The Board considered a review of the project. Certain issues in measurement that remain troublesome to Board members, constituents, or both were noted. They are:

Deposit Floor

In finalising IAS 39, the Board concluded that the fair value of a liability with a demand feature cannot be less than the amount that a holder can demand. While the computation of this amount is straightforward, the resulting measurement would be unanimously rejected by the industry and the actuarial profession. Their arguments would be similar to those expressed by some banks in the context of macro hedging. The insurance industry is, however, in a somewhat different position than other financial institutions:

a. The accounting for insurance contracts is done on a closed-book basis. While the issue of renewal premiums (below) is difficult, insurers limit their computations to future inflows from a closed book of contracts. In contrast, banks argue that the fair value of deposit liabilities extends beyond the runoff of existing account balances.

b. Insurers incur significant costs in selling and underwriting contracts - well in excess of the costs incurred by other financial institutions. While this does not overcome the conceptual arguments for the deposit floor, it does magnify the effects of that floor.

c. Insurers recognise the risks of allowing early withdrawal, and the accompanying risks of allowing renewal at a fixed premium. They price their contracts with those risks in mind. However, they expect that policyholders will not all demand their surrender values at one time. They consider that possibility as unlikely - as unreasonable as the possibility that all policyholders will die tomorrow or file claims tomorrow equal to the maximum allowed under the contract.

d. The Board will have to consider the application of a deposit floor to insurance contracts and, if it decides that the floor does not apply, how insurance contracts differ from other financial contracts.

Intangible and Deferred Acquisition Costs (DAC)

Existing deferral-and-matching accounting models often record an amount equal to front-end costs as deferred acquisition costs or DAC. The Board previously rejected capitalisation, reasoning that those costs do not meet the definition of an asset. That decision was framed in the context of a measurement that was not constrained by a deposit floor. The significant initial costs incurred by an insurer might be characterised as an investment in an intangible asset. The intangible asset is unlikely to satisfy the recognition requirements of IAS 38. The Board will have to consider whether it should be constrained by those requirements in resolving the issues in this project.

Accounting for a separate intangible poses a number of practical and conceptual problems. Prospective measurement systems developed by actuaries treat the insurance contract as a single set of cash flows, the present value of which produces a single net amount. Allocating that amount to separate asset and liability balances, as in US GAAP or embedded value conventions, is an arbitrary process. Neither balance can be said to be the 'right' amount.

Instead, an accounting convention governs the liability, and the asset is the amount necessary to produce a net amount equal to the present value.

Renewal Premiums

Most existing measurement models for long-duration contracts are based on the present value of future premium inflows and claim outflows. As highlighted in the Issues Paper, the DSOP, and several Board discussions, those inflows and outflows do not meet the definition of assets and liabilities when considered alone. In January, the Board tentatively concluded that a measurement model should include those flows. However, some Board members expressed concerns about the implications of that decision for other accounting measurements.

Risk Premiums

Given the absence of observable market information, most measurements of insurance liabilities will be derived from estimates of the present value of future cash flows. The insurance industry and actuarial profession are accustomed to such measurements and should have the ability to make the estimates in most situations. However, an estimate of fair value includes the amount that marketplace participants demand for accepting risk. Many constituents have complained that the very absence of market information that leads to use of present value also makes it difficult or impossible to develop reliable estimates of the risk premium. In January, the Board concluded that this amount should be determined by calibrating the measurement against amounts that would be charged for new contracts with similar exposures and characteristics. Some may see this as an answer to a 'which market' issue, while others see it as a practical expedient. Either way, this answer has been criticised by actuaries as inconsistent with their understanding of the conceptual model.

Discount Rates and Investment Margins

The Board previously concluded that the measurement of an insurance contract should not include estimates of future investment margins. In effect, this means that the discount rate used to compute the present value of future cash flows cannot be based on expected asset returns. The rate, therefore, is the risk free rate, adjusted by the credit standing of the book of contracts. Again, most insurers and actuaries criticise this conclusion as inconsistent with the pricing of their contracts and the amounts at which observable prices for transfers (usually in business combinations) are computed.

Participating (with Profits) Contracts

In much of the world, long-duration contracts include a participating feature that requires the insurer to apportion its income between policyholders and shareholders. However, this feature works in a variety of ways, depending on the contract and the jurisdiction. For example:

a. The amount that must be apportioned varies with contracts and jurisdictions and, in some cases, is 100 percent.(In those cases, the shareholders, usually a parent company, benefit by charging the insurance subsidiary for services provided.)

b. An insurer may choose to apportion more than required by the contract and may be constrained to do so by competitive pressures, past practices, or promises communicated to policyholders.

c. Once apportioned, the amount may not be declared as a dividend to individual policyholders for some years. As a result, some policyholders may benefit from income earned before they entered into their contracts. Policyholders who die or surrender their contracts may or may not have rights to part of the apportioned amount.

d. In some jurisdictions, a dividend declared to individual policyholders cannot be taken back. In others, amounts may be allocated to individual contracts but can be taken back in the face of losses.

e. The basis for apportioning between policyholders and shareholders is often based on regulatory computations that do not mirror financial-reporting requirements.

The Board has yet to consider questions of participating contracts in depth. The issues described in the preceding paragraph suggest to some that those contracts may have characteristics of equity instruments, not unlike those of a non-controlling ownership interest.

Unbundling

Every insurance contract has elements of insurance and investment. An insurance contract always has a time element, in that premiums are 'invested' by the insurer in advance against the possibility of claims for future events. The insurer's ability to invest premium inflows in advance of claim outflows is a major part of its profitability. In some contracts, including most long-duration contracts, the cash flows attributable to the investment element are predominant. This suggests to some that the best accounting solution would be to separate, or unbundle, the contract into pure insurance and investment elements.

It was noted that there will be sufficient staff and Board resources to complete the project, and it is anticipated that there would be an exposure draft in June 2005.

The staff proposed to develop a 'knowledge book', which compares and contrasts the model developed to date (the fair value model) with other insurance accounting models currently employed in various national jurisdictions. This document would initially be compiled from materials developed by the IASC Steering Committee and by staff during the course of the project to date. It would then be added to over to time.

The Board accepted that proposal.

Staff noted that the Board would devote a block of two to three days to a re-education of the issues.

The Board agreed that some form of field visits should take place during the project.