Internal:Training/IFRS17/The premium allocation approach
🔗 Recall. In the previous page, you learned how the income statement under IFRS 17 presents insurance revenue based on service delivered, not premiums collected, along with insurance service expenses and insurance finance income or expense. Now we turn to a simpler way of reaching those same results: the premium allocation approach, a shortcut available for many of the contracts you will encounter in practice.
🎯 Objective. In this page, you will learn:
- The conditions a group of insurance contracts must meet before the insurer may use the premium allocation approach instead of the general model.
- How the PAA measures the liability for remaining coverage and the liability for incurred claims, and why it is considered a simplified method.
- Which elements of the general model the PAA keeps, which it skips, and the practical consequences of those differences.
When PAA is available: the eligibility test
🚪 A lighter path, not a free pass. The general model you studied in the previous pages is the default measurement approach under IFRS 17. It requires the insurer to calculate fulfilment cash flows, discount them, add a risk adjustment, and track a contractual service margin that evolves over time. For contracts that last many years, such as life insurance or long-tail liability covers, that rigour is essential. But many contracts, especially in property and motor insurance, run for just one year. For those shorter contracts, IFRS 17 offers a simpler route called the premium allocation approach, or PAA. Think of it as a shortcut through familiar territory: you arrive at broadly the same destination but with far fewer calculations along the way.
📏 The eligibility test. An insurer may use the PAA for a group of contracts only if it can demonstrate, at initial recognition, that the PAA liability would not differ materially from the one produced by the general model. In practice, this condition is automatically satisfied when the coverage period of every contract in the group is one year or less. A standard 12-month motor policy in Germany or a one-year home insurance contract in Belgium passes this test without further analysis. If the coverage period exceeds one year, the insurer must perform a more detailed assessment to prove the results would be sufficiently close. Most non-life books at an insurer like AXA consist of annual contracts, which is why the PAA is so widely used across the industry.
⚠️ Common misconception. Some learners assume the PAA is available simply because a contract looks "simple" or has a low premium. Simplicity of product design is not the criterion. The test is about whether the measurement result would materially differ from the general model. A complex product with a short coverage period can qualify, while a straightforward product with a five-year term may not.
🤔 Think about it. If the PAA is a shortcut, what exactly does it simplify? How does the insurer measure the liability without tracking a CSM that absorbs changes in estimates at every reporting date?
How PAA works: simplified measurement
📦 Starting with the liability for remaining coverage. Under the PAA, the insurer does not calculate all four building blocks at initial recognition. Instead, it starts by recording the liability for remaining coverage (LRC) at the amount of premiums received, minus any acquisition costs already paid. As time passes, the insurer reduces this LRC in proportion to the coverage provided, which directly becomes insurance revenue in the income statement. Imagine AXA writes a group of annual home insurance contracts in Spain, collecting €6,000,000 in premiums on 1 January. After deducting €300,000 of acquisition costs, the LRC starts at €5,700,000. Each quarter, roughly one quarter of that balance is released as revenue, reflecting the coverage delivered during those three months.
🔓 No CSM to track. This is the heart of the simplification. Under the general model, the insurer must maintain a contractual service margin that absorbs favourable and unfavourable changes in estimates of future cash flows, then release it through coverage units. Under the PAA, that entire mechanism disappears for the liability for remaining coverage. The premium itself serves as a proxy for the expected value of service to be delivered. Because the coverage period is short, the difference between tracking a full CSM and simply running down the premium over time is immaterial. This saves significant actuarial and accounting effort at every reporting period.
⚠️ Common misconception. Learners sometimes believe the PAA eliminates the need to think about claims altogether. It does not. The PAA simplifies the liability for remaining coverage, but once a claim event occurs, the insurer must still measure the liability for incurred claims (LIC) using the full fulfilment cash flow approach: probability-weighted estimates of future payments, discounted to present value, plus a risk adjustment. The simplification applies to the "promise still outstanding" side, not the "claims already happened" side.
🏥 Handling incurred claims. When claims are reported or estimated during the coverage period, the insurer sets up a liability for incurred claims. For a group of Spanish home insurance contracts, suppose a severe hailstorm in March generates claims with an estimated present value of €800,000 and a risk adjustment of €50,000. The insurer records a LIC of €850,000. This part of the measurement is identical to the general model: the insurer uses the same fulfilment cash flow techniques to value obligations that have already crystallised.
🤔 Think about it. If the PAA removes the CSM and simplifies the LRC, does the insurer also get to skip discounting and the risk adjustment on the remaining coverage side? Or are those still required?
What you keep and what you skip vs. the general model
✂️ What the PAA removes. The most significant simplification is the removal of the CSM for the liability for remaining coverage. Under the general model, every change in estimated future cash flows that relates to future service must be routed through the CSM, creating a complex adjustment at each reporting date. The PAA sidesteps this entirely. The insurer also does not need to project and discount the fulfilment cash flows for remaining coverage at each reporting date, because the premium-based LRC serves as a sufficient approximation. Additionally, the insurer is not required to calculate a risk adjustment on the LRC, though it must still do so for the liability for incurred claims. The net effect is a dramatic reduction in the volume of calculations and actuarial judgments needed at each period end.
📋 What the PAA keeps. Despite the simplifications, several important requirements remain. The insurer must still apply the onerous contract test. If at any point during the coverage period the group is expected to generate a loss, the insurer recognises that loss immediately, just as under the general model. The liability for incurred claims continues to require full fulfilment cash flow measurement, including discounting and a risk adjustment. The insurer must also still follow IFRS 17's grouping requirements: portfolios, profitability groups, and annual cohorts apply in exactly the same way.
⚠️ Common misconception. A frequent misunderstanding is that the PAA is a completely different accounting model. It is not. The PAA is best understood as a practical simplification of the general model for the remaining coverage portion only. Everything on the incurred claims side, plus the onerous contract test and grouping rules, works the same way. If you master the general model, you already understand most of the PAA; you simply set aside a few steps that do not materially affect the result for short-duration contracts.
🔍 A practical comparison. Consider a group of one-year motor insurance contracts in Germany. Under the general model, the insurer would project all cash flows over the remaining months, discount them, compute a risk adjustment, derive a CSM, and release it through coverage units each quarter. Under the PAA, the insurer simply starts with the premiums received, deducts acquisition costs, and recognises revenue proportionally over the year. When a claim occurs, it measures the LIC with full rigour. The income statement result is materially the same, but the operational cost of producing it is far lower. For non-life portfolios dominated by annual contracts, this efficiency gain is substantial.
Takeaways
📌 Key takeaways.
- The PAA is available when the coverage period is one year or less, or when the insurer can demonstrate its results would not materially differ from the general model.
- The PAA simplifies the liability for remaining coverage by using premiums received (less acquisition costs) as the starting point and releasing that balance as revenue over the coverage period, eliminating the need to track a CSM.
- The liability for incurred claims is still measured using full fulfilment cash flows, discounting, and a risk adjustment, and the onerous contract test and grouping rules continue to apply.