Why Are We Implementing IFRS 17? A Practitioner Revisits the Case

Halfway through an implementation, somebody in the steering committee always asks it out loud: why are we doing this again? Usually it is a CFO watching the consulting bill, or an actuary on their third weekend in a row. The answer deserves more than “because the IASB said so.” It also deserves more than the marketing gloss on the front page of the Effects Analysis. What follows is how I explain it when asked — what the standard actually fixes, what it costs to fix those things, and which parts of the construction I find genuinely elegant after having built against it on four continents.

The shortest version: IFRS 4 was always an interim standard. It was issued in 2004 to force enhanced disclosure while the IASB took longer to land on a comprehensive model. IFRS 17 is the comprehensive model. Implementing it completes a project the IASB started over two decades ago.

1. What IFRS 4 left unresolved

IFRS 4 permitted an enormous range of accounting policies for insurance contracts. Two insurers writing the same product in the same jurisdiction could measure the liability differently, recognise profit on different patterns, and present revenue on non-comparable bases. Analysts covering the sector routinely rebuilt embedded-value walks by hand before they could compare two companies. The non-GAAP layer sat on top of GAAP because GAAP was not doing its job for insurance contracts.

That is the gap IFRS 17 closes. The standard is not a tweak — it is the first IFRS insurance model that tells you how to measure, how to present, and how to disclose, with enough specificity that two insurers writing identical business should produce comparable numbers.

2. The measurement improvements that actually matter

Inside the standard there is a list of measurement improvements. Not all of them land equally. The ones that matter most in practice:

Updated assumptions, always. Under IFRS 4 it was possible to carry locked-in assumptions through a contract’s life even as experience diverged. IFRS 17 requires best-estimate cash flows remeasured every reporting period. This alone changes how actuarial teams are staffed and how model-run cadence is designed.

Discount rates that reflect the liability. Rates now reflect the characteristics of the insurance cash flows, not the expected return on backing assets. It sounds technical. In practice it is the single biggest reason asset-liability matching strategies had to be rewritten for IFRS 17 — the old linkage between asset yield and liability discount is gone.

Explicit risk adjustment. Risk margins existed under IFRS 4, but they were rarely comparable between insurers because the calibration basis was not explicit. IFRS 17 forces the insurer to disclose the confidence level behind the risk adjustment. That makes it comparable in a way it never was before.

Options and guarantees at market prices. Embedded financial options and guarantees are measured consistently with observable market prices where those exist. A number of pre-IFRS 17 models did not measure these explicitly at all. The gap closed here is large for with-profit and variable annuity books.

3. The presentation changes that drive behaviour

The measurement changes are not the only thing that moves behaviour. The presentation changes do too.

Revenue is now earned, not cash-received. Investment components — amounts the insurer will repay regardless of whether the insured event happens — are stripped out of revenue. That deflates reported top-line for contracts with significant savings features, sometimes dramatically. The number is more honest, but it is also a smaller number than the industry was used to reporting.

Insurance service expenses only include items that genuinely reflect service delivery. Deposit repayments move out. Changes in liability from new business or methodology changes move out. What remains in the line is a cleaner measure of the cost of providing insurance.

The contractual service margin absorbs unearned profit and releases it over coverage. Profit emergence is now matched to service delivery rather than to cash. For long-duration contracts this is a structural change in reported earnings pattern.

4. Where comparability improves — and where it does not

The marketing line is that IFRS 17 improves comparability across companies, across contracts, and across industries. That is broadly true, but the mechanism is worth being precise about.

Across companies: the biggest gains are inside multinationals whose subsidiaries used to consolidate to different insurance accounting policies. Under IFRS 17 they must apply one policy. That is real.

Across contracts: within a single insurer, groups of contracts that behave differently are now measured differently in ways investors can see. Onerous contracts are identified at initial recognition and visible in disclosure. Under IFRS 4 an onerous portfolio could hide inside a profitable one for years.

Across industries: revenue recognition now looks more like the rest of IFRS. Insurance contracts are not quite like services or goods, but the principle — revenue earned in exchange for service — is the same principle applied elsewhere. For analysts that reduces the adjustment needed to compare insurers against asset managers or banks.

The place comparability does not fully improve is in the transition measurement. Full retrospective, modified retrospective, and fair value transition approaches produce materially different opening CSMs on long-duration books. Two insurers writing identical business will still differ on day-one equity depending on transition choice. That is a limitation of the standard, not a feature.

5. What I actually find elegant

Most of what gets written about IFRS 17 focuses on the cost and the disruption. It is worth saying out loud which parts of the standard I think are genuinely well-designed.

The CSM mechanism is elegant. Unearned profit sits as an explicit liability and is released on a coverage-unit basis. It is internally consistent, it prevents front-loading of profit, and it produces a clean reconciliation between profit recognised and service delivered.

The B120 check is elegant. Revenue for a group of insurance contracts should equal premiums received, adjusted for a financing effect and net of investment components. That equation lets you validate the top-line of your disclosures against the most primitive thing in an insurance contract — the premium paid. If the equation does not hold, something in your build is wrong. It is rare for an accounting standard to give you a closed-form validation check that crisp.

The locked-in versus current discount-rate split is elegant. Movements that reflect passage of time unlock the CSM at locked-in rates. Movements that reflect financial conditions flow through insurance finance expense at current rates. The split is what makes the profit pattern stable against rate movements, and it is the bit of the standard I had the most respect for after I first coded against it.

6. Is it worth the cost?

The cost is real. Most mid-sized insurers I have worked with spent multiples of their initial IFRS 17 budget over the life of the programme. The actuarial model rebuild, the accounting-engine licence, the data layer, the reporting cycle redesign, and the audit cost in the first two years after go-live add up quickly. Nobody in the industry disputes this.

What insurers who are now post-implementation will say, usually quietly, is that the new numbers are better numbers. Profit pattern is more stable. The sources of movement are more explainable. Disclosures that used to take weeks of narrative to defend are now anchored to a measurement framework with one set of principles. None of that shows up as a line item on the implementation invoice, and it is the hardest thing to commit to in the business case. But it is the thing that remains after the consultants have gone home.

Closing thought

If the question is whether IFRS 17 was worth implementing in the abstract, the answer is that it was always going to be implemented. The standard exists because IFRS 4 was an interim solution and the insurance industry eventually had to have proper accounting. The more useful question, from inside a live build, is which of the improvements matter most for the business that is implementing it — and which parts of the construction can be genuinely relied on to produce better numbers once the dust settles. The measurement improvements, the CSM mechanism, and the B120 check are three I would pick out. They are worth the trouble.