Driving divorce: Navigating trusts as a vehicle for matrimonial asset division

June 1st, 2024

Picture source: Getty/iStock

While a family trust is often created for legitimate estate planning purposes, it is so that it ‘may turn out to be a handy vehicle for “divorce planning” because transferring property to the trust reduces the scope of a person’s estate and consequently limits (or even eliminates) any claim the person’s spouse may have against him or her on divorce’ (J Heaton ‘The extension of mediation, and piercing the trust veneer on divorce in South Africa’ in B Atkin (ed) The International Survey of Family Law (Bristol: Jordan Publishing Limited 2015) at 313).

The last two decades have seen an increase in claims that the assets of an (allegedly) abused trust should be considered for the purposes of dividing matrimonial property at divorce. Initially, such claims were confined to divorces in which a redistribution order could be sought – in terms of s 7(3) – (6) of the Divorce Act 70 of 1979 – because the spouses were married with complete separation prior to the introduction of the accrual system via the Matrimonial Property Act 88 of 1984 (the MPA). In 2006, the Supreme Court of Appeal (SCA) handed down the leading judgment of Badenhorst v Badenhorst 2006 (2) SA 255 (SCA) in which it held that ‘[t]o succeed in a claim that trust assets be included in the estate of one of the parties to a marriage there needs to be evidence that such party controlled the trust and but for the trust would have acquired and owned the assets in his own name’ (my italics). Such control – as evinced by the trust deed and the manner of trust administration during the marriage – was required to be of the de facto (factual) as opposed to the de iure (legal) variety. The Badenhorst court concluded that the facts revealed a ‘classic instance’ of the respondent’s ‘full control of … the trust [which he had used] as a vehicle for his business activities’ while ‘[paying] scant regard to the difference between trust assets and his own assets’. This led to the conclusion that the value of the trust’s assets be considered in computing the appellant’s redistribution claim (Badenhorst at paras 9 – 13, emphasis supplied). While the application of this so-called ‘control test’ to marriages that qualify for a redistribution order is settled, the position relating to marriages subject to the accrual system is less certain, as is indicated by two recent conflicting judgments of the SCA, namely PAF v SCF 2022 (6) SA 162 (SCA) and MJK and Others v IIK 2023 (2) SA 158 (SCA). This article attempts to shed some light on this issue, as well as on another crucial development occasioned by the PAF case.

Attacking trusts in marriages involving the accrual system – the legal position prior to the PAF case

The judgment in REM v VM 2017 (3) SA 371 (SCA) removed the doubt which had persisted as to whether the (value of) the assets of an abused trust could in principle be taken into account for the purposes of an accrual claim at divorce (see the case law from my previous article, ‘Trust assets and accrual claims at divorce: The SCA opens the door’ 2017 (Aug) DR 22 at 23). Swain JA formulated the following test for this to occur, namely that ‘[t]he respondent had to prove that the appellant transferred personal assets to these trusts and dealt with them as if they were assets of these trusts with the fraudulent or dishonest purpose of avoiding his obligation to properly account to the respondent for the accrual of his estate and thereby evade payment of what was due to the respondent, in accordance with her accrual claim’ (at para 20, my italics). The emphasised words have been interpreted as requiring two separate steps, namely first proving, by virtue of the ‘control test’ in the Badenhorst case, that the trust is the alter ego of a spouse, because he or she has dealt with its assets ‘as if’ they were trust assets. Secondly, it must be proved that this alter ego trust was fraudulently or dishonestly employed to evade the obligation to render a proper accounting of that spouse’s accrual (F du Toit, BS Smith and A van der Linde Fundamentals of South African Trust Law 2ed (Durban: LexisNexis 2023) at 185). The PAF case, however, calls this approach into question.

The PAF case

In this matter it was not disputed that the respondent’s estate showed the smaller accrual, thus entitling her to an accrual claim against the applicant on their pending divorce. However, approximately three weeks before the divorce trial was due to commence, the applicant created a trust in the British Virgin Islands with his brother (a solicitor) as sole trustee and his daughter as the sole beneficiary, ostensibly to provide for her maintenance needs. A day later he donated £ 115 000 (then the equivalent of R 2,2 million) to the trust. The court a quo found that this transaction was fraudulently intended to deprive the respondent of her rightful accrual claim. The SCA agreed, based on the following key conclusions:

  • The High Court could only have ordered the value of the donation to form part of the applicant’s estate in consequence of a piercing of the trust’s veneer resulting from his abuse of the trust form (at para 30).
  • Although a spouse’s accrual claim arises only when the marriage is dissolved (see s 3(1) of the MPA), both spouses have a ‘protectable contingent right’ while the marriage subsists ‘which the law will protect in circumstances of irregularity and a lack of bona fides’ (my italics). An inaccurate reflection of a spouse’s true accrual defeats the rationale behind the legislature’s provision of the accrual system in the first place (at para 35).
  • It follows that a court is empowered to conduct an in-depth factual inquiry to assess allegations of trust form abuse at divorce. This is because piercing the veneer of a trust occurs in consequence of the exercising of a common-law power – derived from the principles applicable to piercing the corporate veil in company law – that exists independently of any statutory discretion conferred by the Divorce Act or the MPA (at paras 32-37).
  • Crucially, the ‘control test’ in the Badenhorst case is not limited to marriages contemplated in s 7(3) of the Divorce Act; it may be applied to ‘among others, marriages subject to an accrual system’. This is because both a redistribution order and the accrual system have ‘as their objective, equitable and fair patrimonial consequences of a marriage’ (at para 42).
  • However, the ‘control test’ could not be applied in casu, because the applicant exercised no control over the trust. But the PAF court noted that the Badenhorst court had contemplated this eventuality in a proviso that it had added to this test, to the effect that trust assets could still be considered despite being placed beyond the control of a spouse if that spouse had done so with the intention of frustrating the other spouse’s patrimonial claim. A conspectus of the evidence in PAF indicated that the applicant had done precisely this, given: (i) the timing of the trust’s creation and the donation to it; (ii) that it was, without explanation, founded in a foreign jurisdiction, thus complicating the respondent’s ability to challenge the status quo; (iii) that he had not consulted with the respondent on the trust’s creation despite their history of mutual decision-making on important financial matters and the detrimental effect of the donation on her contingent right to accrual sharing; and (iv) that there was no immediate need to provide for his daughter’s maintenance by way of a trust, as this issue would be fully canvassed at divorce (at paras 45 – 53).

The impact of the PAF case

The PAF case now makes it clear that a court has the discretion to consider all the evidence before it to adjudicate an accrual claim involving an allegation that the trust veneer should be pierced due to trust form abuse. The case broke further new ground by showing that instances in which a court will be prepared to do so are no longer confined to the typical alter ego scenario (ie, where the aggrieved spouse ‘challenges the control of the trust concerned’ (WT and Others v KT 2015 (3) SA 574 (SCA) at footnote 5, my italics)), but may even occur where the errant spouse exercises neither de iure nor de facto control over the trust, provided of course that the evidence supports this conclusion. This constitutes a significant step towards combating unscrupulous ‘divorce planning’.

Equally significantly, the PAF case provides valuable perspectives on the typical alter ego scenario, especially because of the court’s conclusion that the ‘control test’ in the Badenhorst case is also applicable to marriages subject to the accrual system. Indeed, this conclusion calls the accuracy of the test laid down for trust-veneer-piercing in such marriages by the REM case (quoted above) into question.

To explain: In Honoré’s South African Law of Trusts (E Cameron, MJ de Waal and P Solomon, 6ed (Cape Town: Juta 2018) ‘Honoré’) at 311, 312) the eminent authors identify at least two types of trust form abuse. The first is ‘when the trust form is used in a dishonest or unconscionable manner to evade a liability, or avoid an obligation’ (Van Zyl and Another NNO v Kaye NO and Others 2014 (4) SA 452 (WCC) at para 22, endorsed by the REM case at para 17). However, the authors postulate that neither dishonesty nor unconscionability should necessarily be seen as prerequisites, because trustees’ failure to respect the so-called ‘core idea’ of the trust (by blurring the separation between their control over the trust and the beneficiaries’ enjoyment of trust benefits (Land and Agricultural Bank of South Africa v Parker and Others 2005 (2) SA 77 (SCA) at paras 19 and 22)), and their failure to perform the ‘basic principles of trust administration’ should also suffice. (These ‘basic principles’ are that trustees must exercise an independent discretion, except in matters of law; they must adhere to the terms of the trust deed as far as is legally permitted; and they must act in accordance with the fiduciary obligation imposed by s 9 of the Trust Property Control Act 57 of 1988 (see Honoré at 305)). It is significant that the authors identify the Badenhorst case as ‘possibly the leading case’ on this second form of abuse (Honoré at 312).

However, if one considers the test laid down in the accrual context by the REM case, it becomes clear that the court required both these forms of trust abuse to be proved by the respondent. It is my suggestion that this test is too arduous, and that the PAF case shows that the REM court should merely have required the respondent to comply with the ‘control test’ in Badenhorst in order to succeed in her claim (see BS Smith ‘Prescient precedent: PAF v SCF (SCA) and a new paradigm for testing whether a trust has been abused to manipulate a spouse’s accrual (or redistribution) liability at divorce’ (2023) 140.4 SALJ 697 at 705-711 (Smith 2023a) for a detailed explanation of what follows). This is because, as the PAF court correctly held, where the accrual system is applicable, both spouses have a ‘protectable contingent right’ that exists during the marriage and a duty to properly account for their respective accruals at divorce (the PAF case at para 35, read with s 7 of the MPA). Where the application of the ‘control test’ shows that a spouse’s true accrual has been manipulated by means of a trust, this will (not may) affect either spouse’s accrual liability or entitlement (as the case may be). As such, compliance with the ‘control test’ must automatically entail that a spouse’s obligation to properly account for his or her accrual has been evaded. Cadit quaestio. (It should be noted that matters are different in the case of a redistribution claim because there a spouse must, over and above complying with the ‘control test’, also satisfy the preconditions set by s 7(4) of the Divorce Act). The conclusion that the ‘control test’ should suffice in the accrual context is also borne out by case law such as LW v CW (WCC) (unreported case no 12866/2014, 26-8-2020) (Salie-Hlophe, J) that was decided after – and without applying – the test laid down in the REM case. (On this case, see H Harnett and H Kruuse ‘Proper financial planning and divorce: Failure to consider the antenuptial contract’ (2023) TSAR 822 at 830, who share my view that the REM test is ‘more onerous’ than the ‘control test’ in the Badenhorst case).

Whether the preceding argument will receive judicial imprimatur is unclear. This is because merely five weeks after the PAF case was decided, a different coram of the SCA decided the matter of MJK and Others v IIK 2023 (2) SA 158 (SCA). Although the latter court considered the PAF case, it held that the ‘control test’ in Badenhorst was only applicable to redistribution orders (at paras 35-39). I have critiqued this case elsewhere (see B Smith ‘Trust assets and accrual claims in divorce proceedings: The Supreme Court of Appeal muddies the waters by misconstruing its own precedent – MJK v IIK 2023 (2) SA 158 (SCA)’ (2023) 86 THRHR 491 (Smith 2023b)). For present purposes it will suffice to say that the MJK court made no reference to the PAF court’s express finding to the contrary. This superficial engagement with precedent draws the accuracy of the MJK court’s conclusions into question, which should, therefore, be treated with circumspection.


The PAF court’s perspicacious engagement with the ‘control test’ (and its proviso), coupled with its granting of a wide discretion to future courts faced with allegations of trust form abuse in the accrual context, constitutes a significant gain for spouses adversely affected by unscrupulous ‘divorce planning’. The case also prompts the formulation of a consolidated test for piercing a trust’s veneer that can be applied in both the redistribution and accrual settings (see Smith 2023a 711-714 for a suggested test). On the other hand, the MJK court’s failure to properly engage with the PAF case amounts to a loss (Smith 2023b 502-503). Practitioners should be alerted to these diverging SCA judgments on the scope of the ‘control test’. Although I have expressed my clear preference for the approach in the PAF case, the correct legal position remains uncertain.

Prof Bradley Smith BCom LLD (UFS) is a Senior Researcher at the Independent Institute of Education’s Varsity College and an Extraordinary Professor of Private Law at the University of the Free State.

This article was first published in De Rebus in 2024 (June) DR 16.