Retirees and insurers fear different longevity risks, and everyone pays for it
Updated: 5 hours ago
Retirees who manage their own retirement income prudently have to plan for a longer-than-average lifespan or risk outliving their money. An insurer could remove almost all of that risk. So why do so few people buy annuities?
Life annuities get very little love from retirees and institutions:
they are expensive due to the nature of their guarantee
retirees find them unattractive for various reasons (money lost on early death, giving up control of their savings)
institutions consider them risky, low-margin, and capital-intensive (often ring-fenced in separate entities or strategically volume-limited)
Putting aside the complex behavioral biases behind the annuity puzzle, the reality is that the need for longevity risk coverage will increase in the coming years.
In this context, we offer the following perspectives:
retirees and institutions face two different types of longevity risks, but
the type that costs institutions the most to protect is the type that matters least to retirees.
There are two longevity risks, not one
Lifespan risk: How long will a retiree live? To age 75? To age 100? Longer?
Trend risk: What if mortality estimates are wrong and people end up living longer on average than actuaries expect?
These two risks look very different from the perspective of a retiree versus that of an institution.
Lifespan risk: huge for one retiree, but almost nothing for a large group of them
A 65-year-old man has a 50% chance of reaching 87.6 and a 5% chance of reaching 98.6. If he plans cautiously, say, to age 98.6, his savings will have to stretch about 25% further. That equates to a 20% reduction in annual income… and he still has a 50% chance of dying before age 87.6, without ever having benefited from his cautious planning.
Pooling changes the calculus, however. By pooling retirees together, the law of large numbers quickly takes over.

Looking at the top of the shaded region, which represents a 1-in-200 outcome, we see that the cost of lifespan risk protection for our retiree on his own (pool size of 1) roughly equates to a whopping 32% of his retirement assets. For a pool of 10,000 people, the cost is only 0.4%. A huge difference!
Trend risk: small for a retiree, but an existential threat to insurers
Now let’s consider what could happen if everyone lives longer than expected. This is a systemic risk — one that cannot be diversified. Solvency II, the European insurance regulation, requires insurers to hold enough capital to survive a 1-in-200 event, which it calibrates as an unexpected 20% drop in mortality rates.
For an individual retiree on his own, a shock of that level equates to about 3.6% of assets, reducing his income commensurately. This would be unwelcome, but it wouldn’t be a disaster, and most people could live with it and feel thankful for the gift of long life.
For an insurer, however, that same 3.6% hits every annuity contract at once, and it cannot be diversified away, so it has to be backed by expensive regulatory capital and additional risk margins.
Guaranteeing the trend risk is expensive
Holding capital against trend risk adds roughly 16 basis points a year to an insurer’s costs, or about 1–2% on the price of an annuity. So, insuring the trend risk costs an insured annuity buyer about 1–2% of the price, which in turn reduces his income by about that same amount, for certain. By contrast, the loss to a retiree who goes without insurance would be about 3.6% of assets, in a 1-in-200 outcome! If mortality turns out as expected, it would be zero. One would need to be quite risk averse to find that trade-off attractive.
The misalignment between individuals and institutions in one table
Individual (65-year-old male) | Institution (10,000 lives) | |
|---|---|---|
Lifespan risk (1-in-200) | 32.2% | 0.4% |
Trend risk (1-in-200) | 3.6% | 3.6% |
Effect of each risk on the value of lifetime income
The table tells the story. The individual’s greatest exposure, by far, is lifespan risk, which can be handled cheaply and easily by pooling. The insurer’s greatest exposure, by far, is trend risk, which can only be handled by expensive capital reserving. Each side worries most about the risk that the other side can handle easily. It doesn’t have to be that way.
Pooling is where the value is. Guaranteeing the trend risk is an expensive add-on, with limited benefits for retirees.
The solution: pool the lifespan, share the trend
Longevity risk-sharing pools efficiently align the risk exposure:
Lifespan risk is pooled: people get lifetime income without the spending cut that comes with going it alone.
Trend risk is shared: if everyone lives significantly longer than expected, income adjusts gradually, by perhaps a few percentage points spread over several years (as opposed to the risk an insurer would face in eating through its reserves and potentially going bankrupt).
No solvency capital is required: by adjusting smoothly to long-term mortality trends, a risk-sharing pool can remain solvent without the need to hold aside any risk reserves, at a fraction of the cost of a guaranteed annuity.
Guarantees still make sense for some retirees, such as those with no spending flexibility and no other sources of income. But for most people, paying to transfer trend risk to an institution means paying for the wrong insurance.
What this means for pension and insurance leaders
Sell certainty where it adds value. Capital limits constrain the capacity for guaranteed annuities and defined benefit (DB) plans. To make the most efficient use of risk capital, focus on the risks that matter most.
Unbundle longevity risk. Offer lifespan pooling on its own and make trend protection a separately priced option for those who want it.
Make low-cost lifetime income the default for defined contribution (DC) decumulation. Longevity risk-sharing pools give members most of the benefit of an annuity at a fraction of the capital cost.
Read the full paper: Longevity risk transfer: Addressing the misalignment between institutions and individuals, with the full methodology and results for women and other ages.
Want to design a longevity risk-sharing pool? Talk to Nuovalo.
Figures: Canadian Pensioner Mortality 2024 (CPM2024) tables, 4% interest, life annuity bought at 65 with a 15-year guaranteed period, 20,000 Monte Carlo simulations.
Pascal Winter and Richard K. Fullmer, Nuovalo Ltd.






Comments