We’re far from the days when reserves were just actuarial estimates tucked away in footnotes. Today, financial statements reflect real-time economic reality. The GMM is at the core of this transformation. But understanding it isn’t easy—especially when the standard reads like a legal thriller written by actuaries.
How Does the GMM Work? A Breakdown for Non-Actuaries
The GMM is the default framework for measuring insurance contracts under IFRS 17. It applies to all contracts unless they qualify for the Variable Fee Approach (VFA) or the Premium Allocation Approach (PAA). Think of it as the “full treatment” model—comprehensive, detailed, and, honestly, a bit exhausting.
You start with an estimate of future cash flows: claims, benefits, premiums, expenses. These aren’t static numbers—they’re probability-weighted, adjusted for inflation, taxes, and even behavioral changes. Then you discount them using current rates—no more hiding behind long-term assumptions. That gives you the present value of fulfillment cash flows. But wait: that’s only part of the picture.
Then comes the contractual service margin (CSM). This is your profit buffer. It’s unlocked gradually as services are provided over time. You don’t recognize all profit upfront—no more front-loading. Instead, you release it as you deliver coverage. The CSM adjusts for changes in estimates, but with a twist: favorable changes go straight to profit; unfavorable ones hit the balance sheet immediately. That asymmetry? That’s by design.
What’s Included in the Fulfillment Cash Flows?
Let’s get technical. The fulfillment cash flows cover everything tied to fulfilling the contract. Claims payments—obviously. But also administrative costs, investment expenses, reinsurance recoveries, and even taxes linked directly to the contract. Not all taxes—just the ones directly attributable.
These cash flows are probability-weighted. That means you don’t take best-case or worst-case scenarios. You run simulations—thousands of them—factoring in lapse rates, mortality, morbidity, inflation, and even pandemic risks. If you’re pricing a life annuity today, you’re modeling outcomes over 30 years. And that’s where the uncertainty grows.
How Is Discounting Handled in the GMM?
Here’s where IFRS 17 departs from the past. Discount rates must reflect current market yields, adjusted for the timing and currency of cash flows. No more “lock-in” rates from inception. If rates drop from 4% to 2.5%, your liabilities go up—even if the policy was written years ago.
But—and this is important—you don’t just use government bond yields. You include a “top-up” for illiquidity if the cash flows don’t match available instruments. Say your liabilities peak in 15 years, but the longest liquid bond is 10 years. You extrapolate. The top-up can’t exceed 10 basis points per year beyond liquid maturity. (Yes, that’s specific. And yes, people argue about it.)
The Contractual Service Margin: Profit Recognition Without the Games
The CSM is the GMM’s engine for profit recognition. It starts as the difference between the present value of future cash flows and the premiums received. If you collect $100 in premiums and expect to pay out $90 in benefits and costs (discounted), your initial CSM is $10. Simple, right?
But it’s never that simple. The CSM adjusts for changes in estimates. If claims turn out lower than expected, you increase the CSM and recognize more profit over time. If they go up? You reduce the CSM—or even create a loss if it goes negative. No deferral. That’s a big shift from IFRS 4, where companies could smooth losses over years.
And here’s the kicker: the CSM is amortized over the coverage period. Not evenly. Not linearly. Based on the pattern of coverage. A car insurance policy? Most risk is in the first year—so most profit is recognized early. A 20-year term life policy? Risk spreads out. So does profit.
But what if you revise your estimate of future cash flows? Say, after a hurricane season worse than expected. You adjust the CSM. Favorable changes increase it; unfavorable ones decrease it immediately. That asymmetry prevents companies from building hidden reserves during good years to cover bad ones later. We’re far from it now.
How Is the CSM Adjusted for Changes in Estimates?
Changes in estimates are split into two buckets: those that affect future service and those that affect past service. Future changes go through the CSM. Past changes? They hit profit or loss directly. But how do you distinguish between them?
Take a health insurance policy where medical costs rise unexpectedly. If the increase affects future claims, it reduces the CSM. If it covers claims already incurred but not yet paid (a timing issue), it’s recognized immediately. The line is thin. Actuaries spend hours debating it. And that’s exactly where disagreements arise.
Can the CSM Go Negative?
Yes. And when it does, it becomes an onerous contract liability. No more hiding losses in a deferred profit pool. If your revised estimates show you’ll lose money on a block of policies, you recognize it now. Full stop.
Some insurers hate this. They argue it creates volatility. But I am convinced that transparency beats smoothing. Investors can make better decisions when they see the real economics. Volatility isn’t the enemy—misleading accounting is.
GMM vs PAA: When to Use Which Model?
Not every contract needs the full GMM treatment. Short-duration contracts—like motor or home insurance—often qualify for the Premium Allocation Approach. It’s simpler: you allocate premiums over time, with a small adjustment for claims development.
The PAA is allowed when the coverage period is one year or less, or when the company demonstrates that using GMM wouldn’t make a material difference. That’s the escape hatch. Many general insurers use it. But only if they can justify it.
The difference? A motor insurer using PAA might show steady profit margins year after year. The same company using GMM could swing wildly—especially after a major weather event. The PAA smooths; the GMM reveals. Which is better? That depends on what you value—stability or truth.
Criteria for Using the PAA Instead of GMM
To qualify for PAA, the contract must meet one of two tests: duration or materiality. If the contract period is ≤12 months, you’re in. If it’s longer, you must prove that GMM and PAA produce materially similar results. That requires analysis—often stress testing across multiple scenarios.
Some companies run both models just to compare. The cost? Significant. But the alternative—being forced into GMM—is worse. Especially for portfolios with long-tail liabilities, like liability insurance. One UK insurer found a 12% difference in reported profits between models. That’s not material. That’s a red flag.
Frequently Asked Questions
Is the GMM Required for All Insurance Contracts?
No. The GMM is the default, but contracts can qualify for PAA or VFA. The VFA applies to investment-linked products—like unit-linked life policies—where fees are variable and based on policyholder returns. If you’re managing assets for clients and taking a cut, VFA may apply. But only if you meet strict criteria around fee variability and risk transfer.
How Does GMM Affect Financial Statement Volatility?
It increases it. Cash flow estimates change. Discount rates fluctuate. The CSM adjusts. As a result, insurers’ profits can swing more under GMM than under IFRS 4. A French reinsurer reported a 23% drop in net income in Q1 2023—solely due to rate changes affecting GMM valuations. No real losses. Just accounting. That said, the market is adapting. Investors now expect this volatility.
Can Companies Switch Between GMM and PAA?
Only if circumstances change. You can’t flip back and forth. If you start with GAA and later find PAA is appropriate (or vice versa), you must apply it prospectively. And you must disclose why. Regulators watch this closely. One German insurer tried switching models mid-year—only to be challenged by auditors. The issue remains: consistency matters.
The Bottom Line: Is GMM Worth the Pain?
Let’s be clear about this: the GMM is complex. It demands more data, more systems, more actuarial input. Implementation costs for some insurers exceeded €150 million. Ongoing compliance isn’t cheap either. But—and this is a big but—it delivers better information.
I find this overrated: the idea that simplicity should trump accuracy in financial reporting. Yes, PAA is easier. Yes, it reduces noise. But at what cost? We’re not just accounting for money. We’re reflecting risk, obligation, and performance.
The GMM forces companies to confront their assumptions. It links profit recognition to service delivery. It aligns insurance accounting with the rest of the financial world. That’s not easy. But it’s right.
Do experts disagree? Absolutely. Some say the CSM amortization rules are too rigid. Others argue the discounting methodology is unrealistic in illiquid markets. Data is still lacking on long-term impacts. Honestly, it is unclear how well it will hold up in a sustained crisis.
Yet, one thing’s certain: the GMM is here to stay. And that changes everything.