by Ernest Mazansky, Director, Werksmans Tax (Proprietary) Limited 

Introduction

The general anti-avoidance rule – or the GAAR as it is “fondly” referred to in the tax world – has had a lot of media attention in the past short while. The GAAR has been the feature of tax legislation for many decades, though it has gone through various iterations. The latest version, which came into force in 2006, was designed to counter a number of principles that had been handed down by the courts that allowed taxpayers to escape the grasp of that legislation.

Interestingly it has taken some twenty years before some substantive and detailed decisions have been handed down by the courts on the “new” version.

Of late, there have been three important decisions:

• First, there was the decision of the SCA in the Erasmus case where the taxpayer was successful, and so I understand it, the Constitutional Court declined to hear the appeal. Although that related to the situation where SARS had invoked the GAAR, essentially the dispute was more procedural than substantive.
• The second recent decision was that the ABSA case where the Constitutional Court handed down a far ranging decision that actually went beyond the dispute that it was adjudicating. In the course of the judgment they have, in certain respects, effectively overturned 100 years of jurisprudence in South Africa, including decisions that were strongly influenced by a decision of the House of Lords where Lord Denning handed down the judgment. As many might know, Lord Denning was one of the most revered judges in Britain’s judicial history.
• The latest case is a decision of the Cape Tax Court in a case where six appeals, ie relating to six appellants, were heard concurrently (and, if it does not go on appeal, will likely be cited as company AF Proprietary Limited & Others v C:SARS). In this case the Tax Court dealt with each and every aspect of the GAAR and found against the taxpayer in regard to all of the aspects.

The GAAR

It is not my intention in this article to give an exposition of the GAAR or an analysis of any of the abovementioned court decisions. Rather my focus here is to place the GAAR in its proper context in the Income Tax Act, 1962 (the Act) and, indeed, where other fiscal legislation also has a GAAR of their own, an example being the VAT Act (though it is not as wide-ranging as the GAAR in the Act).

Essentially, for the GAAR to apply, four criteria must be met as follows:

• There must be a “transaction, operation, scheme, agreement or understanding (whether enforceable or not)” – now referred to as an “arrangement”.
• The arrangement must lead to an avoidance, deferral or reduction of tax – now referred to as a “tax benefit”.
• The sole or main purpose of the arrangement must be to obtain the tax benefit.
• Finally, the arrangement must include at least one of various so-called tainting elements, a very common one being that it was undertaken in an abnormal way or created rights and obligations that would not normally be created between persons dealing at arm’s length.

It is only when all four tests have been met that the GAAR can be invoked by SARS. Unlike in most aspects of the Act, SARS has the burden of proof as to whether there is an arrangement, that there is a tax benefit and that at least one tainting element is present. Having discharged its burden that there is a tax benefit there is a rebuttable presumption that the arrangement was entered into solely or mainly to obtain the tax benefit, and it is here where the onus lies on the taxpayer to prove otherwise.

The GAAR in context

The GAAR is designed to upset tax avoidance schemes that the Act does not otherwise deal with in its other provisions. As is evident from the four requirements above there is no specific rule in the Act that the taxpayer might have infringed. It is set out in broad and general terms to create a set of guidelines, so to speak, and if one falls within those guidelines the GAAR becomes applicable. What is very important to note is that the GAAR, by its nature, is only invoked by SARS when there is no other provision in the Act that SARS can rely on to tax the taxpayer. If the arrangement involves receiving amounts that were not taxable, there would be no specific provision in the Act to prevent this or unwind it; or there would be no specific anti-avoidance measure to counter its practice (and the Act is replete with these specific anti-avoidance measures). Put differently, the GAAR is only resorted to if the taxpayer and its arrangements are fully compliant with the provisions of the Act and the various interpretations in respect thereof. There would be nothing in the taxpayer’s actions or in the taxpayer’s income tax returns that SARS could point to and say “this is wrong and you have underpaid tax for the following reason”. And, interestingly, this is why historically SARS could never impose penalties when the GAAR was invoked, because there was absolutely nothing wrong with the tax return filed (this changed some years ago when the penalty provisions in the Tax Administration Act, 2011 were specifically amended to make a 75% penalty compulsory when the GAAR was successfully invoked, though a number of commentators believe that it can still be challenged).

And what is important to note is that the GAAR does not apply automatically – it has to be specifically invoked by SARS. So it is not a case, as with specific anti-avoidance rules, where a taxpayer can try and undertake a transaction but the Act sets out in words of one syllable that this will not work. If the taxpayer ignores the anti-avoidance provision and prepares its return on that basis, then the return is wrong and SARS can tax the taxpayer and impose penalties in the normal course.

Not so with the GAAR.  Before that can happen a number of steps have to be taken. Shortly stated these include –

• initially identifying the possibility of an impermissible avoidance arrangement;
• seeking documentation and explanations from the taxpayer;
• then sending to SARS a specific notice in terms of section 80J of the Act (not surprisingly, colloquially referred to as a section 80J notice) stating that they believe that the GAAR can apply and setting out SARS’s reasons therefor. The taxpayer is then given the opportunity to submit reasons to SARS why the GAAR should not apply.
• Only thereafter is SARS permitted to invoke the GAAR and raise the relevant assessment.

It is therefore evident that there is nothing inherently unlawful where a taxpayer undertakes an arrangement that has the effect of reducing an income tax liability. And nothing will happen in relation to that taxpayer unless and until SARS invokes the GAAR. Until then the taxpayer’s affairs can rightfully be said to be in apple pie order.

A balancing strategy

When a taxpayer seeks to do something that is clearly not permitted in the Act, we as advisors are very quick to explain this to a client and tell the client why it cannot be successfully done. Oftentimes, of course, there might be a different way to achieve the same end result without breaching the provisions of the Act, and that is where we get to deliver good news to the client instead of bad.

And if we have told the client that the proposal is wrong and the client goes ahead and does it anyway, then clearly that client is taking a risk, because if SARS does attack it there is very little that the client can say to defend itself.

Not so with the GAAR. In the case where the GAAR is a possibility we obviously bring that risk to the client’s attention, and then it is up to the client to decide whether or not to take the risk. And if the client wishes us to give an opinion in the proposed transaction or structure it would be remiss of us not also to include an analysis of the GAAR provisions in relation to the facts. But what is different there from the previous circumstance is that the client is secure in the knowledge that he or she is not breaking the law when undertaking the relevant arrangement. As I said, completing a return on this basis means that the return is 100% correct and the client’s affairs remain in apple pie order. It is only when SARS takes the necessary steps to invoke the GAAR that the picture changes.

And truth be told, invoking the GAAR and seeing it through until the end is very resource-intensive for both SARS and the taxpayer, and generally history has shown that the outcomes of a GAAR dispute when adjudicated by the courts are far murkier when assessing the prospects of success, compared to other types of tax disputes.

It follows, therefore, that per force, SARS has to pick and choose very carefully which cases it wishes to tax based on the GAAR. It is not as simple and straightforward as, for example, taking the view that a particular item of expenditure does not meet the tests for deductibility, and then let the dispute runs its course. It is very much more than that. And this inevitably means that the risks of being attacked under the GAAR are, from a purely practical point of view, much smaller than in relation to being attacked under any other provisions of the Act. And in this regard it is not dissimilar from transfer pricing disputes, which are also resource-intensive. There is simply a limit to how many of these SARS (or any tax authority anywhere in the world) can take on at any point in time, so that it is necessary to choose the ones they are most likely to win, and where the amounts of tax make the effort worthwhile.

This, of course, does not mean that taxpayers should be reckless or cavalier in their approach. Thee consequence of the GAAR being invoked means that the taxpayer is in for a very uncomfortable ride. But taxpayers, especially in business, are by nature risk-takers as opposed to risk-avoiders. It then becomes an ordinary business risk to decide whether or not to go ahead and face the possibility of being attacked under the GAAR, but at the same time knowing full well that no unlawful steps have been taken in the interim.