Court name
Supreme Court
Case name
Ellis and Others v Noabeb
Media neutral citation
[2019] NASC 582
Headnote and holding:

The appellants are the trustees of a trust known as Eldo Trust (the trust). In their capacities as such and after the respondent had failed to enter an appearance to defend the action brought against him, moved an application for judgment by default in the High Court. The appellants claimed from the respondent payment of certain amounts owing as a result of a written agreement entered into between the parties. The appellants also prayed for an order declaring that the respondent forfeit all monies already paid to the appellants in fulfilment of the agreement and in addition that the respondent be ejected from the immovable property ‘owned’ by the trust. 

The parties entered into an agreement in terms of which the respondent would, amongst others, become the sole trustee and only beneficiary of the trust that has as its asset an immovable residential property. The parties agreed that as consideration for becoming the beneficial owner of the property and the sole trustee as well as beneficiary of the trust on ‘the consummation date’, the respondent would pay in instalments an amount of money towards the purchase price of the property. In the meantime, he would pay occupational rental in exchange for undisturbed occupation and possession of the immovable property. On the consummation date, the appellants would resign and appoint the respondent as the sole trustee and sole beneficiary of the trust.

The High Court raised concerns over the net effect of the agreement.  It directed the legal practitioners for the appellants to present argument on the question whether the agreement was valid or lawful and/or enforceable. This the High Court questioned in light of the consideration that the agreement had contemplated the respondent becoming the sole trustee and sole beneficiary of the trust. The appellants, were also directed to deal with two further questions; firstly, whether the agreement concluded by the parties was a simulated transaction devised to avoid the obligation to pay transfer duty or stamp duty on the immovable property and secondly, whether it was not against the good morals of society for the appellants to claim forfeiture of all monies paid by the respondent in fulfilment of the agreement and the outstanding balance as well as his ejectment from the property.  After hearing argument, the High Court dismissed the application and declared the agreement null and void and of no force and effect.

Before the oral arguments were heard, the appellants abandoned the forfeiture claim as well as the order seeking the ejectment of the respondent from the property and as such, the High Court was not called upon to determine this issue. On the first question, wstructure contemplated by the agreement the contractual scheme entered into by the parties, viewed as a whole, constituted a thus void ab initio on the basis that it was in fraudem legis of a statute imposing transfer duty on an immovable property. The appellants appealed to the Supreme Court against the order of the High Court declaring the agreement to be null and void and of no force and effect.

The Supreme Court held that a situation where the sole trustee would become the sole beneficiary of the trust while not invalidating the trust, creates an undesirable state of affairs as the enjoyment and control of the trust property are not functionally separated. The court went on to state that it was the separation element that served to secure diligence and independence of judgment on the part of the trustee in dealing with the trust property.

The court further held that at the hearing of the appeal the trust was properly constituted with the three appellants being the trustees and also some of the beneficiaries. It is only on the ‘consummation date’, being the date upon which the respondent would have complied with all his financial obligations towards the trustees, that the composition would change with the respondent then becoming the sole trustee and beneficiary. It is only at this stage and onwards that what the court has described as an ‘undesirable situation’ would take effect.

The court held that the second amendment agreement once executed would not delete the clause in the existing trust deed that makes provision for the appointment of an additional trustee or trustees, to a maximum of five trustees. The court also held that although the statute governing trusts in Namibia does not empower the Master to appoint trustees in the absence of provision in the trust instrument, our courts still have powers to restrict and prevent the abuse of a trust form. Exercising those powers, the court directed the respondent to appoint an independent trustee to close the gap of what could result in an undesirable situation. 

The court restated the principles applicable to the test for a simulated transaction. It stated that the applicable test required a consideration of whether the parties intended the agreement to have the legal effect apparent from its terms.

It was held that the payment of the consideration was not only in exchange for appointment as the sole trustee and sole beneficiary of the trust, but also for the purpose of ‘the new beneficiary to have occupation, use and enjoyment of the immovable property’. The parties’ intention was evidently to structure the transaction in such a way that payment of transfer duty on the sale and transfer of immovable property is avoided, which is a legitimate arrangement that the law currently allows.

The court further held that there was no evidence that the agreement was not what it purported to be. There was no suggestion of the agreement not being a genuine transaction to transfer beneficial ownership of the residential property to the respondent. The appeal succeeded and the order of the Court below was set aside.

Judge
Shivute CJ
Mainga JA
Hoff JA

APPEAL JUDGMENT


SHIVUTE CJ (MAINGA JA and HOFF JA concurring):

Background

[1]        The three appellants are the trustees for the time being of a trust known as Eldo Trust (the trust). They have appealed in their capacities - by virtue of being trustees - as owners of an immovable property, being Erf 1325, Hofmeyer Street, Khomasdal, Extension 5, Windhoek (the property).

[2]        The trust was founded on 8 February 2006 with the principal object of identifying and mobilising resources for charitable purposes. Initially the trust had two trustees namely; one D Meyer and a certain E Bahr. The beneficiaries were the Clinic Clowns (Pty) Ltd (‘a private entity providing clowning services free of                    charge’) and ‘such other charitable institutions, persons or groups of persons as determined by the trustees’.

[3]        On 20 April 2010, the trust deed was amended. The amendment substituted the first trustees for the three appellants. The amendment further made the appellants beneficiaries of the trust together with ‘such current and future employees of the trustees as they may determine from time to time’.

[4]        During September 2012, the appellants and the respondent – an employee of theirs - entered into a written agreement, the material terms of which may be summarised as follows:

(a) As consideration for becoming the sole trustee and only beneficiary of the trust as well as for the purpose of purchasing an immovable property from the trust, the respondent would pay N$815 000 in instalments, into the trust account of Ellis and Partners Legal Practitioners for the benefit of the trustees, to                which amount or the remaining balance the trustees would be entitled ‘on the consummation date’;

(b) ‘The consummation date’ was defined in the agreement as meaning ‘the date upon which the new trustee has paid all the instalments, together with any interest due and owing as well as occupational interest’;

(c)  The respondent would honour the consideration by way of instalments as set out in clause 5.2 of the agreement, namely N$200 000 on the effective date, N$205 000 on the first instalment date, N$205 000 on the second instalment date and N$205 000 on the third instalment date;

(d) As from ‘the effective date’ – being 1 May 2012 - prior to the date of registration and against payment of occupational rental of N$5000 per month, the respondent, as new beneficiary would enjoy undisturbed occupation and possession of the property;

(e) On the consummation date, the respondent would be appointed as the sole trustee by the appellants, whereupon the appellants would resign as trustees, leaving the respondent as the sole trustee and sole beneficiary of the trust;

(f)  In the event of the respondent committing a breach of the agreement, the appellants would be entitled to invoke any of their rights as per the agreement after affording the respondent seven days written notice to remedy such breach. This included the right to cancel the agreement; the right to repossess the  n            property and the forfeiture of all monies paid by the respondent in terms of the agreement.

[5]        The appellants alleged in the summons dated 15 October 2013 that pursuant to the agreement, they had given undisturbed occupation and possession of the property to the respondent, but that the respondent was in breach of the agreement in that he had failed to pay the full sum of the agreed consideration. The appellants further averred that the respondent was also in breach of the agreement by failing to pay occupational rent as of 3 May 2013, leaving a balance of N$24 500. It was also alleged that the appellants had cancelled the agreement by written notice dated 27 August 2013 addressed to the respondent. Accordingly, the appellants claimed from the respondent payment of the amounts of N$234 500 and N$24 500, being the alleged balance on the consideration and occupational rental, respectively. In addition, the appellants initially claimed forfeiture of all payments made by the respondent, confirmation of the cancellation and an order ejecting the respondent from the property.

[6]        Subsequent to the letter cancelling the agreement and the service of the summons, the respondent made further payments towards the purchase price, which payments were accepted by the appellants. On 17 February 2014, the appellants’ legal practitioners wrote to the respondent informing him that in light of the payments he had made, the appellants would waive their claims for his ejectment from the premises and the forfeiture of the money paid pursuant to the agreement. They insisted though that the respondent pay the outstanding amounts in respect of the consideration and for occupational rent as agreed.

[7]        The following day the respondent’s legal practitioners addressed a letter to the appellants’ legal practitioners requesting to be furnished with the outstanding amounts referred to in the appellants’ letter of 17 February 2014. Counsel for the appellants submitted that the correspondence between the parties manifests the desire on the part of the parties to continue with the agreement and that on that understanding the appellants had therefore abandoned their cancellation of the agreement. Counsel contended in the alternative that ‘there is consensus between the parties that the obligations as originally agreed upon will be given effect to’.

[8]        The respondent failed to enter an appearance to defend the action and consequently the appellants sought a default judgment against him.

[9]        The High Court mero motu directed counsel for the appellants to address argument on the question whether the agreement relied upon was valid or lawful and/or enforceable in light of the consideration that it had contemplated the respondent becoming the sole trustee and sole beneficiary of the trust. The court called for arguments on the enforceability of the agreement on the further ground that the agreement appeared to have been simulated and in fraudem legis aiming at avoiding the payment of transfer duty or stamp duty.

[10]      The High Court also directed the appellants’ legal practitioners to address it on the question whether it was not contra bonos mores for the appellants to claim forfeiture of all monies paid by the respondent pursuant to the agreement and the outstanding balance as well as his ejectment from the property.

[11]      After hearing argument, the High Court dismissed the application for judgment by default and declared the agreement null and void and of no force and effect. It is against this order that the appellants are now appealing. The respondent took no part in the proceedings in the High Court or in this court. The court thus only heard argument from the appellants’ counsel.

Issues before the High Court

[12]      In determining whether the agreement entered into by the parties was valid and enforceable, the High Court had to decide the following issues:

(a) Whether the agreement between the parties negated the whole form and nature of trusts?

(b) Whether the agreement concluded by the parties was simulated and in fraudem legis, aimed at simply evading the payment of transfer duty?

[13]      The third issue the court was initially going to decide, namely whether it was contra bonos mores for the appellants to claim forfeiture of all monies paid by the respondent pursuant to the agreement as well as his ejectment from the property, was ultimately abandoned as it was conceded in the High Court that such an approach would be contra bonos mores (contrary to the good morals of society). It is therefore not surprising that the appellants, in the letter of 17 February 2014 referred to above, abandoned the claims of the ejectment of the respondent from the house as well as the forfeiture of the money paid by the respondent in terms of the agreement.

Reasoning of the High Court

[14]      In deciding the first issue, the court analysed the legal nature of a trust. It approached the question from the situation where a person may be a sole trustee and sole beneficiary. The court acknowledged that although this may happen in some instances, it may also lead to an abuse of the trust form. The court stated that to safeguard against potential abuse, the control of the trust property must be kept separate from its enjoyment, with the control being exercised on behalf of another. The court acknowledged the principle that although a trustee is a person holding or administering the trust for someone else, in law, there is nothing preventing a trustee from also becoming a beneficiary.

[15]      With reference to various leading cases, the court proceeded to distinguish between instances where a trust was created initially with only one trustee, who was also the sole beneficiary, on the one hand and where the sole trustee became the sole beneficiary after the establishment of the trust, on the other. The court went on to state that in the former instance, no trust comes into existence whereas the latter situation does not invalidate the trust.

[16]      With reference to the judgment of the Supreme Court of Appeal of South Africa in Land and Agricultural Bank of South Africa v Parker & others 2005 (2) SA 77 (SCA), the court concluded that a sole trustee cannot become the sole beneficiary of a trust. It stated that such a situation would embody an identity of interests that is inimical to the trust idea, and no trust would come into existence. The court reasoned that it was the separation element that served to secure diligence and independence of judgment on the part of the trustee in dealing with the trust property. The court found that the agreement entered into by the parties in this matter was fashioned in such a way that it negated the whole notion of trusts and was thus offensive to the nature of trusts.

[17]      The court further held that the agreement was a clear instance where a trust had been debased and abused to achieve the transfer in ownership of immovable property without the consequence attendant upon such a transaction to pay transfer duty.

[18]      Turning to the second question, the court held that the contractual scheme constituted a transaction simulated to disguise the real agreement between the parties. The court reasoned further that the scheme devised by the parties was meant to achieve an objective different from the one set out in the agreement, namely the sale and transfer of the immovable property. In support of its conclusion on this aspect, the court adopted the approach set out by this court in Strauss & another v Labuschagne[1] and the Supreme Court of Appeal of South Africa in Commissioner for the South African Revenue Service v NWK Ltd.[2] The court then concluded that the contractual scheme in this matter was devised to avoid paying transfer duty on the property.[3] This was done under the guise of transferring control over the trust.

[19]      The court observed that although counsel for the appellants submitted during oral argument that the effect of the agreement was that the respondent would purchase the property from the appellants for a consideration, the transaction itself was structured in an entirely different way, mainly to subvert the payment of certain statutory amounts.

[20]      In dismissing the application for default judgment, the court concluded that the contractual scheme entered into by the parties, viewed as a whole, constituted a simulated transaction and thus void ab initio on the basis that it was in fraudem legis. I digress to point out that the High Court did not have sight of the trust deed in this case as such a document did not form part of the record. We specifically requested to be furnished with a copy of the trust deed, which I found useful in understanding the context of the transaction and which assisted in deciding the appeal.

Issues for determination on appeal

[21]      The issues to be considered and decided in this appeal are essentially the same points of law the High Court decided, namely whether the agreement between the parties negates the form and nature of trusts and whether the agreement concluded by the parties was simulated and in fraudem legis, simply aimed at avoiding the payment of transfer duty.

[22]      Mr Tötemeyer, who argued the appeal together with Mr Dicks, pointed out at the outset that his clients were not appealing against the refusal by the court below to grant a default judgment. As noted above, counsel submitted that subsequent events to the letter of demand and service of summons on the respondent indicated that the parties were desirous of giving effect to the agreement as originally concluded. The appeal was rather against the order of the High Court embodied in paragraph (b) that reads ‘the agreement entered into between the parties attached to the particulars of claim is declared null and void and of no force and effect’.

[23]      In light of counsel’s stance on the issues the court is called upon to decide on appeal, I enquired from counsel whether it was necessary for the court to decide the second question of whether the agreement was in fraudem legis in the event that the first question in the appeal was answered in the affirmative. Counsel submitted that it was necessary to do so for clarity and legal certainty, particularly on the question whether the NWK approach should be followed by our courts as decided by the High Court. I agree that it may be necessary to decide this important issue in light of the facts that it was decided by the High Court and argument on it was heard by this court. It is also a matter that is likely to arise in the future given that undoubtedly many transactions in the country are structured in a manner similar to the present agreement.

[24]      I will thus proceed to deal with the issues on appeal as identified above, starting with the question whether the agreement debases the notion of trusts.

Does the agreement between the parties negate the form and nature of trusts?

Concept of trust

[25]      In dealing with this issue, it is important to address briefly the core elements of a trust. In Namibia, the Trust Moneys Protection Act[4] and the common law regulate trusts. The common law on trusts is the same in both Namibia and South Africa. However, the applicable pieces of legislation in the two countries differ markedly. In South Africa, the Trust Property Control Act[5] repealed the Trust Moneys Protection Act which, as just mentioned, is still applicable in Namibia. It is, to be noted at the outset that the applicable legislation in Namibia is woefully antiquated, and fails to comprehensively regulate modern trusts. In my respectful view, it is a matter that requires urgent legislative intervention. The position obtaining in South Africa is vastly different with substantial legislative reforms in respect of trusts having occurred in that country. There the Master and the courts have been given extensive powers to restrict and prevent abuse of a trust form. The Master in particular has been given wide powers to appoint trustees in the absence of provision in the trust instrument, and ‘to appoint any person as co-trustee of a serving trustee where he considers it ‘desirable’, notwithstanding the provisions of the trust instrument’.[6]

[26]      The concept of ‘trusts’ is widely regarded as an invention of the English jurisprudence.[7] Honoré’s South African Law of Trust defines a trust ‘as a legal institution in which a person, the trustee, subject to public supervision, holds or administers property separately from his or her own, for the benefit of another person or persons, the beneficiaries, or for the furtherance of a charitable or other purpose’.[8] Accordingly, in a wider sense a trust will come into existence when one person has handed over or is bound to hand over the control of his property (the founder) to another (the trustee), which property is to be administered by the trustee for the benefit of someone other than the trustee or in pursuance of an impersonal object.[9] The core concept of trusts relates primarily to the separation between the control of trust property and its enjoyment.[10]

[27]      Trusts are widely used for various reasons. They have been mainly popular for their perceived tax benefit implications. Writing on the popularity and the wide usage of trusts, the learned authors of Honoré’s South African Law of Trusts note that a trust ‘is an all purpose institution, more flexible and wide ranging. . . ’[11] They also observe that trusts can be used to serve almost an indefinite variety of objects.[12] Their wide usage thus ranges from avoiding the dissipation of assets,[13] avoiding conflicts of interests, keeping assets from the reach of creditors or family,[14] to minimising tax liability amongst other uses. It is this wide usage of trusts that also tends to attract abuse. As such, many cases that end up in court relate primarily to allegations of abuse of trusts. It is in this context that one must understand and approach the current matter.

Counsel’s submissions

[28]      In his submissions concerning the issue of whether a person can be a sole trustee and at the same time also a sole beneficiary of a trust, counsel essentialy cited the same cases he relied on in the court a quo. Counsel firstly relied on dicta made at paras 31 and 32 of Groeschke v Trustee, Groeschke Family Trust & others,[15] judgment.

[29]      Counsel argued that the finding by the High Court that the dicta in those two paragraphs were obiter was incorrect. He contended that there was no evidence in the Groeschke matter that the ‘alternative trustee’ administered the trust together with the deceased. In the submission of counsel, the deceased effectively administered the trust as its sole trustee and beneficiary.

Application of the legal principles

[30]      It seems to me that Groeshke goes further than the submissions made above. In that case, the court also stated the following in paras 36 and 37:

‘[36]     Conceivably, the appointment of the second respondent as "alternative trustee" in paragraph 1 of the resolution presents a dilemma only at face value. The choice of wording by the deceased in paragraph 1 was unfortunate, but not fatal. As pointed out above, the intention of the deceased as expressed in the resolution was to remove the applicant as the capital and income beneficiary of the trust and to appoint himself as beneficiary, but without himself resigning as a trustee.

[37]      In accordance with clause 4.8 of the deed of trust an "alternative trustee" is appointed to serve "in the place and stead of another trustee . . . during that trustee's absence or disability to act as a trustee". However, nothing in the resolution permits one to infer that the deceased wanted to appoint the second respondent because he intended to absent himself, or because he was somehow disabled to act as a trustee. On the contrary, the inference to be drawn from the wording of the resolution as read with the deed is that the deceased intended to continue to serve as a trustee whilst at the same time to enjoy the benefits of a beneficiary. Therefore, he could only have intended to appoint the second respondent as an additional trustee, not as an alternative trustee. If I am obliged, in case of doubt as to his intentions, to incline to a construction of the amended deed that would render the amending transaction by the deceased operative rather than inoperative, then the circumstances relevant to the resolution (and logic) therefore dictate that I should construe the phrase "alternative trustee" in paragraph 1 of the resolution so as to mean that the deceased intended to appoint the second respondent as an "additional trustee".’ [Emphasis added].

[31]      Groeschke hinged, amongst others, on the proper interpretation of s 4(2) of the Trust Property Control Act, 1988 and on a question of a sole trustee of an existing trust ‘caused by circumstances to become its sole beneficiary’ and not on the effect of an agreement designed to make the sole trustee of a trust also its sole beneficiary. The court therefore in my view was not confronted with facts similar to those in this appeal as argued by counsel.  Moreover, as the court pointed out in the Groeschke matter, the only logical conclusion that could be reached in that case was that the deceased could only have intended to appoint the second respondent as an additional trustee, not as an alternative trustee. As to the position of a sole trustee becoming the sole beneficiary of a trust, the court reasoned that such an eventuality would not invalidate the trust because s 7 of the (South African) Trust Control Act gives the Master, irrespective of the terms of the trust instrument, the power to appoint co-trustees to any serving trustee.

[32]      On a proper construction of the present agreement, the parties intended to have, from the consummation date, only one trustee who is also the sole beneficiary. This much is clear from clause 6.1.2 of the agreement. Our Trust Moneys Protection Act 34 of 1934 does not have provisions similar to s 7 of the South African Trust Control Act. This, of course, does not mean that our courts are powerless to prevent the abuse of a trust form.[16] This aspect will be dealt with in detail later in the judgment. It becomes apparent that the facts in the present matter are distinguishable from those in the Groeschke case. Thus the High Court’s finding to that effect is undoubtedly correct. In the Parker matter on which the appellants also relied, Cameron JA pointed out, at para 19, that a situation where a sole trustee is also the sole beneficiary would embody an identity of interests that is inimical to the trust idea and no trust would come into existence.

[33]      It should be emphasized, however, that presently the trust is properly constituted with the three appellants being the trustees and some of the beneficiaries. It is only on the ‘consummation date’, being the date upon which the respondent would have complied with all his financial obligations towards the trustees, that the composition would change with the respondent then becoming the sole trustee and beneficiary. Does the change in the composition of the trust offend the nature of trusts? This appears to strike at the core of the issue at hand. Given the circumstances of this matter, can one conclude that the agreement between the parties offends the nature of trusts? Oakley[17] opines as follows regarding this matter:

‘Once a trust has been validly constituted, it cannot thereafter fail for want of a trustee since anyone in whom the trust property comes to be vested, other than a bona fide purchaser for value of a legal estate in the property without notice of the interests of the beneficiaries (or, in the case of registered land, the statutory equivalent) will be bound by that trust and so will be a trustee of the property.’

[34]      The above statement was written with a particular focus on the position of the English law and the applicable trust legislation. However, the principle appears to accord with the principles set out in Parker. Once an inter vivos trust has been validly created and is operational, it cannot be invalidated by a subsequent agreement that seeks to change, for example, the composition of trustees and beneficiaries in a fashion that is at odds with the well-established principles of trusts.

[35]      The main reason for the above approach seems to be rooted on principles of business efficacy. Trusts venture into business transactions, borrow money, and act as surety for certain businesses. As a result of their flexibility or elasticity, trusts are used for invariably many business transactions, depending on the nature of the trust in question. Their transactions can have serious implications on other business entities or bona fide individuals who may contract with the trust.

[36]      When a trust has fewer than the required number of trustees by its trust deed, it appears (based on the Parker case) that its actions are a nullity. The nullity relates to the actions or decisions taken on behalf of the trust and not the trust itself. It relates to the capacity of the trust to act and not its existence. This seems to be the essence of what was decided in the Parker matter. In the relevant parts of the Parker judgment, Cameron JA had this to say in this respect:

‘[11]     It follows that a provision requiring that a specified minimum number of trustees must hold office is a capacity-defining condition. It lays down a prerequisite that must be fulfilled before the trust estate can be bound. When fewer trustees than the number specified are in office, the trust suffers from an incapacity that precludes action on its behalf.

[12]      This is not to say that the trust ceases to exist. Nor is it to say that the trust obligation falls away. Counsel for the bank cited passages from Honoré’s establishing that a trust will not be allowed to fail for want of a trustee, and that the administration of a trust proceeds even when not all the trustees can be appointed in the precise manner envisaged in the trust deed. This is to confuse the existence of the rights and obligations that constitute the trust estate with the question whether and in what manner the trust estate can be bound. It is axiomatic that the trust obligation exists even when there is no trustee to carry it out. The Court or the Master will where necessary appoint a trustee to perform the trust. But it does not follow that a sub-minimum of trustees can bind a trust.’ (Emphasis added and reference to authorities omitted).

[37]      It would then appear that an agreement purporting to amend the trust deed to have only one trustee, who is also the sole beneficiary, does not necessarily invalidate the trust, unless such an agreement has the effect of creating a new, albeit invalid trust. It is not necessary for the purposes of this appeal to decide whether the agreement has the effect of creating a new trust.

[38]      However, a situation where the sole trustee would become the sole beneficiary of the trust while not invalidating the trust, creates an undesirable state of affairs where, to borrow from Cameron JA in Parker, the enjoyment and control are not functionally separate. According to clause 6.1 of the agreement entered into between the parties, on the consummation date, the parties will execute a second amendment agreement in terms of which they would appoint the respondent the sole trustee. This will be done by deleting clause 4.5 of the Trust Deed that reads:

‘Save for periods pending an appointment, there shall at all times be no less than two (2) trustees. If the number of trustees falls below two (2) and the vacancy is not filled within two (2) months after it has occurred, then the remaining trustee of the trust shall appoint a trustee to fill such vacancy. Until the filling of such vacancy, the remaining trustee shall, if necessary, be entitled and empowered to act alone.’

[39]      This clause will be deleted in its entirety and will be replaced by the following clause:

‘There shall at all times be at least 1 (one) trustee. In the event that there are more than 1 (one) trustee and a trustee vacates the office, the remaining trustee(s) shall be entitled to appoint a trustee to fill the vacancy. Until the filling of the vacancy, if  necessary, the remaining trustee(s) shall, if necessary, be entitled and empowered to act alone.’

[40]      Moreover, the second amendment agreement to be executed will no doubt embody clause 4.2 of the Trust Deed as there is no indication that this clause will be deleted in the proposed second amendment. Clause 4.2 of the Trust Deed makes provision for appointment of additional trustee or trustees, to a maximum of five trustees. This clause arguably serves as the saving grace for the trust to be purged of the undesirable situation as the power to appoint additional trustees makes it possible for the court, in the exercise of its powers to prevent the abuse of the trust form, to direct the trustee, after the consummation date, to appoint an independent trustee to close the gap of the undesirable situation.

Whether the transaction was in fraudem legis?

[41]      I now turn to the second question of whether the agreement was in fraudem legis. The court a quo found that the agreement was in fraudem legis of the Transfer Duty Act as the parties sought to avoid the payment of transfer duty. The question that should be decided, differently put, is whether the agreement relied on by the parties was a simulated transaction designed to avoid the payment of the applicable tax.

[42]      Counsel submitted that if the agreement would achieve an avoidance of the provisions of the Transfer Duty Act, it would be in the nature of things perfectly fine as the law countenances this. He also submitted that the intention of the parties was to sell the property to the new trustee who would also become the sole beneficiary of the trust.

[43]      This court had occasions to consider transactions which are said to be in fraudem legis. In Strauss & another v Labuschagne[18] for example, the court had to decide whether the transaction between the parties constituted an agreement in fraudem legis of the provisions of the Agricultural (Commercial) Land Reform Act 6 of 1995. O’Regan AJA, writing for the Court in that case, referred in para 45 to a passage in Zandberg v Van Zyl,[19] where Innes J set out the test for determining whether a transaction was a simulation or a genuine agreement by looking at the intention of the parties. At page 309, Innes J stated the position as follows:

‘The Court must be satisfied that there is a real intention, definitely ascertainable, which differs from the simulated intention. For if the parties in fact mean that a contract shall have effect in accordance with its tenor, the circumstances that the same object might have been attained in another way will not necessarily make the arrangement other than it purports to be. The enquiry, therefore, is in each case one of fact, for the right solution of which no general rule can be laid down.’

[44]      O’Regan AJA also referred to Commissioner of Customs and Excise v Randles, Brothers and Hudson Ltd.[20] In that case, in discussing the phrase ‘disguised’ transaction used by Innes J in the Zandberg matter, Watermeyer JA pointed out that a transaction is not necessarily disguised because it is devised to evade a prohibition of the law or to avoid tax liability. The learned judge of appeal proceeded to elaborate on this aspect from page 395 to 396 as follows:

‘A disguised transaction in the sense in which the words are used above is something different. In essence, it is a dishonest transaction: dishonest, inasmuch as the parties to it do not really intend it to have, inter partes, the legal effect which its terms convey to the outside world. The purpose of the disguise is to deceive by concealing what is the real agreement or transaction between the parties. The parties wish to hide the fact that their real agreement or transaction falls within the prohibition or subject to tax, and so they dress it up in a guise which conveys the impression that it is outside of the prohibition or not subject to the tax. Such a transaction is said to be in fraudem legis, and is interpreted by the Courts in accordance with what is found to be the real agreement or transaction between the parties.

Of course, before the Court can find that a transaction is in fraudem legis in the above sense, it must be satisfied that there is some unexpressed agreement or tacit understanding between the parties.’

[45]      A determination of whether the transaction in this case is a simulated transaction requires (as pointed out in the Strauss case) a consideration of whether the parties intended the agreement to have the legal effect apparent from its terms. It is by no means an easy undertaking. Indeed, this difficulty was also acknowledged by Innes CJ in the Dadoo & others v Krugersdorp Municipal Council 1920 AD 530 at 544.

[46]      In Dadoo’s case, Innes CJ went on to make the following seminal remarks  at 548:

‘But an Act thus construed may nevertheless be evaded; parties may genuinely arrange their transactions so as to remain outside its provisions. Such a procedure is, in the nature of things, perfectly legitimate. There is nothing in the authorities, as I understand them, to forbid it. Nor can it be rendered illegitimate by the mere fact that the parties intend to avoid the operation of the law, and that the selected course is as convenient in its result as another which would have brought them within it. An attempted evasion, however, may proceed on other lines. The transaction contemplated may in truth be within the provisions of the statute, but the parties may call it by a name or cloak it in a guise, calculated to escape those provisions. Such a transaction would be in fraudem legis; the Court would strip off its form and disclose its real nature, and the law would operate.’

[47]      The dicta above have been cited with approval in many cases over the years, including by this court in the Strauss matter. Counsel submitted that the High Court erred in relying for its decision on the judgment in the NWK case, which placed a slightly different emphasis on the principles relating to simulations. It was contended that NWK does not detract from the well-established Dadoo principles nor did it lay down a principle that every avoidance scheme, which has as its sole or principal purpose the avoidance of tax, is to be treated as a simulated transaction.

[48]      This court in the Strauss case expressly left open the question whether ‘the difference of emphasis’ in NWK was now the new approach and followed the earlier authorities on the issue of simulation. O’Regan AJA at para [50] said the following on this score:

‘NWK, concerned as it was with the question of tax avoidance and evasion, shifts the focus of analysis from the question whether the parties intend to give effect to the actual terms of the impugned transaction, to the question of whether the impugned transaction makes any commercial sense at all. It is not necessary now to decide whether the difference of emphasis that underlies the approach in NWK should be adopted in Namibia. For the purposes of this case, it is appropriate merely to consider, on the approach established in Dadoos’s case and followed since, whether on the factual record before it, the respondent has shown that the contractual scheme was more likely than not a simulated transaction.’

[49]      The High Court reasoned that on both an application of the traditional approach adopted in Strauss and the slightly different principles adumbrated in NWK, the scheme devised by the parties in this case was a sham designed to disguise the underlying contract in order to avoid the paying of transfer duty in terms of the Transfer Duty Act consequent upon the transfer of the immovable property. The court concluded that the question of whether or not the scheme devised by the parties amounted to a transaction simulated to disguise the real agreement between the parties to avoid the obligation to pay transfer duty was similar to the one decided in the NWK case. Therefore, the NWK approach should be adopted by our courts in cases of transactions allegedly simulated to avoid or evade the payment of tax.

[50]      It is thus necessary to have a closer look at the approach adopted in NWK and to examine how the principles set therein were received in subsequent cases. NWK concerned the question whether a loan agreement that was part of a series of transactions entered into between NWK and a bank as well as subsidiaries of the bank, all designed to disguise the true nature of the transaction between NWK and the bank, to enable NWK to avoid or reduce its liability for tax, was a genuine sales agreement or a simulation. The parties had concluded interrelated contracts in terms of which a subsidiary of the bank had lent R96,4 million to NWK, repayable upon the delivery of a specified quantity of maize over a period of five years. NWK then claimed interest deductions on the capital of R96,4 million.

[51]      The Commissioner for Inland Revenue argued that the true substance of the transactions was that NWK had borrowed only R50 million and that a series of grain transactions between the parties were simulated to enable NWK to claim increased tax deduction. In agreeing with the contention advanced by the Commissioner, the court first reviewed leading cases discussing the principles that one may arrange one's affairs so as to 'remain outside the provisions of a particular statute', and that a court would not be deceived by the form of a transaction but that it should instead examine its true nature and substance.

[52]      Writing for the court, Lewis JA observed that while the principles may not necessarily be in conflict, decided cases did not consistently approach the issue of a party's intention in concluding a contract.[21]  The learned judge of appeal then made the following key observations at para 55 that appear to have become a source of controversy in subsequent cases:

‘In my view the test to determine simulation cannot simply be whether there is an intention to give effect to a contract in accordance with its terms. Invariably where parties structure a transaction to achieve an objective other than the one ostensibly achieved they will intend to give effect to the transaction on the terms agreed. The test should thus go further, and require an examination of the commercial sense of the transaction: of its real substance and purpose. If the purpose of the transaction is only to achieve an object that allows the evasion of tax, or of a peremptory law, then it will be regarded as simulated. And the mere fact that parties do perform in terms of the contract does not show that it is not simulated: the charade of performance is generally meant to give credence to their simulation.’ (Emphasis added.)

[53]      In respect of the transaction under consideration, the court held that the transaction was ‘artificially engineered and specifically designed to conceal’ the true loan amount.[22]  In coming to this conclusion, the court examined the ‘true nature and substance’ of the agreement and came to the conclusion that in substance the parties could not have intended a loan.

[54]      In subsequent decisions of South African courts and in legal commentary, NWK was explained, criticized and qualified. In Bosch & another v Commissioner,  South African Revenue Services[23], for example, Davis J, writing for the Full Bench of the Western Cape High Court, observed that in respect of the views expressed in the dictum, beyond the finding that the parties had not created genuine rights and obligations but had constructed a loan for R94,5 million as opposed to R50 million purely to enable the taxpayer to obtain a greater tax benefit, there was nothing in the judgment supporting the argument that the NWK reasoning had been intended to alter the ‘century old’ principles regarding the determination of a simulated transaction for the purposes of tax. The learned judge added:

‘[84]     In my view, the key paragraph relied upon by the respondent in the NWK case needs to be read within this context so as to ensure that the body of precedent is read coherently rather than reading NWK as being an unexplained rupture for more than a century of jurisprudence.’

[55]      Davis J went on to explain that the intention of the remarks in para 55 of NWK was to give a court pointers in which it should direct the enquiry in determining whether a transaction is in fraudem legis. He stated that in determining this question, a court should examine the real commercial sense of the transaction in question, adding that if there is no commercial rationale, then the avoidance of tax as the sole purpose of the transaction would be a powerful justification for approaching the transactions in the manner done in NWK.

[56]      Davis J also referred to the analysis of the NWK judgment by Eddie Broomberg in an article titled ‘NWK and Founders Hill’ in The Taxpayer 60 (2011, October). In this article, Broomberg was reported to have commented on the NWK judgment in the following critical terms:

‘[T]he stare decisis rule should not have been ignored by the court because there is no apparent ground for asserting that all the preceding judgements and cases involving alleged simulated transactions were wrong in applying the principle laid down in Zandberg v Van Zyl, Randles Brothers et al. in any case, it was not necessary to flout the stare decisis rule because the legislature had already closed the door on the mischief which the court was seeking to avert. Moreover, it was beyond the power of the court to adopt the NKW rule, since to do so was to usurp the function of the Legislature, and finally, the adoption of the NWK approach could result in discrimination.’

[57]      While concurring with Davis J’s reasoning and the order he had proposed in the Bosch matter, Waglay J expressed reservations about Davis J’s interpretation of the NWK judgment. He took issue with Davis J’s conclusion at para 85 that ‘without an express declaration to that effect, NWK should be interpreted to fit within a century of established principle rather than constituting a dramatic rupture.’ In Waglay J’s view on this score, NWK was ‘a dramatic reversal of what has been a consistent view of what constitutes a simulated transaction.’[24] He added that considered in its entirety, NWK ‘does in fact lay down the rule that any transaction which has as its aim tax avoidance will be regarded as a simulated transaction irrespective of the fact the transaction is for all purposes a genuine transaction’.

[58]      Waglay J also agreed with Broomberg’s view that the NWK judgment seeks to hold previous judgments involving alleged simulated transactions wrong. The learned judge was of the view that NWK did not authoritatively mark a departure from the existing (or the then existing) precedent for the following reasons that he gave in para 103:

‘Before one is bound to a precedent-setting judgment and is obliged to follow it, the judgment must be clear and unequivocal, it must be plain, unmistakable and explicit in its rejection of previous judgments which it seeks to reverse, and it must be applicable to the facts in the matter before the court confronted with its possible application. While I do not believe that the reversal must be express, the reasoning should demonstrate a departure from previous binding judgments. NWK does not in my view do so. It does not provide any reasons why the judgments aptly dealt with by Davis J in paras [79] to [83] are no longer good law. This is further compounded by the troubled equivalence in the judgment of the phrases “tax avoidance” and “tax evasion” — two very distinct concepts.’

[59]      Waglay J then concluded his brief judgment with the following cutting remarks in para 105:

‘Having regard to the above, NWK cannot be read to serve as a precedent in this case where evasion is not the issue. In any event, any transaction which has as its purpose tax evasion is unlawful as tax evasion constitutes a criminal offence in terms of the Income Tax Act.  NWK cannot therefore be authority for setting aside a transaction as simulated by reason of being a vehicle for tax evasion as this is automatic in terms of the law. On the other hand, if the words “evasion of tax” are to be substituted with “avoidance of tax”, then the dictum goes against the accepted practice in our income tax law which permits transactions aimed at tax avoidance.  Furthermore, the confusion[25] created by the judgment militates against it serving as a precedent binding upon the lower courts.’

[60]      In Roshcon (Pty) Ltd v Anchor Auto Body Builders CC & others 2014 (4) SA 319 (SCA), the Supreme Court of Appeal of South Africa felt constrained to deal with an aspect of the case, which the court characterised as ‘a misconception regarding the proper approach to simulated transactions’. Writing for the unanimous court, Wallis JA rejected the notion that the NWK judgment had developed or clarified the test on simulated transactions laid down in previous judgments, thereby taking South African law in a new direction in this regard. He traced the foundation of South African law on simulated transactions to locus classicus such as Zandberg v Van Zyl; Dadoo Ltd & others v Krugersdorp Municipal Council; Commissioner of Customs and Excise v Randles Bros & Hudson Ltd, and noted that nothing said in subsequent judgments of that court dealing with simulated transactions had altered the original principles set out in those cases in any way or purported to do so. The learned judge of appeal went on to observe in paras 32 and 33 as follows:

‘However, in a number of them dealing with income tax, the courts have been called upon to apply these principles in a different context. The earlier cases dealt with cases of agreements being dressed up in a particular form where the underlying intention of the parties was inconsistent with that form. In the income tax cases a different problem arises. [T]he parties seek to take advantage of the complexities of income tax legislation in order to obtain a reduction in their overall liability for income tax. There are various mechanisms for doing this, but they all involve taking straightforward commercial transactions and adding complex additional elements designed solely for the purpose of claiming increased or additional deductions from taxable income, or allowances provided for in the legislation. The feature of those that have been treated as simulated transactions by the courts is that the additional elements add nothing of value to the underlying transaction and are very often self-cancelling. . .’

[61]      Wallis JA explained further that the analysis of the transactions in NWK showed that a number of ‘unrealistic and self-cancelling features’ had been added to what was otherwise a straightforward loan. They served no commercial purpose and were included solely for purpose of concealing the nature of the loan and inflating the deductions that NWK could make against its taxable income. It was in that context that the unrealistic elements were stripped off the transactions to reveal the real agreement.

[62]      The limitation of the application of the dictum in para 55 of the NWK judgment was reiterated by Wallis JA when the judgment of the Full Bench in the Bosch matter was taken on appeal with the leave of the Supreme Court of Appeal. In para 40 of the judgment,[26] the learned judge dealt with the contention by the Commissioner of Inland Revenue based on the approach set out in para 55 of NWK as follows:

‘That submission involved a misunderstanding of the judgment in NWK, as was pointed out in Roshcon. There I stressed that simulation is a question of the genuineness of the transaction under consideration. If it is genuine then it is not simulated, and if it is simulated then it is a dishonest transaction, whatever the motives of those who concluded the transaction. The true position is that 'the court examines the transaction as a whole, including all surrounding circumstances, any unusual features of the transaction and the manner in which the parties intend to implement it, before determining in any particular case whether a transaction is simulated'.  Among those features will be the income tax consequences of the transaction. Tax evasion is of course impermissible and therefore, if a transaction is simulated, it may amount to tax evasion. But there is nothing impermissible about arranging one's affairs so as to minimise one's tax liability, in other words, in tax avoidance. If the revenue authorities regard any particular form of tax avoidance as undesirable they are free to amend the Act, as occurs annually, to close anything they regard as a loophole.’

[63]      It is thus clear from the post-NWK cases that other than setting the bar high or developing a new standard for the test of simulated transactions, all that NWK does is to require a court faced with the question of whether a transaction in question is genuine or simulated, to examine the transaction as a whole, taking into account unusual features including the question whether or not it makes commercial sense. It cannot be overstressed that the commercial rationale for the transaction in question is, of course, only a part of the considerations that come into play in determining whether the transaction is genuine or not. NWK has therefore not replaced the principles set out in the original authorities on the point surveyed by the Supreme Court of Appeal of South Africa in Roschcon above.

[64]      It is also important in this respect to distinguish between tax avoidance and tax evasion. Tax avoidance refers to a situation in which a taxpayer, within the provisions of the tax statute, arranges his or her affairs so that the tax obligations are minimized or completely avoided.[27]  On the other hand, tax evasion is an egregious situation where a taxpayer unlawfully arranges his affairs in a way that attempts to escape from tax liability which he ought to pay.[28] It refers to all those activities deliberately undertaken to free oneself from tax that the law charges upon income.[29]

[65]      Tax evasion is an offence and it is something a court of law can obviously not countenance. As to tax avoidance on the other hand, parties may genuinely arrange their affairs so as to remain outside the provision of any particular statutory provision. That this is so was stated as early as the 1920s in Dadoo’s case at 548 that such a procedure is, in the nature of things, perfectly legitimate and cannot be rendered illegitimate merely because the parties intended to avoid the operation of the law. This must, as already noted, be distinguished from an attempted tax evasion, which is a different and serious matter. Indeed, that there is no equity about tax or a presumption about tax is not only a principle of South African law, but one that has been recognised also under English law. For example, in Commissioners of Inland Revenue v His Grace the Duke of Westminster 1936 AC 1, Lord Russel of Kilowen at 24-25 quoted from an earlier case, Partington v Attorney-General (1869) LR 4 HL 100 at 122 where it was said:

‘[A]s I understand it the principle of all fiscal legislation is this: if the person sought to be taxed comes within the letter of the law he must be taxed, however great the hardship may appear to the judicial mind to be. On the other hand, if the Crown, seeking to recover the tax, cannot bring the subject within the letter of the law, the subject is free, however apparently within the spirit of the law the case might otherwise appear to be.’  

[66]      It is against these principles that the agreement between the parties should be considered. The question for decision then is whether the agreement was a genuine transaction or a simulated transaction designed to evade the application of the Transfer Duty Act?  One of the considerations relevant to determining whether a contract is simulated or not is whether it seems ’anomalous’ or has an ’air of unreality.’ O’Regan in Strauss summed up this consideration as follows:

‘In determining as a matter of fact, whether a particular contractual arrangement is simulated or not, the courts have considered whether the arrangement has an “air of unreality”, “accords with reality” or contains anomalies or is “startling”. Where an arrangement seems anomalous or unreal, it is more likely that a court will conclude that it is a simulated arrangement disguising a different but tacit agreement.’ (Reference to authorities omitted).

[67]      In an attempt to answer the question whether the transaction was genuine or not, it is necessary to revisit some of the pertinent clauses of the trust deed and the agreement. In terms of the transaction, (a) existing trustees will resign and the new trustee would become the sole trustee; (b) the new trustee would become the sole beneficiary of the trust, and (c) the beneficiary was to have occupation, use and enjoyment of the immovable property. As noted above, the new trustee was also expected to perform by paying instalments as set out in para [4] of this judgment.

[68]      It is, however, important to emphasise that it is only on ‘the consummation date’ that the respondent was to be appointed the sole trustee and beneficiary of the trust. As earlier noted, ‘the consummation date’ is the date upon which the respondent would have paid the instalments, interest thereon as well as occupational interest for the purchase of the house in full.

[69]      Furthermore, from the effective date being 1 May 2012, the new trustee was to pay to the original trustees an occupational rental of N$5000 per month. Such payments, including the instalments were to be used, to the extent required, to settle any outstanding amount on the mortgage bond. Counsel submitted that the effect of the agreement was to enable the new trustee to purchase the immovable property. This submission appears to have been borne out by the pertinent clauses in the agreement. The object of the agreement is set out in clause 3. Clauses 3.1.1, 3.1.2 and 3.1.3 respectively provided that the parties were desirous of ‘the new trustee to become the sole trustee; ‘the new trustee to substitute and replace the beneficiary’, and ‘the new beneficiary to have occupation, use and enjoyment of the immovable property’.

[70]      For the attainment of these objects, the parties agreed that after the consummation date, the new trustee would be appointed a sole trustee and beneficiary. As consideration for becoming sole trustee and sole beneficiary and for the purpose of having occupation, use and enjoyment of the immovable property, he would pay the consideration by instalments into the trust account of Ellis Shilengundwa Incorporated. The phrase ‘original trustees’ is defined in the agreement as meaning the three appellants.

[71]      It is apparent from the agreement that the payment of the consideration was not only in exchange for appointment as the sole trustee and sole beneficiary of the trust, but also for the purpose of ‘the new beneficiary to have occupation, use and enjoyment of the immovable property.’ The parties’ intention was evidently to structure the transaction in such a way that payment of transfer duty on the sale and transfer of immovable property is avoided, but this seems to me to be a legitimate arrangement that as the law now stands, appears to countenance. As was pointed out by Wallis JA in the Bosch matter above, if the revenue authorities regard any particular form of tax avoidance undesirable, they are at liberty to have the Act regulating tax amended to close what they may perceive to be a gap in resorting to transactions similar to the one under consideration. 

[72]      There is no evidence that the agreement is not what it purports to be. There is no suggestion of the agreement not being a genuine transaction to transfer beneficial ownership of the residential property to the respondent. There is no air of unreality about the transaction except for the fact that the parties structured it in a way that takes advantage of the tax law as they understand it to be. There does not appear to me to be any simulation. 

[73]      There is, of course, that lingering undesirability of the conflation of interests in that at the consummation date, there would be no functional separation of ownership and enjoyment as the trustee would become the sole beneficiary and trustee. But as noted earlier, this does not invalidate the trust. As was pointed out in the Parker judgment, courts do have the power and may invoke it to ensure that the trust form is not abused. Although the trust in question is not a commercial trust per se, this power has to be invoked to ensure that the trust functions in accordance with the principle of accountability and reasonable expectations of outsiders who may deal with it.[30]

[74]      It is true that in terms of clause 4.2 of the agreement, the trustee has discretion to appoint additional trustees. As he is endowed with discretion in this regard, he may decide to maintain the status a quo. Should he elect not to exercise his discretion, there is every possibility that the identity of interest may be perpetuated. This would be highly undesirable. It is therefore necessary that he should be directed to exercise his discretion upon the consummation date.

[75]      In light of the conclusion arrived at on the appellants’ principal argument, it is not necessary to deal with their alternative contention regarding what the parties referred to as the ‘unwinding provisions’ in clause 15 of the agreement.

[76]      In the result the appeal must succeed.

Order 

[77]   The following order is made:

(a) The appeal succeeds.

(b) The order of the High Court embodied in paragraph (b) of that court’s order is  set aside and the following order is substituted therefor:

      ‘The agreement concluded by the parties was neither simulated nor was it in fraudem legis.’

(c)  The trustee is directed to appoint, within thirty (30) days from ‘the consummation date,’ as that phrase has been defined in the memorandum of agreement entered into between the parties, an additional independent trustee so as to ensure that there shall at all times be not less than two (2) trustees constituting         the Eldo Trust.

(d) The trustee is further directed to furnish the registrar of this court with proof of the appointment of the additional trustee within seven (7) days of such appointment.

(e) No order as to costs is made.

 

 

_________________

SHIVUTE CJ

 

 

________________

MAINGA JA

 

 

________________

HOFF JA

 

 

APPEARANCES

APPELLANTS:                                       R Tötemeyer (with him G Dicks)

                                                               Instructed by Ellis Shilengudwa Inc

RESPONDENT:                                     No appearance 


[1] 2012 (2) NR 460 (SC).

[2] 2011 (2) (SA) 67 (SCA).

[3] Contrary to the Transfer Duty Act 14 of 1993.

[4] Act 34 of 1934.

[5] Act 57 of 1988.

[6] Section 7(2) of the Trust Property Control Act. See also the Parker case at para 34.

[7] Cameron et al  Honoré’s South African Law of Trusts. 5 ed (2002 Juta and Co. at 24).

[8] Op. cit. at  page1, para 1.

[9] Id. At page 3, para 1.

[10] Id. At 17, para 6.

[11] Op. cit. at 14, para 5

[12] Id.

[13] Ex Parte Estate Graaff 1947 (4) SA 496 (C).

[14] Ex Parte Easton NO 1948 (2) SA 535 (C).

[15] 2013 (3) SA 254 (GSJ).

[16] Parker, at para 34.

[17] Oakley, A. (1998). Parker and Mellows: The Modern Law of Trusts 7 ed. Sweet & Maxwell: London.

[18] 2012 (2) NR 460 (SC).

[21] Para 45.

[22] At 74C-D.

[23] 2013 (5)SA 130 (WCC).

[24] At para 100.

[25] Waglay J explained in a footnote that the confusion related to the fact that Davis J held that NWK did not intend to depart from the established precedent, while Broomberg contended that it did. 

[26] Commissioner, South African Revenue Services v Bosch & another 2015 (2) SA 174 (SCA).

[27] South African Income Tax Guide, 2015/2016, 18.8

[28] Id.

[29] Id.

[30] Cf. Parker at para 37.