Estate Planning — Conradie Family

What it is: Section 4(q) of the Estate Duty Act lets a deceased spouse leave any amount to the surviving spouse with full deduction from the dutiable estate. No estate duty payable on that portion.

Why first death is R0: Your March 2021 will leaves the residue to Nicolette. Everything in your dutiable estate (R178m) flows to her under s4(q) — net dutiable becomes R0, duty becomes R0.

Cornwall 228 / 228B (clarified): Same erf (Erf 796, title deed T85781/2016) — both halves pass to surviving spouse under clause 5.1 → full s4(q). No carve-out, no clause 5.2 trap. Earlier version of this analysis incorrectly treated 228B as a separate property; corrected 2026-05-01 per Elmar.

What happens at death: SARS treats death as if you sold every asset you own at market value, the day before you died. Capital gain = market value minus base cost. Tax on the gain = capital gains tax (CGT). Inclusion rate 40% × top marginal 45% = ~18% effective on the gain.

Section 9HB(2)(a) rollover: If an asset is bequeathed to your spouse, the deemed disposal is suspended. She inherits your original base cost, and the CGT is only triggered when she sells (or when she dies — whichever comes first).

Your situation at first death: CGT ≈ R0 because everything that goes to Nicolette qualifies for the rollover. Cornwall 228B is the only carve-out, and it has no gain (mortgage > value).

The basic abatement (s4A): Every estate gets R3.5m exempt from estate duty before any duty is calculated. Above that, 20% on the next R30m, 25% above R30m.

Portability (s4A(2)): If first-dying spouse leaves everything to surviving spouse (full s4(q)), the first abatement is "rolled over" to the second-dying estate. Second estate gets R7m abated instead of R3.5m. This is automatic — no special election needed.

What this means for you: At second death today's R177m aggregate becomes R170m net dutiable. Duty = 20% × R30m + 25% × R140m = R6m + R35m = R41m.

Phase 1 — freeze (executed 11 December 2025): You and Nicolette transferred R109.5m of personal investments + loan claims into Werda Operations in exchange for Class B shares. Section 42 of the Income Tax Act made this a tax-neutral roll-over — no CGT triggered, base costs preserved.

What moved out of your personal estates: PSG Creator FoF, Preserver FoF, Securities; Peregrine High Growth + Pure Hedge; loan claims to ENC Trust + IEG Trust + YellowNickel.

What you got in exchange: 15 Class B shares each. Base Value R94.2m (Elmar) + R15.3m (Nicolette) = R109.5m combined.

The MOI cap (clauses 7.3.4 + 7.3.5): Class B shares are NOT entitled to participate in WO's net assets on liquidation. Their entitlement = "B Share Entitlement" formula: Base Value minus any distributions received. All growth in WO accrues to Class A (Werda Investments → ENC Trust). Your kids inherit the growth; you and Nicolette are capped at the original R109.5m. CGT-on-death on Class B ≈ R0 (deemed market value ≤ base cost).

Phase 2 — evacuate (year 5+, planned): Once enough time has passed to satisfy GAAR / substance-over-form scrutiny, WO declares dividends in specie of its underlying portfolio — actual assets, not cash — up to its Class A shareholder, Werda Investments. Tax treatment:

• Dividends tax: Exempt under s64F(1)(a) — dividends between resident companies. No 20% withholding.
• WI's hands: Receives the assets at WO's base cost (rolled over per various dividend / asset-for-share provisions). No CGT on transfer.
• WI is wholly trust-owned. Assets now sit inside the trust's wholly-owned holdco — fully outside your personal dutiable estates.

What this does to Class B value:

• On paper, B Share Entitlement still says R109.5m.
• In practice, WO no longer has assets to pay it. Companies Act s46 solvency-and-liquidity test blocks any redemption when WO is asset-less.
• "Market Value" definition (MOI clause 1.2.12) = "price obtainable from a willing buyer to a willing seller." If WO is empty, no rational buyer pays anything material → Class B market value at death approaches R0.
• Estate duty deemed-disposal at market value → Class B contribution to dutiable estate collapses to near-zero.

The MOI already fixes the price. Per clauses 7.3.4 + 7.3.5 + the B Share Entitlement formula (clause 1.2.5), the redemption amount = Base Value − Distributions received. So the price isn't the question — funding and process are.

What a buy-sell agreement adds: A binding contract between shareholders + the company that says: "When a Class B shareholder dies, the company OR Werda Investments MUST redeem/buy at the B Share Entitlement, and the deceased's estate MUST sell. Payment timing = X months. Source of funds = Y."

The mechanism for Class B:

• Trigger: Death of Class B shareholder.
• Buyer: Werda Operations (redemption) or Werda Investments (buy-out).
• Price: B Share Entitlement per MOI — already locked. No valuation dispute possible.
• Timing: Within 12 months of death — to align with estate duty filing deadline.
• Funding: Life cover policy on each Class B holder, owned by the buyer (Werda Investments / trust). Pay-out funds the redemption. Cash flows from insurer → buyer → deceased's estate → SARS for duty.

What it is: A separate contract from the MOI, governing the relationship between shareholders. It usually covers:

• Pre-emption rights: If a shareholder wants to sell, who gets first refusal at what price.
• Tag-along / drag-along: If majority sells, minority can join (tag) OR is forced to join (drag).
• Reserved matters: Decisions requiring all shareholders' consent (issuing new shares, changing rights, big asset sales).
• Deadlock resolution: What happens if shareholders disagree.
• Death / disability triggers: Auto-buyout provisions (this is where the buy-sell sits).
• Restraint of trade: What you can / can't do with related companies.

WO right now: Werda Investments holds Class A (76.92%, controlling), Elmar holds Class B (11.54%), Nicolette holds Class B (11.54%). No shareholders' agreement. Class A and Class B share rights are different (only the MOI defines them).

Initial concern (now retracted): The first version of this analysis treated "Cornwall 228B" as a separate property acquired after March 2021, which would have fallen under will clause 5.2 (forced sale to ENC Trust on first death).

Corrected understanding (2026-05-01, per Elmar): Cornwall 228 and Cornwall 228B are the same property — Erf 796, Irene Ext 10, Cornwall Hill Country Estate, held under title deed T85781/2016. The "228B" designation in Portfolio.xlsx is an accounting / structural split (likely tracking a second dwelling or building improvement separately for cost basis / mortgage allocation), not a separate erf or title deed.

Will clause 5.1 covers everything on Erf 796. Both 50% interests pass to surviving spouse → full s4(q) deduction → no estate duty on the property at first death.

The rule: Section 3(2)(i) of the Estate Duty Act says: amounts payable from a pension fund, provident fund, or retirement annuity are NOT deemed property of the estate. They flow directly to nominees per the fund's beneficiary nomination form (subject to s37C trustee discretion).

Why this matters here: R19.4m of Safair Pension Fund is the single biggest source of clean liquidity at first death. It bypasses estate duty entirely AND bypasses the executor fee (no fee on assets that don't flow through the estate).

Tax treatment of the lump sum on death: Recipient pays tax per retirement lump-sum table:

• R0–R550,000: 0%
• R550,001–R770,000: 18%
• R770,001–R1,155,000: 27%
• R1,155,001+: 36%

On R19.4m to a single recipient: ~R6.71m tax (~35% effective). Top bracket dominates.

Question: When Elmar retires from Safair, take the lump sum or preserve?

Answer: preserve as long as possible. Pension fund is the most tax-efficient holding vehicle in SA.

Three compounding advantages of staying inside the fund:

1. Tax-free internal compounding. No income tax on interest, no CGT on disposals, no dividends tax. Outside the fund: same R19m earning ~8% returns gets taxed every year (interest 18-45%, dividends 20%, CGT 18% on disposal). Over 10 years of compounding the tax drag outside is 30-50% lower returns vs inside.
2. Estate duty bypass (s3(2)(i)). Stays outside dutiable estate at death. Saves 20-25% on the way out.
3. Executor fee bypass. Pension flows direct to nominees, never through the estate, so no 4.025% executor fee. On R19m that's R0.78m saved.

The unavoidable 36% bite at death: Top bracket of retirement lump-sum table. R19m to single recipient = ~35% effective. Splitting saves a bit on the tax-free R550k band but top-bracket dominates the bulk:

• 1 recipient (Nicolette only): ~R6.71m tax, ~35% effective
• 3 kids equally: ~R6.17m tax, ~32% effective (saves R0.55m)
• 4 ways (Nicolette + 3 kids): ~R5.90m tax, ~30% effective (saves R0.81m)

Why the bite is still cheaper than drawing:

Practical retirement-day moves at Safair:

• Don't take cash. Transfer to a preservation fund — same tax-free environment, no compulsory drawdown, retains s3(2)(i) protection on death.
• Avoid retirement annuity (RA) if you can — RA forces you to annuitise 2/3 at age 55+. Preservation fund is more flexible.
• Two-pot system (Sept 2024+): the "savings component" (1/3 of new contributions) is accessible pre-retirement but taxed at marginal rate — worst path. Don't touch unless emergency.

Practical nomination strategy:

• Don't nominate ENC Family Trust — fund trustees override under s37C (a trust isn't a "dependant").
• Pragmatic split: ~R5m–R10m to Nicolette as first-death liquidity buffer (18-24 months while offshore probate completes); balance to children's testamentary trusts equally. Cayla via her existing testamentary trust (still minor); Rowen + Jared via direct nomination as adult dependants.
• Fund trustees still apply s37C discretion — they look at factual dependency. The form gives strong guidance but isn't binding.

The target: R59.1m PSG Global USD funds (Malta-domiciled), held personally by Elmar. Currently fully exposed to second-death estate duty + CGT + Malta probate complexity.

Bottom line: revised lead recommendation is "DO NOTHING NOW" pending fresh Stephan engagement on foreign assets. The structural complexity, tax leakage, and control-loss risks make all three earlier recommendations less attractive on close inspection. Phase 2 dividend strip on WO is the bigger lever (R36m saving). Focus there first.

Recommended structure: Spamer Guernsey OffshoreCo + Foreign Trust + B-shares (per CDH-Spamer 2020 memo at SecondBrain/sources/family-docs/werda-structure/tax-memorandum-sibongile-manganyi-140420.md; Stephan has implemented this template before).

Elmar settles (no donation, no donations tax)
        ↓
Foreign Trust (Guernsey)  ──── subscribes A-shares 80% ────→  OffshoreCo (Guernsey)
                                                                    ↑
Elmar  ───────  subscribes B-shares 20% (nominal value)  ──────────┘
  │
  └────  lends R59m to OffshoreCo (a COMPANY, not a trust)
            ↓
       OffshoreCo invests in USD portfolio (PSG-Global-equivalent)

How this beats the trust-policyholder route:

Tax leakage comparison (R59m loan, 10y, 8% growth):

Why this works:

• Internal tax rate: 30% individual cap (with natural-person beneficiaries — your kids + Nicolette qualify). Beats personal mix of interest 45% + dividends 20% + CGT 18%. Equivalent to a high-earner's effective rate but with simpler admin and tax deferred to maturity.
• Estate duty: bypass. Policy is a trust asset, not Elmar's. Trust assets are outside personal dutiable estate. Domestic-policy s3(3)(a) "deemed property" rule does not apply because (a) the policy is offshore (not "domestic") and (b) policyholder is the trust, not Elmar.
• No Malta probate at Elmar's death. Policy isn't his asset.
• No executor fee on R59m. Out of estate → no 4.025% fee → R2.4m saved.

The 27% corporate rate via WI is no longer the recommendation. Reasons:

• SARB approval pathway absent. Pacific Padel approval (1475/25, 3 Sept 2025) cited Section B.2(C)(i)(f) — explicitly the FDI lane for operating-company stakes. SARB pushed back hard during the application: CDH (Stephan Spamer + Howmera Parak) had to request urgent in-person meeting in Aug 2025 to motivate. Took 3 months June→Sept. An endowment policy doesn't fit any auto-approval category for SA companies.
• 3% internal-rate edge not worth the EXCON risk. If application fails or is materially delayed, the structure is stuck. Trust route uses an established lane.
• Reconsider WI route only if Stephan finds a category in his read of the current AD Manual Section B.3 — and even then weigh against the certainty of the trust pathway.

SA endowment internal tax — comparison table:

Steps in order (Spamer Guernsey OffshoreCo route):

1. Stephan refresh of CDH 2020 memo for Elmar's specific structure — covers post-2024 s25B + 2023 AIT + Pacific-Padel-precedent learnings. ~R150-250k legal.
2. Settle Foreign Trust in Guernsey — independent trustee company (e.g. Saffery Champness, Estera, Trident Trust). R50-100k setup + ~R100-200k/yr admin.
3. Incorporate OffshoreCo in Guernsey simultaneously. R20-50k setup + ~R30-60k/yr admin.
4. Foreign Trust subscribes 80% A-shares of OffshoreCo at incorporation (R0 in co — no value to subscribe at).
5. Elmar subscribes 20% B-shares at nominal value (e.g. R20k for 20 shares).
6. Sell PSG Global personally → CGT today on accumulated gain (~R8m × 18% = ~R1.4m exit cost). One-time.
7. Elmar lends R59m to OffshoreCo (USD-denominated, term ~10y, interest ~4-5% USD commercially defensible). EXCON via Investec — likely simpler than corporate offshore policy because loan to foreign company is a recognisable category.
8. OffshoreCo deploys capital into USD portfolio (PSG Global equivalent or directly into PSG via Investec / OMI / etc.).
9. OffshoreCo earns returns; tax-free at Guernsey level (0% effective). Cash distributions UP to Elmar (B-shares) and Foreign Trust (A-shares).
10. Elmar's B-share dividends: tax-free under s10B(2) participation exemption — can be used to offset interest income on the loan.
11. Foreign Trust's A-share dividends: stay in trust, no SA tax (s7(8) attribution doesn't trigger because dividends don't flow to Elmar; if distributed to other SA-resident beneficiaries they're taxed in their hands per conduit principle).
12. On Elmar's death: B-shares redeemed at structured value (Spamer designs the redemption-on-death machinery). Loan repaid from OffshoreCo to estate. Foreign Trust continues with 100% of OffshoreCo for kids' benefit.

The "money back in SA" concern — manageable:

• Path A (recommended): WI holds offshore USD account. Endowment payout flows offshore-to-offshore. WI is SA-domiciled but its assets can be offshore (with SARB approval). Money stays USD, stays offshore.
• Path B (default if path A not approved): payout converts to ZAR. FX cost on conversion + ZAR exposure. Avoidable but only if you plan path A in advance.

Risks / open items for CDH (Stephan Spamer):

• Refresh of 2020 memo essential. Sibongile-Manganyi-Rath template predates 2024 s25B conduit narrowing + 2023 AIT process changes + post-2021 loop relaxation + Pacific Padel SARB precedent. Stephan to redraft for Elmar's specific facts.
• s7(8) attribution: Income / gains attributable back to Elmar while alive if structure is "by reason of donation, settlement or other disposition". Spamer's design relies on B-shares being subscribed at MARKET value (or nominal at incorporation) and the loan being arm's-length — to avoid s7(8) trigger. Critical drafting point.
• s31 transfer pricing on the loan: Loan from Elmar to OffshoreCo must be at arm's-length terms. 4-5% USD on a portfolio-backed term loan to a Guernsey investment co is defensible. 0% would be a problem. Document the rate-setting methodology.
• s8E / s8EA "hybrid equity instrument" anti-avoidance: If B-shares look debt-like (mandatory redemption, fixed yield, security), entire dividend gets re-characterised as taxable income. Spamer specifically structures B-shares to retain equity character. Don't deviate from his template.
• Estate duty s3(3)(d) on B-shares: Even if redeemable on death (so not "owned" at moment of death), s3(3)(d) deems property in estate if deceased was "competent to dispose" of it. Spamer's drafting limits Elmar's disposition rights — argument is defensible but SARS may challenge. Worst case: B-share market value at death is in dutiable estate (still small if dividends have stripped value out earlier).
• CFC s9D: OffshoreCo not a CFC because Elmar holds only 20% (Foreign Trust holds 80%, non-resident). If shareholding shifts above 50% SA-resident, CFC kicks in — regular monitoring needed.
• GAAR: Mainstream offshore investment structure when commercial rationale documented (asset diversification, FX hedging, multi-generation planning). Stephan's signed s223 opinion already exists for the template — this is well-trodden ground. Lower risk than novel structures.
• SARB approval for outbound loan: Loan from SA resident to foreign company is a recognised EXCON category. Investec EXCON team handle. Cleaner pathway than the offshore-policy question that killed Option WI.
• Beneficiary admin cost reality: Independent Guernsey trustee (Saffery Champness / Estera / Trident) needs annual fees + audit + bookkeeping. Realistic ~R150-300k/yr all-in.

The rule: If you lend money to a trust at less than the SARS official rate of interest (currently ~9%), the difference between actual interest charged and the official rate is deemed a donation. Donations tax kicks in above the R100k annual exclusion (rate 20% up to R30m, 25% above).

Where this hits you:

• Werda Ops → ENC Trust loans (R4.8m, ceded from you + Nicolette in S42)
• Werda Ops → IEG Trust loan (R6m)
• Werda Ops → YellowNickel loan (R14.1m)

WO is a company not a person, so s7C technically doesn't apply between WO and another entity in quite the same way as person-to-trust — but interest must still be charged at arm's length to avoid transfer pricing / sham concerns.

The case: The Supreme Court of Appeal held that where a trust has only one trustee, or where trustees act as if assets are personally theirs (no trust formalities), courts can pierce the trust veil — treating trust assets as personal assets. This is fatal to estate planning: the whole reason ENC Trust holds Werda Investments is to keep it OUT of your dutiable estate.

Your current trustees: Elmar + Nicolette + Stephan Spamer (3 trustees, fine).

The fix: Trust deed amendment naming a pool of successor trustees (e.g. Johan Hendré Conradie + Albertus Marais + a professional trustee company). Always 2+ trustees post any death event.

Situs: "Where the asset is located for legal purposes." Different jurisdictions have different rules. Each jurisdiction where you hold assets requires its own probate process — separate court, separate forms, separate fees, separate timeline.

Your offshore footprint:

• Malta — PSG Global Funds (R59.1m). Malta-domiciled fund. May require Malta sealed letters of administration before redemption.
• Jersey — Lloyds account (R0.5m). Jersey "grant of probate" required if estate above £10k threshold.
• New Zealand — Pacific Padel shares (R1.8m, but inside WO so company-owned, not personal — easier).

Cost impact: Each jurisdiction adds 6–18 months to settlement timeline + 2-5% in legal/admin fees + estate duty equivalent in some jurisdictions (UK has IHT at 40% above £325k for non-residents holding UK situs assets — but Jersey is outside UK IHT scope).

The current single-point-of-failure: Elmar is sole director of both Werda Operations and Werda Investments. If you become incapacitated (stroke, coma, dementia) or die, neither company can sign anything until either:

• The Master of the High Court appoints a curator / executor (months), or
• The shareholders convene and appoint a new director.

Meanwhile: no bank instructions, no tax filings, no SARB reports, no payments.

The fix — alternate director or s66(8) deputy:

• Alternate director: A pre-named person who automatically steps in when the primary director can't act. Requires MOI amendment.
• Section 66(8) deputy: Less formal — director can authorise another person to act on the director's behalf for specified matters.

Either way, name Nicolette + Stephan + a third independent (e.g. Albertus Marais) so single-incident scenarios don't paralyse the entire structure.

The statutory tariff: Section 51(1)(b) of the Administration of Estates Act + Master's tariff regulation. Executor entitled to 3.5% of gross asset value of the estate + VAT (15%) = effective ~4.025% on assets that flow through the estate.

Your will (clause 3.1): Caps executor fee at the statutory tariff. Good — without this clause some professional executors negotiate above-tariff. Clause 3.2 also reduces fee on insurance proceeds to 10% of normal — also good.

The size impact:

• First death @ R178m gross = ~R7.2m executor fee + R0.16m Master's fees
• Second death today @ R177m = ~R7.1m
• Second death +10y @ R382m = ~R15.4m

The capacity question: Your will names Nicolette as primary executor → brother Johan as backup. Brother as backup for a R200m+ multi-jurisdiction estate is under-resourced. Worth considering a professional co-executor (CDH or trust company) — fee impact is marginal because tariff caps it.

The core problem: At second death the family will need ~R49m today, ~R65m at +5y, ~R87m at +10y in cash within 12 months to settle estate duty + CGT + executor fees. The Class B shares (R109.5m face value, fixed by MOI) cannot be turned into cash without a buy-sell mechanism + funding. PSG Global funds (R59m) are offshore and need Malta probate first. Working through the WO portfolio means dividends tax + CGT layers.

Why life cover is the cleanest solution:

• Pays out within weeks of death certificate (vs months/years for offshore probate).
• If structured correctly (policy owned by trust, beneficiary = trust, not estate), proceeds bypass dutiable estate AND executor fee.
• Funds the buy-sell — the trust uses the cover proceeds to redeem Class B from the deceased's estate at the MOI-fixed price. Estate gets cash to pay duty. Class B is extinguished.
• Premium cost is small relative to the duty saving (typical R40m cover at age 50, healthy = R10k–R30k/month).

The revised sizing logic (R30m–R45m):

• Today's second-death cash need: R49.6m. Pension nomination R19.4m bypasses + R3.7m personal cash → ~R26m liquid before forced sales. Cover R30m bridges the gap.
• +5y need: R65m. Pension + cash + some WO portfolio liquidation. Cover R35m bridges.
• +10y need: R87m. Cover R45m bridges, rely on offshore + WO realisation for the rest.
• This is materially smaller than the R40m–R60m sizing in the initial Marietjie memo — that range was based on Class B growing at 8% (which the MOI prevents).