Contingencies Online - September/October 2009

TORTOISE: Thanks for the pep talk, but you can say that about any system. After all this discussion, I’m still stuck because fair values can be useful only if they provide meaningful performance information. I thought that a significant goal of financial reporting is to enable a user to assess expected future performance. Is this measure of market assessment the right benchmark to use? Why not internal objectives used by management (verified by a “use test,” as proposed in Solvency II)? Wouldn’t a better indicator of expected costs be one that’s based on an objective measure of the risks involved?

ACHILLES: Well, I agree with you on that one. The traditional actuarial concept of release from risk can come in handy. In fact, if properly applied, it can be a really good approach for the assessment of performance. But you have to remember that since the primary users of financial statements are expected to be investors or market participants, maybe the best performance metrics should be based on their views instead.

TORTOISE: I might be able to buy that, but first you have to convince me that market-estimated or market-based measures are practical and will be used internally.

ACHILLES: That’s right. Those writing accounting standards do have to get their act together to provide a decent framework from which performance measurement can make sense. Still, maybe actuaries can develop some neat approaches for the purposes of disclosure (that is, disaggregations) to convey meaningful benchmarks for performance measurement. We need effective disclosure of judgments and risks to enable users to make sense of financial statements. These will continue to be a challenge to insurers, no matter what standard is used. Ideally, internal and external benchmarks should be the same. Now that’s something to work toward!

TORTOISE: Another thing I can’t understand is a mixed-at-tribute model in which fair value is only one of the attributes. It’s bad enough to use fair values at all. Isn’t it worse to use different attributes for different assets and liabilities so everything doesn’t line up?

ACHILLES: Another great point. Why don’t those setting accounting standards adopt a single method for everything so that it would all be in sync?

TORTOISE: (Sigh) I think they’d like to. Discussions are under way to reduce the number of measurement methods used in financial instrument accounting. And reviews of insurance and service contract measurement are ongoing. But it’s tough to ignore history in some areas and drastically change so many things when most stakeholders, including politicians, disagree with you. Accounting standards-setters used to declare their ultimate goal was the broad adoption of fair values for all financial assets and liabilities. Is that still the objective?

ACHILLES: I’m not sure that will ever be achieved, even though I know that some participants are still trying.

TORTOISE: Don’t you think we’ll need further guidance on modeling fair values in areas such as discount rates and margins? If so, who’ll provide it?

ACHILLES: I hope that the accounting standards boards provide a clear set of objectives. I’m confident that actuarial practice will develop according to whatever framework is chosen. We need actuarial leadership—otherwise, useful information won’t be provided at all. U.S. actuaries are already challenged in developing fair values in response to the FASB’s SFAS 157, business combinations and the fair value option. I’m looking forward to following future developments in this area, both in the U.S. and internationally. I think that fair values, whether or not they apply to insurance contracts, are here to stay. Actuaries will have to master their measurement.

TORTOISE: Thanks! Now go take a shower. Although this sort of makes sense, it requires a lot of work and understanding, as well as more research and practical development.