Retirement plans don't fail on the income side. They fail on the cost side, specifically, a stack of unbudgeted costs that arrive in waves in the second half of retirement.

By Murray Anderson, Head of Retail, Prescient Investment Management

The headline numbers in South Africa's retirement debate are well-rehearsed. Only about 6% of South Africans can retire independently. Sanlam Corporate's 2025 Benchmark data puts the effective age of comfortable retirement closer to 80 than 65. These are real problems. But almost every debate about retirement in this country lives on the accumulation side: did the client save enough? Did they preserve? Is their drawdown too aggressive?

The plans that fail in real life rarely fail on the income side. They fail on the cost side, specifically, a stack of unbudgeted costs that arrive in waves in the second half of retirement. These are the expenses our cashflow models systematically undercount. Six of them deserve particular attention.

1.  The retiree inflation basket isn't CPI.

The price basket of an active, independent South African retiree — medical aid, chronic medication, domestic help, private security, backup power, borehole water where needed, municipal rates — has inflated materially above headline CPI for more than a decade. Medical schemes alone announced 2025 contribution increases of between 9.3% and 12.75%. A plan that inflates retirement income at CPI is understating real costs from year one.

2.  The frail-care step-up is a forced purchase, not a lifestyle choice.

A move from independent living to assisted living to full frail care can see monthly costs step up three- to fivefold, and the step-up isn't discretionary. Frail care in the Western Cape now runs around R30,000 to R55,000 a month; specialised dementia care can reach R75,000. Medical aids cover very little of this. Few cashflow plans model the step function; most keep inflating a single line item.

3.  Retirement villages aren't a destination — they're an ongoing cost.

Life-rights schemes don't appreciate like freehold property. Levies at established Western Cape retirement villages have been escalating at between 8% and 12% annually in recent years — well above CPI — and resale can take 12 to 24 months. The "two-body problem" — one spouse needing frail care while the other doesn't — plays havoc with levy structures that assume a single household. We recommend villages as a safe landing, but the economics deserve the same scrutiny we give to an equity fund on a platform.

4.  Cognitive decline is a financial planning risk, not a medical one.

Alzheimer's South Africa estimates more than 190,000 South Africans currently live with dementia; roughly one in three people over 85 will face some form of cognitive impairment. Yet most retirement plans make no provision for diminished financial capacity — no trusted contact on file, no enduring mandate, no protocol for how the advisor responds when a client can no longer manage their own decisions. That is not planning; it is hope.

5.  Reverse dependency is a very South African problem.

A meaningful share of our affluent retired clients are supporting adult children financially, often across borders. Grandchildren's school fees, international airfares, top-up living costs, even deposits on offshore property. In a high-unemployment, high-emigration economy like South Africa's, the probability that a retiring parent carries some form of ongoing financial responsibility for an adult child is arguably higher here than almost anywhere in the developed world. These expenses are real, recurring, and they virtually never appear in a standard retirement plan. They materially shorten portfolio life.

6.  The last five years carry their own cost shock.

Executor's fees of up to 3.5% plus VAT on gross estate. Capital gains tax on deemed disposal at death. Estate duty. Final medical bills. The closing-down costs of a life. These are often the single largest drawdown a portfolio will ever take, and they are almost universally modelled as zero.

The investment half of the same problem

Every one of these cost shocks leads to the same investment conclusion, and it is a conclusion most retirees resist at exactly the wrong moment. If a retiree's real cost basket compounds at CPI plus three to five percent, the only way a portfolio keeps up is by delivering real returns of at least CPI plus four, net of fees. Over a thirty-year retirement horizon, that is not a low-risk outcome. It requires meaningful, sustained exposure to growth assets well into retirement, not only up to it.

The historical evidence is uncomfortable for conservative portfolios. Allan Gray's long-run South African data shows that a 4% starting drawdown with no local equity exposure had roughly a 30% probability of lasting thirty years. The same drawdown with 60% equity exposure had around a 90% probability. Ninety One's research lands in the same place: living annuities need at least 60% in growth assets, and de-risking into retirement is rarely a prudent strategy. Prescient's own analysis of domestic multi-asset portfolios is consistent with these findings.

"I can't afford to lose money now." It is the most understandable sentiment in wealth management, and mathematically one of the most destructive.

The client is defending against a short-term drawdown while accepting a thirty-year shortfall against compounding costs. Under-risking is the quiet twin of over-drawing; both end at the same destination.

The cost side of this piece and the investment side are the same argument. Financial plans fail because costs rise faster than CPI, and because portfolios are positioned as if they won't.

The uncomfortable question for our industry

Every one of these costs is knowable. None of them is exotic. Every one of these investment decisions is already well-researched. Yet most of our planning tools and most of our client conversations are still built around an accumulation-era question: will they have enough on day one of retirement? The sharper, harder question is: will they still have enough when the fourth cost shock hits, through a portfolio still carrying the growth to outrun them?

Accumulation and decumulation are different disciplines. Advisors are trained  and paid  for the first. The profession has not yet fully matured into the second. If the retirement conversation in South Africa is going to change and it must it won't be because we got the accumulation phase right. It will be because we finally got honest about the cost side, and built portfolios with enough growth to meet both goals.

Murray Anderson is Head of Retail at Prescient Investment Management. This opinion draws on a retirement panel discussion that took place with industry colleagues on 23 April 2026.