Three years after IFRS 17 took effect on January 1, 2023, KPMG's review of 44 insurers' 2025 annual financial statements found that most non-life insurers now publish an IFRS 17 combined ratio, and that the calculations differ enough across companies that direct peer comparison remains unreliable.
The figure behind that is not the 44. It is that the components differ: which items enter the ratio, whether discounting flows through it, and how acquisition costs are allocated.
Key Takeaways
- 44 insurers reviewed by KPMG for 2025 reporting mostly disclose an IFRS 17 combined ratio, but the ratios are built differently enough that the label carries less information than it did under IFRS 4.
- 4% to 5% versus 27% is the risk adjustment as a share of insurance liabilities for Canadian direct writers against reinsurers, an actuarially calibrated figure that is now a leading source of cross-company variation.
- Eight standardized KPIs proposed by the Insurance Bureau of Canada all use insurance revenue as the denominator, which is the specific fix for the comparability problem rather than a general call for consistency.
- 0% to 1% of total insurance liabilities is the loss component for most Canadian direct writers, against roughly 5% for reinsurers, a line item with no pre-IFRS 17 analog at all.
- January 2027 brings IFRS 18 presentation requirements, which will recategorize results before the first round of IFRS 17 metric calibration has settled.
Why the Combined Ratio Stopped Comparing
The standard changed both halves of the ratio at once, which is why the disagreement is structural rather than a matter of preparers being slow.
The denominator moved from earned premiums to insurance revenue, recognized as performance obligations are fulfilled rather than as premium is collected. For an annual P&C policy the practical difference is modest. For long-duration life contracts it is large, because upfront premium is no longer revenue at inception.
The numerator moved too. IFRS 17 merges benefits, claims and expenses into a single "insurance service expenses" line, so isolating a loss ratio from an expense ratio takes decomposition work that the statement does not do for you. On top of that, contract liabilities are measured at present value, so some insurers publish a discounted combined ratio, some an undiscounted one, and some both.
The result is a metric that requires footnote reading before any peer comparison. As one senior Canadian insurance executive put it to Canadian Underwriter: "Our standard, which is supposed to make us all comparable, has made us less comparable than we've ever been in the industry."
The Actuarial Inputs Are Now the Variance
Under legacy standards, assumption changes flowed into reserves with limited disclosure of individual effects. IFRS 17 requires explicit disclosure of how changes in estimates of future cash flows hit the CSM, the loss component and profit or loss, which puts actuarial judgment directly into the comparability problem.
The risk adjustment is the clearest case. The Actuary Magazine reported it at roughly 4% to 5% of insurance liabilities for Canadian direct writers in 2024 and 27% for reinsurers. Part of that spread is genuine portfolio composition. Part is calibration choice, and there is no external anchor that separates the two from outside the company. The same applies to discount rate methodology and to how the loss component is recognized, which ran 0% to 1% of liabilities for most direct writers against about 5% for reinsurers.
Canada shows what a fix looks like because it built one. The Insurance Bureau of Canada's discussion paper proposed eight standardized KPIs, led by the Comprehensive Combined Ratio as the closest proxy to the IFRS 4 combined ratio, with every one of the eight using insurance revenue as its denominator. The MSA Research committee's Net Insurance Service Ratio does the same work at the insurance service result level, and its 2024 benchmarks read cleanly: 85% for the Big Three, 89% for other direct writers, 99% for reinsurers.
Canada's advantage is structural rather than cultural. It mandates IFRS for publicly accountable enterprises and uses IFRS 17 for statutory reporting, so its insurers produce these ratios from one set of templates. Markets running dual-basis reconciliation cannot copy the approach without also copying the reporting architecture underneath it.
What the CSM Absorbs, and When It Stops
The Contractual Service Margin is IFRS 17's genuine innovation, and it is also the reason a smooth earnings series proves less than it appears to.
When assumptions about future cash flows change on profitable contracts, the change is absorbed into the CSM rather than flowing to profit or loss. That is deliberate, and it means the income statement does not always report the period in which the economic event happened. The Footnotes Analyst identifies the mirror case as well: accounting results can move while the underlying economics are stable, because measurement mechanics produce fluctuations that reflect recalibration rather than a change in risk.
The consequence for reading a set of accounts is that neither smoothness nor volatility is evidence on its own. Absorption continues until a group of contracts becomes onerous, at which point the loss component recognizes immediately, so deterioration arrives as a step rather than a slope. An insurer with a large mature CSM from legacy business can report a strong insurance service result while current-year underwriting is weak, because release from prior vintages dominates the number.
None of that is hidden. It is disclosed at the group-of-contracts level, tracked in annual cohorts, and unlocked quarterly, which is exactly why the data infrastructure demands have grown as much as they have. The interpretive burden simply moved from the preparer to the reader, and the reader has fewer of the inputs.
The window for settling this is also about to narrow. IFRS 18 takes effect in January 2027 and will require results to be categorized into operating, investing and financing activities. Insurers still calibrating IFRS 17 metrics will be recalibrating presentation on top of them, with the first exercise unfinished.
Further Reading on actuary.info
- LDTI Year Three: Earnings Volatility Lessons for Life Actuaries
- IFRS 17's CSM Release Ratio Emerges as Life Insurer Report Card: a deeper look at how the CSM release ratio specifically is displacing embedded value as the metric analysts probe first.
- IFRS 17 Implementation 2026: The $20 Billion Accounting Revolution
- IFRS 18 Arrives in 2027: Why Insurers Need 2026 Comparatives Now
- AG 55 First Filing: What Life Actuaries Learned
- Life Insurance and Annuity Market Trends 2026
Sources
- KPMG, “Insurers’ 2025 Annual Financial Statements: Real-Time IFRS 17” (April 2026)
- The Footnotes Analyst, “IFRS 17 Insurance: Economic Versus Accounting Volatility” (September 2025)
- Jesse Resnick, FCIA, FSA, “IFRS 17 KPIs: Interpreting Insurance Financial Results”, The Actuary Magazine (April 2026)
- EY, “How Insurers Should Communicate IFRS 17 KPIs”
- EY, “IFRS 17 Challenges and Opportunities”
- Insurance Bureau of Canada, “New IFRS 17 Metrics Discussion Paper” (June 2025)
- SOA Investment and Risk Management Section, “Managing CSM Under IFRS 17” (March 2026)
- Actuaries Institute of Australia, “IFRS 17 Post Implementation 2025 Survey Report” (February 2026)
- IBIMA, “Transforming Insurance Reporting: Challenges of IFRS 17 Implementation” (2026)
- ICAEW, “IFRS 17 Insurance Contracts Standards Tracker”
- IFRS Foundation, “IFRS 17 Fact Sheet”
- PwC UK, “IFRS 17 Key Performance Indicators”
- Canadian Underwriter, “Why Insurers’ Combined Ratios Under IFRS 17 Can’t Be Compared” (April 2024)