By Richard Hartigan, IFRS 17 Actuary at Hiscox
The Insurer Stands Ready doctrine works for business that earns linearly. Extending the Insurer Stands Ready doctrine to business that earns non-linearly violates the very principles underpinning IFRS 17.
Example:
1 year policy attaching 1 January, covering the US wind peril only. Tornadoes occur all year round and hurricanes in Q3, so the earning pattern is Q1 10%, Q2 10%, Q3 70%, Q4 10%. Strict interpretation of IFRS 17 leads to 50% of the CSM being amortised in the first 6 months (when only 20% of the premium has been earned).
It is important to note that the original development of the CU is thus: “coverage units were introduced to achieve an appropriate allocation of the contractual service margin of a group that contains contracts of different sizes.” (Para 22(a) of TRG staff paper AP3, May 2018). It appears possible that the CU was developed for one purpose (possibly with Life business in mind, and probably for a valid purpose) and then re-purposed for a completely invalid purpose unsuited to non-Life business. Read Para 25-30, and especially Para 29 (ibid): “The staff also acknowledge the calls from TRG members for a principle-based approach—it is not possible to set detailed requirements that will apply appropriately to the wide variety of products.”
In advocating for (or accepting) a principle-based approach the TRG noted:
• “IFRS 17 does not specify a particular method or methods to determine the quantity of benefits. Different methods may achieve the objective of reflecting the services provided in each period, depending on facts and circumstances.” (Para 35(g) of TRG Minutes, 2 May 2018)
• [as one possible method, adhering to IFRS 17 principles] “methods based on expected cash flows. However, methods that result in no allocation of the contractual service margin to periods in which the entity is standing ready to meet valid claims do not meet the objective.” (Para 35(h)(v), ibid)
The soundest path is to completely reject amortising CSM on a linear basis. An insurer may then adopt the method in the second dot-point above, which is (as if by magic) on an earned basis in All But Name: linguistic gymnastics worthy of Sir Humphrey Appleby. One hopes auditors will be sympathetic to a purely principle-based approach.
Intimately connected to the above is the notion that the Risk Adjustment (RA) for Liability for Remaining Coverage (LRC) and the CSM are somehow completely different things, the former to be amortised with the Present Value of Future Cash Flows (PVFCF) (i.e. on an earned basis) and the latter to be amortised on a linear basis.
The way I think about the RA and the CSM is different: in economic terms the RA is the ‘normal’ return (in a Return on Capital (RoC) sense) and the CSM is the ‘super-normal’ return (in a RoC sense). That is, the RA and the CSM are the same thing, split apart solely for the purpose of testing for loss-making (onerous) business.
Once one starts thinking in the above way then amortising CSM on a linear basis (for business that earns non-linearly) obviously becomes a nonsense: the service the insurer is providing is obviously the (expected) payment of claims and if that is non-linear then it must be measured thus.
As a post-script, it also becomes evident that amortising CSM on a linear basis significantly increases the likelihood that LRC outcomes differ materially when comparing the General Measurement Model (GMM) approach and the Premium Allocation Approach (PAA), thus impeding efforts to prove PAA eligibility for business that is not automatically PAA eligible.
About the Author:
Richard Hartigan is a general insurance actuary working in London for a major global (re)insurer, currently seconded to their IFRS 17 project team. He graduated from Macquarie University (BEc) in 1992, and completed a MBA (with Distinction) at Cornell University’s Johnson School in 2003. He is a member of both UK (FIA) and Australian (FIAA) institutes.
|
