or her analysis to sign a statement saying that the data “contains no material inaccuracies.” Going on to quote a long-winded paragraph verbatim, he said, “this is not going to be terribly productive.”

Both Mr. Broadie and Mr. Purple noted that the instructions already require the actuary to state that he or she evaluated the data provided for reasonableness and reconciled it to annual statement loss schedules (Schedule P).

Many actuaries in their opinions already provide the name of the person–usually the insurance company CFO–who provided the data, Mr. Purple noted. The new data certification by the company officer is “just taking that a little further and getting that person to take responsibility.”

“From our standpoint, it puts the company on notice that we expect this to be complete and accurate,” he said. And, in some cases, he said, “it also takes the actuary off the hook,” noting that actuaries who have signed opinions of companies that later became insolvent have told regulators that they didnt realize the data they relied on was inaccurate in the past.

Ms. Sweeny said the assurance being requested in the new data certification “is not going to change anyone's behavior and just adds paperwork.”

“If regulators are concerned about the quality of data, then such issues can be addressed more directly” through audit processes, she said.

Reproduced from National Underwriter Property & Casualty/Risk & Benefits Management Edition, May 27 2002. Copyright 2002 by The National Underwriter Company in the serial publication. All rights reserved.Copyright in this article as an independent work may be held by the author.