Executive summary
- PAA is available automatically where the coverage period of each contract in a group is one year or less.
- Beyond one year, the entity must demonstrate that PAA would not produce a materially different liability for remaining coverage than the general model.
- The liability for incurred claims is measured on full IFRS 17 principles: current estimates, discounting and an explicit risk adjustment.
- Onerous contract testing is required whenever facts and circumstances indicate a group may be onerous, and a loss component is recognised immediately.
- Practical accounting policy choices — discounting exemptions, expensing acquisition cash flows — materially change reported results and must be applied consistently.
Evidencing eligibility
The one-year criterion is mechanical and needs only documentation of contract terms. The materiality criterion is a modelling exercise: the entity builds a general-model measurement for a representative sample and compares the resulting liability for remaining coverage against the PAA result across plausible scenarios. Significant variability in the expected timing of claims within the coverage period is the usual reason for failure.
- Document the sample selection and why it represents the group.
- Test across scenarios, not only the base case; eligibility that holds only in the central projection is not eligibility.
- Reassess when product design, coverage length or claims patterns change materially.
The liability for incurred claims is not simplified
PAA simplifies only the unexpired portion of coverage. Once a claim is incurred the obligation is measured as fulfilment cash flows: probability-weighted estimates, discounted at current rates, with an explicit risk adjustment. For long-tail liability lines this is where the actuarial work sits, and where the difference from legacy undiscounted reserving is largest.
An accounting policy choice permits an entity not to discount incurred claims expected to be paid within one year of the claim being incurred. That exemption is helpful for motor damage and property, and largely irrelevant for casualty.
Onerous contract testing under PAA
Under PAA no formal CSM exists, so the trigger for testing is facts and circumstances indicating that a group is onerous: deteriorating loss ratios, adverse large-loss experience, pricing below technical cost, or a change in the reinsurance programme. When triggered, the entity measures fulfilment cash flows for the remaining coverage and recognises a loss component for any excess over the carrying liability.
| Indicator | Evidence | Frequency of review |
|---|---|---|
| Combined ratio above 100 on recent cohorts | Cohort-level actual versus expected loss ratios | Quarterly |
| Rate reductions in a hardening claims environment | Rate change index versus claims inflation estimate | At each rate review |
| Large loss or catastrophe within coverage period | Event-level exposure and reinstatement analysis | On occurrence |
| Reinsurance restructuring increasing net retention | Net versus gross projected loss ratios | At renewal |
Where PAA implementations go wrong
- Treating eligibility as a one-off exercise rather than a reassessed judgement.
- Applying the one-year discounting exemption to lines where payment routinely extends beyond a year.
- Expensing insurance acquisition cash flows without considering the effect on onerous testing for the same group.
- Failing to align contract groupings with the way underwriting performance is actually monitored, so that onerous indicators never trigger.
- Reporting a revenue figure that cannot be reconciled to written premium, without an explanation the audit committee can follow.
Practical example
A commercial property account writes 120 of premium for a twelve-month coverage period, with acquisition cash flows of 18. Under PAA the liability for remaining coverage begins at 120 less 18 if acquisition costs are deferred, and is released evenly over the passage of time. Halfway through the year the carrying liability is roughly 51.
Suppose a storm season forecast and two large losses lift the expected remaining-period loss ratio to 115 per cent of unearned premium. Fulfilment cash flows for the unexpired coverage are then about 59 against a carrying liability of 51, and a loss component of 8 is recognised immediately — five months before the legacy unexpired risk reserve test would have shown anything.
Limitations and caveats
- PAA reduces measurement effort but not data effort; cohort-level exposure and claims data remain a prerequisite.
- The eligibility assessment is judgemental and can be challenged by auditors where the sample is narrow.
- Loss components under PAA are volatile because they are triggered by observed indicators rather than a continuous measurement.
- Comparability with general-model peers is limited even within the same market segment.
Conclusion
The premium allocation approach is a simplification of one component, not of IFRS 17. Firms that plan for full measurement of incurred claims and disciplined onerous testing implement it smoothly.
Its real benefit is speed of close; its real value to management is the early loss signal it produces when the testing is taken seriously.
References
- IFRS 17 Insurance Contracts — International Accounting Standards Board, 2020
- IFRS 17 implementation resources — IFRS Foundation
Author bio
Jonas Mohamed Osman Abdelghafour, known as Yonas Osman is an actuary, FRM and financial risk professional specialising in banking, insurance, model risk, capital modelling and quantitative risk management.