X hits on this document

PDF document

Draft for discussion at ICAEW IISC meeting, 20th June 2005 - page 39 / 51





39 / 51

that new international accounting standards will be introduced (in 2005) and that the way we value and report insurance earnings will change. Much of the discussion centred on the models and techniques we will need to use to conform. For some contracts, stochastic techniques will be required, and several of the speakers explained how advances are being made in this area. The Baird Report has recommended the use of stochastic modelling to value options and guarantees, and this seems now to be widely accepted. Nevertheless, it was pointed out that some life offices are not capable currently of doing stochastic modelling, and there are question marks over how assumptions should be determined (and validated) in any such model; some also queried whether it was possible to credibly model issues of actuarial discretion, such as the determination of bonus rates.

Speakers from the accounting community urged a dash of pragmatism: a "good" answer would be acceptable in many cases rather than trying to produce an "exact" one. Perhaps we need to set ourselves the goal of being able to calculate " fair enough" values rather than exact fair values if the latter can only emerge from a black box. An approach built around a baseline stochastic model accompanied by specific scenario tests might be best.

Standards Many speakers supported the idea of an Actuarial Standards Board, as the paper proposes, which would issue guidance on standards, assumptions and best practice, to prevent valuations becoming too subjective. This could incorporate guidance on suitable stochastic models and scenarios. During the meeting, several specific issues of approach were discussed including:

  • It appears likely that fair value for insurance contracts will be calculated using an entity specific approach, meaning that lapse and expense assumptions would be specific to the company, not based on market averages.

  • The calculations are likely to ignore the credit rating of the insurer, though the IASB drafts could yet take a different view.

  • Any inherited estate should be assumed to be split 90:10 in the absence of a different explicit agreement.

  • There were different opinions on whether liabilities should be valued including an excess over best estimate assumptions - i.e. a market value margin (MVM) for risk. It was suggested that the latest IASB proposal might allow MVMs for diversifiable risk, but the view from the floor that this was contrary to financial theory and market values. Since there is no obvious objective way to determine such adjustments, some felt that allowing them would open the profession to accusations of retaining too much discretion.

  • Companies should not be allowed a free hand when modelling management discretion. A particular course of action should only be allowed for if supported by clear historical evidence or there are well- documented procedures.

Other questions were left unanswered, but the profession will need to turn its thoughts to these:

  • When we report results, how will we deal with projected glidepaths and smoothing accounts for with- profits? How can actuaries define the rules or guidelines they will follow for recommending bonus rates, smoothing and asset mix - and what do we mean by PRE?

  • What treatment should be assumed for in-force funds under management - should their retention for many years be assumed when valuing an insurer?


Document info
Document views212
Page views212
Page last viewedMon Jan 23 17:11:00 UTC 2017