The ABSA case: the Constitutional Court rules on the GAAR

By Associate Professor David Warneke, Tax Partner

The judgment of the Constitutional Court in ABSA Bank Limited and United Towers (Pty) Ltd v CSARS (CCT 72/24) has crashed like a tsunami against a shore of transactions - present, past and future - washing away most of our general anti-avoidance rule (‘GAAR’) jurisprudence that had been developed over some 60 years.

In a nine to one majority judgment penned by Majiedt J, with the sole dissenting judgment that of Rogers J, the country’s apex court has undoubtedly put paid to many arrangements with strongly tax-beneficial elements. However, it interpreted the terms ‘party’ and ‘tax benefit’ with such a broad sweep that an investor in an arrangement in which their funds are used by other parties in an ‘impermissible avoidance arrangement’ without the investor’s knowledge, may be subject to attack by SARS. Moreover, the judgment’s application to less convoluted arrangements will also be challenging in many instances.

The facts

The facts were briefly that between 2011 and 2015 ABSA and its wholly owned subsidiary United Towers, which I shall refer to collectively as ‘ABSA’, entered into four preference share subscription agreements with a company PSIC Finance 3 (RF) (Pty) Ltd (PSIC3), for a total of R1.9 billion. ABSA received tax-exempt preference dividends on these preference shares. The investments were secured through back-to-back preference shares arrangements with PSIC Finance 4 (RF) (Pty) Limited (PSIC4), although ABSA maintained that it believed those arrangements were with Macquarie Securities South Africa Limited (MSSA), the Macquarie Group (Macquarie) having introduced the bank to these transactions. ABSA and United Towers maintained that, unbeknown to them, the funds flowed beyond PSIC3 into PSIC4 to a trust called Delta 1 Finance Trust (D1 Trust – a non-SA resident trust) and ultimately back to MSSA. The D1 Trust lent the funds to MSSA through interest-bearing notes. The D1 Trust received interest payments from MSSA on these notes, which it then used to acquire Brazilian government bond interest. Because interest on Brazilian government bonds is not taxable in SA in the hands of a SA resident under the SA-Brazil double tax treaty, it was distributed by the D1 Trust to PSIC4 as non-taxable income and up the chain of companies, ultimately to ABSA as dividends on the preference shares in PSIC3. 

So in summary, ABSA subscribed for preference shares and received a tax-exempt preference dividend. However, apparently unbeknown to ABSA, the amount of its funding was involved in a swap transaction designed by Macquarie whereby interest was paid by an unrelated SA resident (MSSA) and returned on-shore as tax-free interest. ABSA benefitted in that its dividend yield on the preference shares was higher than it would otherwise have been because of the tax-free nature of the interest received by PSIC4.

The litigation history

In October 2019 SARS assessed ABSA for additional taxation, alleging that they were parties to an impermissible avoidance arrangement under GAAR. SARS re-characterised the tax-exempt dividend income received by ABSA as taxable interest income. ABSA, on the other hand, argued that it was unaware of the transactions beyond its investment in the preference shares and ancillary agreements with Macquarie and that its involvement was limited to investing in preference shares in PSIC3.

ABSA launched a review in the High Court together with a request for direction under section 105 of the Tax Administration Act (the TAA), on the basis that the assessments were legally flawed because, first, ABSA could not be said to have been a party to an arrangement as it was unaware of the full structure, particularly the parts generating the alleged tax benefit and second, that it did not obtain a ‘tax benefit’ from the arrangement. The High Court entertained ABSA’s request for a review because it found these to be issues of law and set aside the assessments. SARS then appealed to the Supreme Court of Appeal (the SCA).

The SCA confined its judgment to the jurisdictional question under section 105 i.e., whether the High Court had the requisite jurisdiction to review the assessments. The SCA held that the ‘tax benefit’ question was not solely one of law but also of fact because in its view the effect, purpose and normality of a transaction are essentially factual questions. Thus, the dispute did not constitute an exceptional circumstance. It also found that the ‘party’ question also involved disputed facts rather than law. Thus, the SCA held that the High Court had erred in charactering the issues as involving only questions of law and that no exceptional circumstances justified exercising the granting of a direction under section 105. Hence it held that the High Court did not have jurisdiction to entertain the review.

In a much-anticipated judgment issued in March 2025 (United Manganese of Kalahari (Pty) Limited v CSARS and four other cases [2025] ZACC 2), the Constitutional Court granted ABSA leave to appeal and found that the SCA had erred in charactering the issues as factual since they involved only matters of law. This decision thus paved the way for the instant appeal by ABSA to the Constitutional Court that the assessments be set aside on review because of the errors of law allegedly made by SARS. The main questions considered by the Court were:

1. Whether the arrangement was an ‘impermissible avoidance arrangement’ as contemplated in the GAAR;
2. what constitutes a ‘party’ under the GAAR provisions and, in particular, whether it requires knowledge of all steps of an avoidance arrangement (‘the party issue’); and 
3. whether SARS can invoke the GAAR provisions against a taxpayer who is alleged not to have personally obtained a ‘tax benefit’ but had allegedly merely received financial returns from others’ tax benefits (‘the tax benefit issue’).

In what follows I refer to the judgment of the majority, unless otherwise stated.

The Court’s approach

The Court commenced with an examination of the context of GAAR in relation to tax avoidance more generally. For this it drew on established jurisprudence, stating at the outset that “[t]here is nothing illegal in minimizing one’s tax obligations and avoiding paying tax, provided that one does so within the parameters permitted by the law” and that “[a] taxpayer is entitled to choose the most tax efficient method of implementing a transaction where the same commercial result can be attained”. It then moved to a discussion of the historical background to South Africa’s current GAAR provisions, which is the third iteration of such provisions. The current version of the GAAR came into effect in 2006, replacing the now-repealed section 103(1) of the Income Tax Act (the Act). Various discussion papers issued by the executive branch of government- in this case SARS or the Minister of Finance- around the time the current GAAR provisions were enacted, strongly informed the Court’s interpretation of these provisions. 

The approach of the majority in its interpretation of the structure of the GAAR provisions was that section 80A is the heart of the GAAR and that the remaining subsections simply “expand on, provide remedies for and deal with related procedural and administrative aspects of the rule enunciated in section 80A”. Therefore, it did not consider the wording of the other subsections as guides to the interpretation of section 80A. With respect, in my view, this was an incorrect approach. Especially because of the imprecise style of wording employed in the GAAR provisions, the various subsections should be read together to make sense of the whole - besides shedding light on the context of the provisions. 

Was ABSA a party to the impermissible avoidance arrangement?

On the basis that the applicants did not take issue with SARS’ contentions that what had happened downstream was an impermissible avoidance arrangement, the Court proceeded to consider whether ABSA was a ‘party’ to the impermissible avoidance arrangement and whether it derived a ‘tax benefit’ from it.

It concluded that for a taxpayer to be a ‘party’ to an avoidance arrangement simply requires the taxpayer’s conduct to form part of the chain of transactions constituting the arrangement. Knowledge of how the downstream steps operate is not a requirement. This implies that one can be a ‘party’ to an arrangement that one does not even know exists - and is a point with which the minority strongly disagreed.

It appears that the majority was drawn to this counter intuitive conclusion because it believed that a dangerous precedent would otherwise be set, allowing investors to profit from tax avoidance simply by remaining ignorant of the schemes into which their funds were channelled. It stated: “If ‘party’ were to be interpreted to require knowledge, these schemes would be allowed to proliferate beyond the scope of the GAAR simply by actors purposefully keeping investors ignorant of the details of the downstream transaction’s tax avoidant nature”.

The Court was dealing with a situation in which SARS did not allege willful blindness on the part of ABSA. However, such a principle would have been preferable – in other words, that in order to be a ‘party’ to an arrangement in terms of the GAAR, a person must either have had knowledge of all of the transactional steps involved (although not necessarily knowledge of the legal consequences thereof or that there was a tax avoidance feature) or that the person must have been willfully blind to the steps. 

Can the GAAR be invoked against a party who did not personally obtain a ‘tax benefit’ but had allegedly merely received financial returns from others’ tax benefits?

Again counter intuitively, on the preliminary question whether SARS could tax a party that had not obtained a ‘tax benefit’, the majority held that the answer to this question is affirmative. The majority reasoned that the deliberate breadth of section 80B(1), which empowers SARS to determine the tax consequences for “any party”, evidences the legislative intent for a wide remedial reach and that had Parliament intended to restrict liability to the party obtaining a ‘tax benefit’, the legislation would have been worded accordingly. The minority disagreed, arguing that SARS’ power to tax under section 80B must be confined to the party that obtained a ‘tax benefit’ - that SARS may not allow a tax benefit to stand in the hands of the party that received it and tax another party which did not. The minority found support for this view in section 80G, which places the burden of proving the sole or main purpose of the arrangement was not to obtain a ‘tax benefit’ on the party that obtained a tax benefit. 

The majority found that notwithstanding that SARS was entitled to tax ABSA even if it had not obtained a ‘tax benefit’, ABSA had actually obtained a ‘tax benefit’.

The correct approach to determine if a ‘tax benefit’ is derived, according to the majority, is to strip a transaction of its avoidance features i.e. to view it from the perspective of the removal of artificial elements that serve no genuine commercial purpose. This is an enquiry into the economic substance of the arrangement rather than its legal form. It held that on this view, ABSA’s receipt of tax-exempt preference dividends was functionally equivalent to earning taxable interest, with the dividend form masking what was, in substance, a loan yield. 

One would have thought that any enhanced yield earned by ABSA would simply have formed part of its preference dividend, which was in any case exempt from tax. This point was expounded upon in the minority judgment, which stated that to determine whether there has been a ‘tax benefit’ one must contrast what occurred with a plausible counterfactual i.e. with what would have occurred if the impermissible component of the arrangement were removed or re-characterised. Why this counterfactual should have been a loan by ABSA rather than an investment in preference shares is not clear. The majority stated that the investment in preference shares was part of the avoidance planning inherent in the arrangement, which also had to be stripped away, thereby converting the tax-exempt dividend into taxable interest. But, with respect, this is unconvincing. Investment in preference shares instead of by loan is a routine economic choice made by taxpayers. Why should ABSA not have invested in preference shares and instead loaned the funds to PSIC3? As the majority noted at the commencement of its judgment, a taxpayer is entitled to choose the most tax efficient method of implementing a transaction where the same commercial result can be attained. If the impugned swap transaction downstream is removed from the arrangement, ABSA is left receiving a lesser quantum of tax- exempt dividends. 

To determine whether a ‘tax benefit’ exists in an arrangement, taxpayers have often argued that a ‘but for’ test should be applied by asking whether ‘but for’ the impugned arrangement or transactions, a tax liability would have arisen.  On the other hand, a different form of the ‘but for’ test centers on a comparison of the tax consequences of the arrangement entered into by the taxpayer and a plausible counterfactual. The majority states that the correct ‘but-for’ test requires assessing whether, but for the avoidance features, a tax liability would have arisen. It is not a comparison of the tax consequences of the impugned arrangement with no arrangement at all. 

Unfortunately, the finding of the Court on especially the ‘party’ and ‘tax benefit’ questions will have far-reaching economic consequences should SARS seek to apply the provisions to the full extent of the judgment. Unfortunately, these principles will not only apply to complex arrangements that are designed to secure a fiscal advantage but also to ordinary transactions. For example, if a bank offers an attractive yield on an investment product and, unbeknown to the investor, the bank is able to offer the attractive yield because of an impermissible avoidance arrangement that it has entered into with a third party, the investor will be a ‘party’ to the avoidance arrangement and SARS may choose to tax the investor because of the ‘tax benefit’ the investor derived. This is but one example. Another example cited by the minority is that a listed company with many shareholders may be able to declare higher dividends because it has arranged its affairs in a tax-efficient way. The majority recognized this risk in stating: “[t]he cognizable risk of overreach by attaching liability to any participant in a transaction later characterised as impermissible, even where the taxpayer’s involvement was commercially ordinary, cautioned against by the applicants, will be dealt with on a case-by-case basis.” 

With respect, from a practical standpoint, this is an unsatisfactory position. Investors require an acceptable level of certainty and need to know prior to entering into a transaction what its tax effects will be. In any event, one may well ask on what basis a court would be able to intervene in a decision by SARS to tax an ordinary investor under section 80B, given the findings of the majority and that this provision expressly grants wide powers to SARS. Besides the requirement that the Commissioner in his discretion make ‘compensating adjustments’, there are no express limitations on the quantum of SARS’ taxing powers under section 80B.

The majority also dealt with the requirement of ‘sole or main purpose’ as a jurisdictional requirement for entry into the GAAR under section 80A. It found that, unlike section 103(1), the test under the new GAAR is objective and not subjective, noting that section 80G requires of the party obtaining a tax benefit to prove that, reasonably considered in the light of relevant facts and circumstances, obtaining a tax benefit was not the sole or main purpose of the avoidance arrangement. Again with respect, the above wording appears to indicate a balancing of the taxpayer’s alleged subjective purpose against the objective facts and circumstances rather than a wholly objective test. It is difficult to understand how a party to an arrangement could bear the  onus of proving something which is wholly objective. In addition, as pointed out by the minority, if the ‘party’ issue is interpreted as widely as by the majority, section 80G may lead to a situation in which a taxpayer without knowledge of downstream steps would bear the onus of proving that the sole or main purpose of the arrangement (held by the majority to be an objective enquiry) was not the obtaining of a tax benefit. 

Both majority and minority judgments contain international law comparatives of GAAR: the minority concluding that the majority findings that a GAAR assessment may be issued against a person who was unaware of the impermissible tax avoidance and obtained an economic advantage but not a tax benefit from it is unprecedented internationally. The majority stated that the minority’s conclusion in this regard is reached only by omitting OECD-style GAARs.

In the result, ABSA’s appeal was dismissed with costs.

Conclusion

The impact of this judgment of our apex court cannot be overstated. It is the first substantive judgment on the new GAAR provisions, that date from 2006. It is alarmingly at odds with interpretation that has become widely accepted. Although the Court was dealing with a form of arrangement of which it clearly disapproved, the principles enunciated will have to be interpreted and applied in advice given on practically all significant transactions.