Let’s be clear about this: IFRS 17 didn’t just change accounting rules. It rewired how we see insurance profitability. And at the center of that shift sits the PAA — a metric so sensitive that a 0.5% change in claims inflation assumption can flip a surplus into a deficit. You don’t need to be a CFO to see why that matters.
Understanding the PAA: How It Fits Into the IFRS 17 Framework
IFRS 17 replaced decades of fragmented insurance accounting with a single, principles-based model. At its core? The measurement of insurance liabilities using current estimates, market-consistent discounting, and explicit recognition of profit over time. The PAA isn’t a standalone figure — it’s one component of the fulfilment cash flows, which also include future premiums, expenses, and the risk adjustment.
Think of it like this: if an insurer sells a 10-year auto policy today, it can’t just book the premium as profit. It has to model what it expects to pay in claims over the next decade, adjust for past patterns, and discount that back. That’s the PAA. It’s not about what was paid last year. It’s about what last year tells us about next year — and the six after that.
What Does “Adjusted” Mean in the PAA?
The “adjusted” in PAA is where it gets tricky. Historical claims data is raw. It includes outliers — a hurricane in Florida, a spike in medical costs in Germany, a fraud ring in Ontario. Insurers must smooth these while preserving real trends. They adjust for inflation (healthcare costs rose 6.3% in the U.S. in 2023), operational changes (a new claims processing system reduced average settlement time by 18 days), and mix shifts (e.g., selling more high-risk policies).
And that’s exactly where subjectivity creeps in. Two actuaries looking at the same dataset might produce PAA figures differing by 4–7%, simply because one assumes a permanent shift in claim severity, while the other sees a one-off anomaly. This isn’t error — it’s judgment. And under IFRS 17, judgment is visible on the balance sheet.
Discounting: Why Timing Is Everything in the PAA
The PAA is a present value — so timing matters as much as amount. A $1 million claim expected in year one is worth more than the same claim in year five. Discount rates must reflect the characteristics of the liability. For non-life insurance, that often means short-term risk-free rates plus a liquidity premium — say, 3.8% in euro-denominated contracts as of June 2024.
But here’s the catch: if interest rates drop, the present value of future claims rises. That means higher liabilities — even if claim frequency stays flat. A 100-basis-point fall in discount rates can inflate PAA values by 12–15% for long-tail lines like liability insurance. That changes everything for capital planning.
The Technical Mechanics: How Insurers Build the PAA Step by Step
Building the PAA isn’t a spreadsheet hack. It’s a multi-layered process blending data science, actuarial modeling, and governance. Most insurers use a bottom-up approach: starting with policy-level claims history, aggregating by cohort, then applying adjustments. The goal? A forward-looking, unbiased estimate — not a rearview mirror.
You might assume past claims predict the future perfectly. We’re far from it. Medical inflation has outpaced general inflation for 14 consecutive years. Cyber claims are doubling every 18 months. Climate change is making property losses less predictable. The PAA must reflect these shifts — or it’s fiction.
Data Aggregation and Cohort Design
Insurers group policies into cohorts — contracts with similar risk profiles and inception dates. A cohort could be all U.S. homeowners’ policies issued in Q2 2022, or European motor policies with deductibles over €1,000. Why? Because claims patterns evolve over time.
Within each cohort, historical claims are pulled — not just paid amounts, but incurred but not reported (IBNR) estimates. These are then trended forward. If bodily injury claims rose 5.7% annually over the past five years, that trend may be projected — but only if the insurer can justify its sustainability. And that’s where internal audit teams start asking questions.
Adjustment Methodologies: From Trend Factors to Structural Shifts
Adjustments aren’t mechanical. They’re strategic. A common method is the chain-ladder technique, but even that requires assumptions about tail factors and volatility. For longer-term lines, actuaries use frequency-severity models: projecting how often claims occur and how much they’ll cost per incident.
Because the future isn’t linear. A 2021 study by the Geneva Association found that inflation in repair costs post-pandemic added 8.2% to auto claims in developed markets — a jump no model had predicted. So insurers now embed scenario buffers. Some add a 1–2% structural uplift for “emerging risk drag.” It’s not precise. But it’s prudent.
PAA vs. Other IFRS 17 Components: Where the Confusion Lies
The PAA doesn’t stand alone. It feeds into the fulfilment cash flows, alongside future premiums, expense assumptions, and the risk adjustment. Yet people don’t think about this enough: the PAA is the largest component in most non-life contracts — often 70–80% of total liability.
Compare that to life insurance, where future premiums dominate. In motor or health lines, claims are the beast. And that’s why PAA accuracy makes or breaks earnings stability.
PAA vs. Risk Adjustment: Two Sides of Uncertainty
The PAA reflects the “best estimate” of claims. The risk adjustment adds a buffer for non-financial risk — uncertainty around that estimate. If the PAA says “we expect $100 million in claims,” the risk adjustment might add $8–12 million to cover the chance it’s wrong.
But here’s a question few ask: if the PAA already includes prudent trending, isn’t the risk adjustment double-counting caution? Some auditors think so. Others argue they serve different purposes — one for expected value, one for risk capital. Experts disagree. Honestly, it is unclear how consistent this is across jurisdictions.
PAA and the Contractual Service Margin (CSM)
The Contractual Service Margin is profit deferred and recognized over time. It’s calculated as the difference between premiums and the total fulfilment cash flows (including PAA) at contract inception. If the PAA is too low, the CSM looks artificially high — and future profits get inflated.
And then reality hits. When actual claims come in above expectations, the CSM gets reduced. That means less profit recognition later. That’s why underestimating the PAA today can trigger profit crashes tomorrow. One European insurer saw its CSM erode by €210 million in 2023 due to underestimated medical trend assumptions. Suffice to say, their actuarial team had a rough quarter.
Frequently Asked Questions
Can the PAA Be Negative?
No. The PAA represents expected outflows — future claims — so it’s always a positive number in the liability calculation. But the overall fulfilment cash flows can be negative if premiums exceed expected costs, which increases the CSM. That said, a negative net liability doesn’t mean no risk — just deferred profit recognition.
How Often Is the PAA Updated?
Every reporting period. IFRS 17 requires a current measurement model. So insurers reassess the PAA quarterly, incorporating new claims data, updated trends, and revised assumptions. A spike in storm claims in Q4? It shows up in the PAA by December 31. No lag, no smoothing over years — real-time accountability.
Do All Insurers Calculate PAA the Same Way?
They don’t. While the standard sets principles, methods vary. Some use gross claims data, others net of reinsurance. Some adjust for large losses, others include them. This lack of uniformity is a known issue. Regulators are watching. The problem is, prescribing a single method might kill relevance. The issue remains: comparability across firms is weaker than many hoped.
The Bottom Line: Why the PAA Is the Hidden Power Player in IFRS 17
I find this overrated: the idea that IFRS 17’s biggest challenge is discount rate selection. It’s not. It’s the PAA — because it combines data complexity, actuarial judgment, and earnings sensitivity in a way no other component does. Get it wrong, and your balance sheet wobbles. Get it right, and you gain credibility with investors who now see through the old smoothing tricks.
My recommendation? Insurers should publish PAA sensitivity disclosures — not just in footnotes, but in earnings calls. Show how a 1% change in claim trend impacts profit. Transparency builds trust. And in a world where a single assumption can move stock prices, that’s not just good accounting. It’s survival.
To give a sense of scale: one global reinsurer found that improving its PAA modeling reduced earnings volatility by 27% over three reporting periods. That’s not a minor win. It’s a strategic upgrade. Because while IFRS 17 is a compliance exercise for some, for others — it’s a lever. And the PAA? It’s the fulcrum.