The decision in Roshcon (Pty) Ltd v Anchor Auto Body Builders CC & Others [2014] ZASCA 40 was handed down on 31 March 2014 as a unanimous judgment of the Supreme Court of Appeal. In fact, two separate judgments were given, but all five justices of appeal on the bench concurred in both judgments.
The significance of the decision is that, to a considerable extent, it clears the fog of confusion that has prevailed since the decision of the Supreme Court of Appeal three years ago in CSARS v NWK Limited 2011 (2) SA 67 (SCA) in regard to what constitutes a simulated transaction.
Anecdotal evidence suggests that, since the NWK judgment, SARS has readily – arguably too readily – attacked bona fide tax-saving arrangements as being transactions that satisfy the criteria for simulation laid down in that case.
The confusion that followed the NWK decision
The decision in NWK professed to lay down criteria in terms of which a transaction would be regarded as simulated (thereby entitling SARS to give effect to the underlying real transaction) that seemed to contradict criteria that had been laid down in the locus classicus, namely, the decision in Zandberg v Van Zyl 1910 AD 302 and had been accepted for a century. In that case, Innes CJ said that –
“Not infrequently, however (either to secure some advantage which otherwise the law would not give, or to escape some disability which otherwise the law would impose), the parties to a transaction endeavour to conceal its real character. They call it by a name, or give it a shape, intended not to express but to disguise its true nature. And when a Court is asked to decide any rights under such an agreement, it can only do so by giving effect to what the transaction really is: not what in form it purports to be. The maxim then applies plus valet quod agitur quam quod simulate concipitur. But the words of the rule indicate its limitations. 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.”
In NWK Lewis JA said (at para [55]) that – “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. . . . 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.”
The latter dictum seemed to suggest that, even if a transaction was ‘genuine’ in the sense that the parties intended to give effect to it according to its tenor and even if it lacked what Innes CJ in Zandberg v van Zyl called a ‘ real intention, definitely ascertainable, that differed from the simulated intention’, the transaction would nonetheless be characterised as simulated if its only purpose was to evade tax.
The difficulty created by the dictum of Lewis JA was that it stood in stark contradiction to the dictum of Watermeyer CJ in Randles, Brothers & Hudson Ltd v Commissioner of
Customs 1941 AD 349 at 395:
“I wish to draw particular attention to the words ‘a real intention, definitely ascertainable, which differs from the simulated intention’, because they indicate clearly what the learned Judge meant by a ‘disguised’ transaction. A transaction is not necessarily a disguised one because it is devised for the purpose of evading the prohibition in the Act or avoiding liability for the tax imposed by it. A transaction devised for that purpose, if the parties honestly intend it to have effect according to its tenor, is interpreted by the Courts according to its tenor, and then the only question is whether, so interpreted, it falls within or without the prohibition or tax.”
The earlier decision made it clear that the intention to avoid tax or a law was not, of itself, sufficient ground for holding that a transaction is a simulation.
The Court’s interpretation of the NWK judgment
The significance of the judgment of Wallis JA’s judgment in Roshcon (in which all the other judges concurred) is, firstly, that he rejected the proposition that the Supreme
Court of Appeal in NWK had taken the law in a ‘new direction’, and secondly, that he affirmed the traditional view that the essence of a simulated transaction is that it involves a disguise.
Wallis JA affirmed that, as was held in Dadoo Ltd v Krugersdorp Municipal Council 1920 AD 530 at 548 –
“… the law permits people to arrange their contractual or business affairs so as to obtain a benefit for themselves that a different arrangement would not permit or so as to avoid a prohibition that the law imposes.’’
The judgment continued (at para [27]) that –
“Whether a particular transaction is a simulated transaction is therefore a question of its genuineness. If it is genuine the court will give effect to it and, if not, the court will give effect to the underlying transaction that it conceals. And whether it is genuine will depend on a consideration of all the facts and circumstances surrounding the transaction.”
In the reported income tax cases on this issue, said Wallis JA, the parties had sought to take advantage of the complexities of income tax legislation to obtain a reduction in their overall liability for income tax.
The feature common to all such stratagems, he said, was that they –
“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.”
The problem in the NWK case was the contention that –
“provided the parties intended to take all the steps provided for in the contractual documents, in other words to jump through the contractual hoops as a matter of form, the court could not find that the transaction was simulated.”
Wallis JA said that this is the aspect that Lewis JA was dealing with when she said that –
“… the test to determine simulation cannot simply be whether there is an intention to give effect to a contract in accordance with its terms. . . . . 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.”
This dictum had been interpreted in some circles as meaning that any contractual arrangement that enabled the parties to avoid tax would be regarded as simulated.
However, said Wallis JA, this was not what Lewis JA intended.
It is impossible to determine whether there was simulation on the basis of generalizations
Wallis JA said (at para [36]) that it was impossible to make generalisations and that each agreement under scrutiny should be viewed in its particular commercial context to determine whether it was simulated.
Thus, the judgment concluded (at [37]) that –
“… the notion that NWK transforms our law in relation to simulated transactions, or requires more of a court faced with a contention that a transaction is simulated than a careful analysis of all matters surrounding the transaction, including its commercial
purpose, if any, is incorrect.”
The correct position, said Wallis JA (at para [37]) is that the court examines each particular transaction as a whole, including all surrounding circumstances, any unusual features, and the manner in which the parties intend to implement it, before determining whether the transaction was simulated.
A welcome clarification
This judgment will be welcomed for its affirmation that the essence of a simulated transaction is not that it has the effect of avoiding tax or that it lacks a ‘commercial purpose’ – as some passages in the NWK judgment seem to suggest – but rather that the essence of a simulated transaction (see para [27] of the Roshcon judgment) is that it is not genuine and that, conversely, if it is genuine, the court will give effect to it.
This reaffirmation of fundamental principles will be particularly welcomed in the financial services sector, where it is common for complex financial arrangements to be entered into for a variety of reasons, including fiscal advantages. It will be easier to defend such arrangements against attack by SARS now that the essential question is whether the
transaction is genuine rather than whether it has a commercial purpose. Merely because a transaction is structured in such a way that it avoids or reduces tax that would have been incurred if it had been structured in a different way does not suffice to make it a simulated transaction.