Living annuity drawdown limits and sequence risk: why the order of returns can matter more than the average
A R15 million living annuity drawing 4% a year looks comfortable on paper. Run the same plan through two poor early years, and South Africa's drawdown rules can turn it into an income shortfall by the mid-eighties. Here is how that happens — and what it suggests about where retirees should focus their attention.
The plan that looks comfortable
Consider a 65-year-old with R15 million in a living annuity. He plans to draw R600 000 a year — R480 000 for essentials, R120 000 for the things that make retirement worth having. That is an initial drawdown of 4.0%. He assumes long-term returns of 9% and inflation of 5.5%, both reasonable for a balanced South African portfolio, and he plans to age 95.
If markets deliver that 9% steadily, the arithmetic is generous: by age 95 his capital could have grown to roughly R59 million in future rands — spending fully funded throughout, with a substantial legacy remaining.
Most retirement conversations stop here. The average return is plausible, the drawdown is modest, the projection is green. What could go wrong?
The rules of the vehicle come first
A South African living annuity is not a flexible bank account. Two regulatory constraints define how income actually works:
You must draw between 2.5% and 17.5% of capital each year. The percentage is set once a year, on the policy anniversary, and applies until the next one. You cannot draw less than the floor in years you don't need the money, and you cannot draw more than the cap in years you do.
The percentages apply to the capital that remains — not to the capital you started with. When the portfolio falls, the rand value of the maximum income falls with it.
The rules do not flex because markets had a bad year. This rigidity is the mechanism through which investment risk becomes income risk.
Living annuity income is also taxed as ordinary income, which means the gross drawdown required to fund a given lifestyle is higher than the spending figure itself — another quiet pressure on the capital base.
Same average, different order: sequence risk
Here is the part that average-return projections conceal. Two retirements can experience the same average return over thirty years and end in entirely different places, depending on the order in which good and bad years arrive.
The reason is withdrawals. A portfolio in accumulation can ride out an early crash — no money is leaving, and the recovery applies to the full base. A portfolio in drawdown cannot. Every rand withdrawn during a downturn is a rand that never participates in the recovery. Losses early in retirement are, in effect, permanently locked in by the income the retiree had no choice but to take.
Return to our 65-year-old. We stress-tested his exact plan under three scenarios using our living annuity stress test:
| Scenario | What changes | Capital at age 95 |
|---|---|---|
| Base case | 9% returns hold throughout | ≈ R59.1 million |
| Bad early markets | −20% in year one, −5% in year two, then normal returns resume | ≈ R3.1 million, with income falling short of spending from age 84 |
| High inflation | 9% inflation for the first five years, then normalising | ≈ R35.3 million |
Deterministic projections from identical starting assumptions; figures in future rands. Illustrative only — not a prediction of any actual outcome.
Read that middle row again. Nothing about the retiree's intentions changed. His average long-term return barely changed. Only the first two years changed — and the outcome moved from a R59 million legacy to a plan that cannot fund his own spending from age 84.
Where the 17.5% cap quietly bites
Notice that in the bad-early-markets scenario, the plan does not "run out of money" in the dramatic sense. Something subtler happens.
As capital shrinks and spending keeps rising with inflation, the drawdown percentage required to fund the same lifestyle climbs — 4%, then 6%, then 10%. Eventually it reaches the regulatory ceiling of 17.5%. From that point, the retiree may draw no more than 17.5% of whatever capital remains, regardless of what his life costs. Each year the capital is smaller, so 17.5% of it is smaller too.
The plan does not fail with a bang. It fails by increments — a gap between income and spending that appears in the mid-eighties and widens every year thereafter.
This is also why a poor early decade is more dangerous than a poor late one: it brings the cap forward into years the retiree is still likely to be alive. A male retiring at 65 in South Africa has a table life expectancy of around 85 — and roughly half of such retirees will live beyond it. An income shortfall beginning at 84 is not a tail risk; it sits in the middle of the distribution.
Inflation is the patient version of the same problem
The high-inflation scenario ends less dramatically, but the mechanism deserves attention. A burst of elevated inflation early in retirement permanently raises the base from which all future spending grows. The portfolio may recover; the spending level does not come back down. Five years of 9% inflation in our example reduced the age-95 outcome by roughly R24 million — without a single negative market year.
This is a governance problem, not a returns problem
The instinctive response to sequence risk is to chase better returns. The arithmetic above suggests that instinct is misplaced. In the scenario that mattered, returns over thirty years were nearly identical — what differed was what happened to withdrawals and decisions during a short, bad window.
The decisions that actually move the outcome are made annually, often under pressure, and mostly in bad years:
Where should this year's drawdown percentage be set, given what markets just did? Which assets should fund the withdrawal, so that depressed assets are left to recover? Which spending is genuinely essential, and what can flex for a year or two? Should other capital outside the annuity carry the load while it heals? And — hardest of all — is the plan still on course, or is this the year that quiet adjustments need to be made?
None of these questions is answered by an average-return projection, and few of them are comfortably answered alone, in the middle of a drawdown, by the person whose lifestyle is at stake. That is the case for treating retirement income as something that is governed — reviewed annually, against scenarios, with someone whose job is to hold the framework steady when the year is not.
See what your own plan depends on
The same stress test used in this article is freely available. It takes about five minutes, models South Africa's 2.5%–17.5% drawdown limits exactly, and shows how your plan could respond to bad early markets, high inflation and a long life.
Run the Living Annuity Stress TestIllustrative only. No advice. No obligation.
What a structured review examines
When we review a living annuity for a client, the stress test is the beginning of the conversation, not the end. A full review considers the annuity alongside everything the projection cannot see: discretionary investments and the sequence in which different pools should be drawn; the tax position, since drawdown timing and asset location change the after-tax outcome; estate intentions, because a living annuity passes to beneficiaries outside the estate and the nomination structure matters; and the spending architecture itself — which commitments are fixed, and where the plan has room to breathe in a poor year.
The objective is not a prettier projection. It is a documented basis for the annual decisions — made calmly, in advance, rather than improvised in the middle of a bad market.
