Full explanation
Overview
An insurer collects a premium today for a promise that might not pay out for thirty years. IFRS seventeen says you cannot book that premium as revenue on day one.
IFRS seventeen is insurance contracts — the culmination of roughly two decades of IASB work, and the most complex standard on this channel. Insurance breaks the normal revenue model completely: the sale and the service are separated by decades, and the ultimate cost isn't even known at the point of sale.
Two building blocks
The general measurement model has two building blocks. Fulfilment cash flows: estimated future cash flows, discounted, plus a risk adjustment for non-financial risk. And the contractual service margin — the CSM — the day-one unearned profit, held back rather than recognised immediately.
The CSM absorbs
The single most-tested idea in this entire standard: the CSM absorbs change. Favourable or unfavourable revisions to future cash-flow estimates adjust the CSM first. They do NOT hit profit or loss immediately — almost the opposite of every other standard on this channel.
There's exactly one place the CSM can't absorb a loss: once it would go negative, it's floored at zero, and whatever's left over is recognised in profit or loss immediately, on day one. That's an onerous contract.
The PAA
For contracts of roughly twelve months or less, the premium allocation approach — the PAA — is permitted: a simplified model close to ordinary deferred revenue. Most general and non-life insurance uses this rather than the full general model.
The VFA
For contracts with direct participation features — unit-linked products, with-profits policies — the variable fee approach applies. The CSM absorbs the entity's share of changes in the underlying investments too, not just cash-flow revisions.
Let's build one cohort of contracts from day one, through a real change in estimate, and then see exactly what happens when that same shock hits a contract with no CSM cushion left.
Pricing the cohort
Meridian Insurance prices a cohort: expected premiums, discounted, five million euros. Expected claims and expenses, discounted, four point two million. Risk adjustment for non-financial risk: three hundred thousand. Net fulfilment cash flows: negative five hundred thousand — a net inflow position.
It becomes CSM
That five hundred thousand euros doesn't hit profit on day one. It becomes the CSM instead — the expected profit, held back, to be released as Meridian actually delivers the coverage over the life of the contracts.
Absorbed, then released
During year one, claims estimates worsen by one hundred fifty thousand euros. That doesn't hit profit either — it reduces the CSM first, down to three hundred fifty thousand. Over a five-year coverage period, evenly spread, that's a seventy thousand euro release to profit this year, leaving a closing CSM of two hundred eighty thousand.
Now the contrast. A second cohort, Cohort B, has only two hundred thousand euros of CSM left. The same kind of unfavourable shock hits it — three hundred fifty thousand euros this time. The CSM absorbs two hundred thousand, floors at zero, and the remaining one hundred fifty thousand euros hits profit or loss immediately.
Two outcomes
Same size shock, two completely different outcomes — purely because of how much CSM cushion each cohort had left. That's why the CSM balance matters so much more than current-period profit alone.
Three new lines
Presentation changes completely too. Insurance revenue replaces premiums received as the top-line number. Insurance service expenses sit separately. Insurance finance income and expense splits the underwriting result cleanly from the investment result — three genuinely different lines, not one blended number.
A confidence level
The risk adjustment itself is disclosed with an equivalent confidence level — so a reader can compare how conservative one insurer's risk appetite is against another's, not just trust a single unexplained number.
Three routes
Transition offered three routes: full retrospective, modified retrospective, or the fair-value approach. In practice, most insurers used fair value — the historical data to do a genuine full retrospective calculation simply didn't exist.
Disclosure requirements are extensive: a full reconciliation of insurance contract balances from opening to closing, including exactly what moved the CSM and why, the risk adjustment's confidence level, and the significant judgements behind every fulfilment cash flow estimate.
A material DTL
A practical consideration most people miss: deferred tax. Most tax regimes tax premiums and claims on a completely different basis than IFRS seventeen's model. Meridian's cohort might carry an IFRS seventeen liability of twelve million euros against a tax base of ten point five million — a one point five million euro temporary difference, a three hundred seventy-five thousand euro deferred tax liability at twenty-five percent. Routinely one of the largest deferred tax lines an insurer reports.
Twenty years
Why this actually mattered: before IFRS seventeen, insurers reported under wildly different local GAAPs, making genuine cross-border comparison close to meaningless. Implementation cost the global industry billions of dollars and roughly two decades of IASB development to land.
Dated, real
Amendment history, dated: the June 2020 amendments deferred the effective date to the first of January twenty twenty-three, excluded credit card contracts from scope, allowed recognition of insurance acquisition cash flows for expected renewals, permitted interim-period relief, clarified CSM allocation, added a risk-mitigation option, and eased transition. Mandatory since 2023.
Similar, not the same
A real regulatory law sits alongside all of this, easy to confuse with IFRS seventeen: Solvency II, the EU's prudential capital regime for insurers. It looks similar on the surface — both use a best-estimate-cash-flows-plus-margin structure — but they are NOT the same numbers. EIOPA's own first-year comparison found IFRS seventeen life liabilities, EXCLUDING the CSM, running about two and a half percent lower than Solvency II's — include the CSM, and IFRS seventeen life liabilities actually run higher instead. Non-life liabilities, where the CSM barely features, came in about nine and a half percent higher. The risk adjustment and the Solvency II risk margin diverge even more. And the CSM itself has no Solvency II equivalent at all — Solvency II is a pure capital-adequacy view, with no concept of deferring unearned profit.
Annual cohorts
A genuine operational headache, not just a technical rule: contracts have to be grouped into ANNUAL cohorts, and split further by profitability — onerous, no significant risk of becoming onerous, and everything else. No averaging across issue years allowed. Insurers had to rebuild core systems specifically to track this level of granularity — a real reason implementation cost billions, not just an accounting footnote.
Still ahead
Worth knowing, live and dated: as of this recording, the IASB hasn't opened a post-implementation review of IFRS seventeen at all — it's currently reviewing IFRS sixteen and IFRS nine instead. IFRS seventeen has been mandatory since 2023, and its own formal review is still ahead, not behind.
Outside the carve-out
IFRS eighteen implications: effective twenty twenty-seven, it changes how insurers present results OUTSIDE the insurance-specific lines — operating, investing and financing categorisation still applies to everything IFRS seventeen doesn't already carve out. Twenty twenty-six is the comparative year for calendar year-end insurers.
For an investor, the CSM balance is effectively unrecognised future profit sitting off the current period's numbers. Its size, and how it moves year to year, tells you more about an insurer's real earnings trajectory than this year's profit figure on its own.
Why no hit?
Question: why doesn't a bad year immediately hit profit for an insurer under IFRS seventeen? Because cash-flow revisions absorb into the CSM first and release gradually over the remaining coverage period — unless the CSM is already at zero, in which case the loss hits profit immediately instead.
Car vs life?
Question: is IFRS seventeen the same for a car insurer and a life insurer? No. Most short-duration general insurance uses the simplified premium allocation approach, while long-duration life and unit-linked products use the full general model or the variable fee approach.
Next up: IFRS nineteen, subsidiaries without public accountability — why a subsidiary that already reports everything to its parent shouldn't have to produce the full public-investor disclosure package on top of that.