I work as head of actuarial function (HAF) and consultant to life and non-life insurers of quite different sizes and sophistication. This gives me a view of the issues that come up consistently across the industry. I plan to publish a short set of questions annually that I think non-executive directors should be putting to their actuarial teams. This is the first edition. It clearly cannot be comprehensive. Every insurer’s circumstances are different, and these questions should supplement, not replace, the ones that arise from your own business.
For guidance on how to assess the performance of your HAF, see Evaluating the HAF toolkit for insurance boards on the Milliman South Africa website.
1. What has the sustained decrease in yields and compression in spreads done to our asset-liability management (ALM) position and the sensitivity of our risk products?
Over the past two years, yields have declined and credit spreads have compressed across most durations, particularly at the long end.1 Some ALM mismatches have been exposed by the movement. Others have been created by the persistent scarcity of long-dated bonds.
On the reporting side, insurers who develop their own International Financial Reporting Standard (IFRS) 17 discount curves have flexibility but must be able to explain divergences from the Prudential Authority (PA) curve. Those who adopt the PA curve directly inherit its features, including the low ultimate forward rate and heavy reliance on the last liquid bond. The treatment of the last liquid point (LLP) and the extrapolation methodology beyond it involves real trade-offs between economic reality, ease of matching, and balance sheet volatility. Directors should understand which choices have been made, why, and whether those choices are still appropriate given how much the curve has moved.
2. What do lower yields and tighter spreads mean for annuity and guaranteed product volumes and margins, and how is the business responding?
Non-profit annuity and guaranteed endowment sales have been a strong feature since yields rose post-COVID-19. That medium-term tailwind is now fading. The drivers are not only monetary easing. Foreign purchases of long-dated bonds, credit spread compression on the back of positive economic news, improved political stability, and credit rating upgrades have all contributed to the shift.
Lower yields compress margins on guarantees and may shift relative attractiveness toward with-profit annuities, living annuities, and other products. The question for the board: Does management have a view on when the product mix needs to adjust, and are the distribution channels positioned for that? The oil price-driven inflation risk adds a further complication. If the loosening cycle reverses, the picture changes again. Brokers need something to sell, and the business needs to be ready with it.
3. Let’s say we have to restate our 31 December 2025 results. What will have caused that?
Instead of asking actuaries to defend the results, ask them to imagine the results were wrong and work backward. What would the cause have been? This forces a different kind of thinking. It surfaces model errors, assumption drift, data quality issues, late reinsurance settlements, or methodology problems that defensive questioning does not reach.
It is particularly relevant in 2026. IFRS 17 is still young enough that the probability of restatement is arguably elevated relative to the mature IFRS 4 regime. The macro volatility through 2025 created unusual experience patterns that test models and assumptions in ways that benign years do not.
4. How is artificial intelligence (AI) being used within the business? Is it well controlled or fragile? What risks are added through use of AI, and what risks are added by not adopting it?
Is adoption top-down or bottom-up? Well controlled and slow and bureaucratic, or chaotic and innovative and fragile? These are different failure modes with different risk profiles. A board should understand which one it is dealing with and may find the answer is both, in different parts of the organisation.
What risks are being added through the use of AI: in data leakage, in model validation, in reliance on outputs that have not been independently checked? And equally, what risks are being added by not adopting AI more broadly: in falling behind competitors, in failing to attract talent who expect modern tooling, in leaving efficiency gains on the table? Push for specifics. Where exactly is AI being used? Who approved it? Is AI being used where it should not be? What governance exists around it? If the answer is vague, that is itself telling.
5. Recent non-life claims experience has been good and reinsurance rates are softening. Are we responding strategically or reactively?
Global reinsurance capacity has expanded and non-life pricing has softened materially at recent renewals.2 At the same time, 2024 and 2025 gross claims experience has been unusually benign following a period of significant catastrophe losses. This combination creates the strategic trap: Good gross experience makes even cheap reinsurance look expensive relative to recent loss ratios, which can tempt boards into increasing retentions at the wrong point in the cycle.
Claims ratios across much of the non-life market are currently below long-term averages. Profits are above. It is a question of when, not whether, competition undercuts margins or claims experience reverts to the mean. Are we increasing retentions because of good gross experience, and if so, what happens when that experience turns? Conversely, if reinsurance is now cheaper, why are we not buying more cover to lock in favourable terms and capacity while they last?
6. How well did our Own Risk and Solvency Assessment (ORSA) scenarios prepare us for what actually happened, and what should we be testing next?
2025 delivered record JSE returns (albeit concentrated in natural resources), a strong rand, massively changing yield curves, trade war escalation, and a volatile oil price. 2026 has already brought international drama in Venezuela and Iran. US tariffs, both the levels and the uncertainty, affect South African businesses and the outlook for the global economy. Oil prices are currently elevated due to the war in the Middle East, risks to the Strait of Hormuz, and the shutting down of oil wells due to full storage infrastructure. South Africa has a new inflation target (3% midpoint). There is a real possibility that the monetary easing cycle, which was until recently expected to continue, may reverse if imported inflation picks up.
How many of these events were covered by prior ORSA scenarios, and at what severity? Did yield curve stress tests adjust the level but not the “twist”? If none of our stress tests came close to the combination of events we actually experienced, that tells us something about how we are setting scenarios. Are the ORSA scenarios broad enough to cover the spectrum of possibilities?
There is also a process question. If the ORSA takes four months from setting the basis and scenarios to reaching the board, the results may already be out of date by the time they inform decisions. Can we re-run key scenarios quickly when the environment shifts, or are we locked into an annual cycle?
7. IFRS 17: What should I be focusing on, are we still finding problems, and is doing it too resource intensive?
Two full years of IFRS 17 reporting. The question for directors is no longer whether it has been implemented but whether the outputs are trusted and useful. Is growth in the contractual service margin (CSM) the right headline measure, or is it masking issues? For new business, how far along are we in developing a meaningful profitability metric: CSM plus loss component, adjusted for full expenses, tax, and risk adjustment versus cost of capital, so that it becomes genuinely comparable to the embedded value measures the market is accustomed to? Can these IFRS metrics replace embedded value measures, and what do we gain or lose with this change?
What methodology choices made at implementation are now being reconsidered? The other comprehensive income (OCI) election is a natural candidate given the yield curve movements of the past two years. And directly: What is the probability of material restatement? Is it still temporarily elevated as the new standard beds down, or is IFRS 17 inherently more restatement-prone than IFRS 4 was in its mature years?
8. What would you be able to do if you had more actuarial resources?
This is a better question than “do you have enough resources,” which will often get a cautious “yes.” There are always trade-offs and limitations within an actuarial team. Certain analysis inevitably gets deprioritised. Investigations get deferred. Model improvements sit in a backlog. The board should know what those trade-offs are.
If you had five more actuaries from tomorrow, how would you deploy them?
This is worth asking of both the first-line actuarial team and the HAF. The HAF has a regulatory responsibility to raise resource concerns, but the framing of this question makes it easier to answer frankly without it sounding like a complaint about management or the CFO.
These questions are intended to start conversations, not interrogations. The best boards treat the relationship with their actuarial team as a dialogue, and the best actuaries welcome being challenged by directors who are asking the right questions.
1 For example, using the Prudential Authority’s prescribed yield curve, the nominal risk-free spot rate at the 30-year point declined from 14.7% at end-December 2023 to 8.9% at end-December 2025. This is compression of approximately 580 basis points. The 10-year point fell from 12.1% to 8.8% (approximately 330 basis points) over the same period. Declines were driven by SARB rate cuts, moderating inflation, lowering inflation target, the Government of National Unity (GNU)-related improvement in sovereign risk sentiment, and the S&P credit rating upgrade in November 2025. The SA-specific factors were amplified by a broader global rotation into emerging market local currency debt.
2 Priya, S., Thakare, S., & Dave, A. (October 2025). South Africa outlook: Opening pathways to greater resilience. Swiss Re Institute. Retrieved 8 April 2026 from https://www.swissre.com/dam/jcr:ad6e0e02-edf3-4400-99f3-081b0890871d/sr-africa-outlook-2025.pdf.