Intergenerational wealth transfer strategies to reduce tax drags

By Ruan Breed *

Paying SARS R2.7 million today could leave your heirs millions richer

A voluntary capital gains tax bill feels like the last thing an elderly investor should sign up for. But run the numbers on what death actually costs a large South African estate, and the picture changes completely.

South Africans spend decades worrying about the tax they pay while they are alive, and almost no time on the far larger tax number that waits on them when they die. On a sizeable estate, death triggers three separate costs in quick succession: capital gains tax on a deemed disposal of assets, executor's costs on the gross value of the estate, and estate duty on the remaining assets. Each one on its own is manageable. Put together, they can eliminate a quarter to a third of a family's wealth in a matter of months and then repeat it all over again a generation later. Yes, it actually costs money to die. 

The fortunate truth is that much of this is optional. With the right structuring done while the investor is still alive, two of the three charges can be legally reduced or eliminated in full, and the third can be prevented from ever repeating. The catch is that the most effective version of the fix requires something few investors have the appetite for: writing SARS a large cheque voluntarily, years before anyone has died.

A worked example makes the case study a little more understandable. The client below is fictitious, but the numbers, rates and legal aspects are real and reflect the 2026/27 tax year.

Meet Estelle*

Estelle* is 79, recently widowed and honest about the fact that her health is failing. Her income needs are covered. She draws income from a R9 million living annuity and keeps a healthy cash reserve. The asset/portfolio that matters is a R55 million discretionary share portfolio built up over decades, with a base cost of roughly R40 million. Her only child, Johan*, is in his early fifties and her real ambition is that the capital ultimately funds her grandchildren.

Her stated objectives were simple: get the money to the next generation with as little tax leakage as possible, make sure there is liquidity when she dies, and avoid the family being stuck in a drawn-out estate administration process.

What doing nothing costs her

Suppose Estelle leaves the portfolio exactly where it is and passes away three years from now. Assuming growth of 9% a year, the shares would be worth about R71.2 million at her death. At that point, the tax and costs start running.

First, section 9HA of the Income Tax Act deems her to have disposed of the portfolio at market value on the date of death. With a R40 million base cost, which crystallises a gain of just over R31 million. At the 40% inclusion rate and a 45% marginal rate that is a CGT liability of roughly R5.6 million. Second, the executor is entitled to remuneration of up to 3.5% plus VAT on the gross estate, which on this asset alone comes to about R2.9 million. Third, because her dutiable estate comfortably exceeds R30 million, estate duty applies at the top marginal rate of 25%, taking a further R15.7 million. Of the R71.2 million, approximately R47.1 million reaches the family and after a winding-up process that, for estates of this size, easily takes eighteen months or longer through the Master's Office.

That is round one only. Whatever Johan inherits personally will run the exact same path on his death: deemed disposal CGT, executor's fees, estate duty.

The counterintuitive move

The alternative Estelle's adviser put on the table sounds, at first, like malpractice. Sell the entire discretionary portfolio now and trigger the capital gain. On the R15 million gain, that means an immediate CGT expense of about R2.69 million, which can be provided for from the funds of the portfolio that has been sold.

The net proceeds of roughly R52.3 million are then reinvested into a local endowment policy wrapper, or several policies, spread across different underlying portfolios so that future withdrawals can be managed in a flexible manner. The important factor is the beneficiary nomination. Instead of naming Johan, Estelle nominates the family's existing inter vivos trust, of which she and Johan are both beneficiaries. 

This nomination changes the previous issues of her death entirely, for mainly four reasons.

One: assets that pass to a nominated beneficiary on a policy do not pass through the estate administration process. There are no executor's fees on that value, no waiting for the Master's Office, and the trust has access to the capital within weeks rather than years. The liquidity problem solves itself and the trust can even fund the estate duty attributable to the policies, from the policies.

Two: there is no CGT event at her death. Because the trust elects to continue the policies rather than surrender them, no disposal takes place and the deemed disposal at death that would have cost R5.6 million on the share portfolio is eliminated. Tax on growth inside the wrapper is paid continuously by the insurer at the individual policyholder fund rates: 30% on income and an effective 12% on capital gains, which, for a taxpayer already at the 45% marginal rate, is an effective saving of 6% compared to an effective 18% CGT rate.

Three: section 7C, the anti-avoidance provision that penalises interest-free loans to trusts, is not in play. The trust acquires the policies by inheritance through a beneficiary nomination, not by way of a loan account, so there is no deemed donation.

Four: this is where the strategy stops being a fee-saving trick and becomes generational planning: the capital lands in the trust rather than in Johan's hands. It never forms part of his personal estate. When he dies, there is no deemed disposal, no executor's fee, and no estate duty on this money. The entire second round of death costs is eliminated at Estelles death.

The policy values still form part of Estelle's own dutiable estate. Nothing about the wrapper eliminates estate duty at the first death, the share portfolio would have been dutiable either way and so are the policies. The saving at her death comes from the CGT roll-over and the executor's fees. The estate duty saving only comes in a generation later. But it must be taken into account that the estate duty would have been payable regardless.

Round one: three years

Figure 1 compares the two routes to Estelle's death in mid-2029. The 'do nothing' line starts higher, she keeps the R2.69 million she would have paid SARS and grows to R71.2 million against the restructured route's R67.7 million. Then death reverses this. After CGT, executor's fees and estate duty, the untouched portfolio delivers R47.1 million to the family. The policy route, bearing estate duty only, delivers R50.8 million. Paying R2.7 million early bought the family R3.7 million within three years, plus immediate access to the capital.

Round two: the next generation

The three-year picture understates the true value from this restructuring. Figure 2 extends the comparison to 2049, assuming Johan survives his mother by twenty years, and both pools of capital keep compounding at 9%.

On the personal ownership route, Johan's inherited R47.1 million grows to about R263.9 million and then his own estate takes the same three hits: roughly R39 million in CGT, R10.6 million in executor's fees and R53.6 million in estate duty, leaving R160.8 million for the grandchildren.

In the trust route, the R50.8 million grows to about R284.7 million and simply stays where it is. No death costs apply because nobody who owns the assets has died. The gap between the two outcomes is R124 million on a strategy whose entire upfront cost was R2.69 million in tax.

The caveats one should be aware of

None of this is a free lunch (there is only one – remember that quote about diversification?). The upfront CGT is actual money surrendered early. If the client dies much later than expected, or if growth disappoints, the break-even stretches out. Endowment wrappers carry a five-year restriction period with limits on access, and their internal tax rates only make sense to investors in the higher brackets (or for investors with a substantial asset base). For someone below the 30% marginal rate the wrapper can be tax-inefficient, unless the capital base/growth is substantial. Trusts bring about their own running costs, governance duties and a punitive 45% flat income tax rate and 36% effective CGT rate on amounts retained within the trust, which makes competent use of the conduit principle and distribution planning essential. And the illustration deliberately simplifies: it applies estate duty at the 25% marginal rate on the assumption that the abatement and the 20% band are absorbed by other estate assets, and it assumes identical gross growth in both structures for comparability.

What the numbers nonetheless demonstrate is a principle that runs against every instinct investors have about tax. Deferral is not always the best strategy in effective portfolio management, especially in cases where succession planning is one of the primary objectives. For an elderly investor with a large unrealised gain, an estate that will be dutied and heirs who will one day face the same three charges, accepting a known, controlled tax cost today can be much cheaper than leaving an uncontrolled decision to compound. The planning conversation with a 79-year-old should never be only about the 79-year-old. It is a conversation about the estate of a child who may still be alive in 2060, and the long term objectives of these funds.

Estate restructuring of this kind sits in the territory where product rules, trust law and four different taxes intersect. Get independent advice, model your own numbers, and stress-test the assumptions, but do it while the person at the centre of the plan is still there to make the relevant decisions.

Assumptions and disclosures

*The case study is hypothetical and all figures are illustrative. Calculations reflect the 2026/27 South African tax year: CGT inclusion rate of 40% for individuals with a 45% maximum marginal rate (18% effective); annual CGT exclusion of R40 000, increased to R440 000 in the year of death; executor's remuneration at the maximum tariff of 3.5% plus VAT; estate duty at 20% on the first R30 million of the dutiable estate and 25% above it, with the section 4A abatement of R3.5 million and the 20% band assumed absorbed by other estate assets; individual policyholder fund taxation of 30% on income and an effective 12% on capital gains. Growth of 9% per annum is assumed in both structures for comparability; in practice the internal tax drag of a policy wrapper differs from that of a discretionary portfolio. Living annuity capital falls outside the estate for estate duty purposes and is excluded from the comparison. This article is general information, not financial, tax or legal advice; readers should consult a qualified adviser about their own circumstances.