FirstRand Limited and Another v National Bank of Abu Dhabi PJSC (Pty) Limited

The judgment

The court's own file

THE SUPREME COURT OF APPEAL OF SOUTH AFRICA JUDGMENT Reportable Case no: 662/2024 In the matter between: FIRSTRAND LIMITED FIRST APPLICANT FIRSTRAND BANK LIMITED SECOND APPLICANT and NATIONAL BANK OF ABU DHABI PJSC (PTY) LIMITED RESPONDENT Neutral citation: FirstRand Limited and Another v National Bank of Abu Dhabi PJSC (Pty) Limited (662/2024) [2026] ZASCA 98 (7 July 2026) Coram: MBATHA ADP, MAKGOKA, SCHIPPERS, HUGHES and UNTERHALTER JJA Heard: 22 August 2025 Delivered: This judgment was handed down electronically by circulation to the parties’ representatives by email, publication on the Supreme Court of Appeal website and released to SAFLII. The date and time for hand-down of the judgment is deemed to be 7 July 2026 at 11h00. 2 Summary: Intellectual Property – Trade marks – Trade Marks Act 194 of 1993 – s 10(4) – whether there is a bona fide intention of using a mark as a trade mark – a factual issue. Superior Courts Act 10 of 2013 – section 17(2)(f) – application for reconsideration of refusal of leave to appeal – whether grounds for reconsideration and for granting leave to appeal established. 3 ORDER On application for reconsideration referred in terms of s 17(2)(f) of the Superior Courts Act 10 of 2013: 1 The application is struck from the roll. 2 The applicants shall pay the costs incurred by the respondent in opposing the application for reconsideration, including the costs of two counsel. JUDGMENT Makgoka JA (Schippers, Hughes and Unterhalter JJA concurring):

[1]This is an application under s 17(2)(f) of the Superior Courts Act 10 of 2013 (the SC Act) for reconsideration of an order by two judges of this Court refusing the applicants’ application for leave to appeal. The applicants sought leave to appeal against an order of the Gauteng Division of the High Court, Pretoria (the High Court). The High Court dismissed the first and second applicants’ opposition to the respondent’s trade mark applications, and the applicants’ subsequent application for leave to appeal.

[2]The first and second applicants are South African companies. The first applicant, FirstRand Limited, is the listed holding company of the FirstRand Group. The second applicant, FirstRand Bank, as the name suggests, operates as a commercial bank in South Africa. I collectively refer to the applicants as ‘FirstRand’. FirstRand is the proprietor in South Africa of a number of trade mark registrations in class 36 of the Trade Marks Act 194 of 1993 (the Act). The trade marks in class 36 cover a wide range of financial services. FirstRand has used its trade marks in relation to, among other things, financial activities and banking. 4

[3]The respondent, National Bank of Abu Dhabi PJSC (Pty) Limited (Abu Dhabi Bank), is a lender bank in the United Arab Emirates (UAE). It offers a full range of retail, corporate, wholesale and investment banking, wealth management and private banking, as well as Islamic banking, brokerage, property management and leasing services. Its international branch network is the largest among UAE banks, spanning 20 countries across five continents, including Asia, the Americas, and Africa.

[4]On 26 April 2017, Abu Dhabi Bank, in terms of s 21 of the Act, filed two trade mark registration applications with the Registrar of Trade Marks, seeking registration of the marks FIRST ABU DHABI BANK and GROW STRONGER FIRST ABU DHABI BANK, respectively, in class 36 for, among other things, banking affairs and related services. The marks were advertised for opposition in August 2018.

[5]FirstRand filed its opposition to the registration of the marks in February 2019. FirstRand’s opposition was initially based on ss 10(4), 10(12) and 10(14) of the Act, alleging that the registration of the trade marks would infringe its registered trade marks. The Registrar subsequently transferred the matter to the High Court in terms of s 59(2) of the Act. 1

[6]FirstRand abandoned reliance on s 10(12) and sought relief only under ss 10(4) and 10(14). The High Court dismissed FirstRand’s opposition and its subsequent application for leave to appeal. In February 2024, FirstRand applied for leave in this Court. On 21 May 2024, two judges of this Court considered the application and dismissed it on the ground that the appeal had no reasonable 1 Section 59(2) of the Act reads: ‘Where proceedings in terms of section 21, 24, 26, 27 or 38(8) are pending before the registrar, the registrar may in his discretion refer the proceedings to the court, and shall refer the proceedings to the court on written application of all the parties to such proceedings.’ 5 prospects of success and that there were no compelling reasons for the appeal to be heard.

[7]Undeterred, on 20 June 2024, FirstRand applied to the President of this Court, in terms of s 17(2)(f) of the SC Act, for reconsideration of the dismissal of its application for leave to appeal. In its application to the President, FirstRand limited its opposition to s 10(4) of the Act, thereby disavowing its earlier reliance on s 10(14). On 29 August 2024, the President referred the matter to the court for reconsideration and oral argument under s 17(2)(d). 2 She also directed the parties to be prepared to argue the merits, if called upon to do so.

[8]At the time FirstRand applied to the President for reconsideration, s 17(2)(f) had been amended to read: ‘The decision of the majority of the judges considering an application referred to in paragraph (b), or the decision of the court, as the case may be, to grant or refuse the application shall be final: Provided that the President of the Supreme Court of Appeal may, in circumstances where a grave failure of justice would otherwise result or the administration of justice may be brought into disrepute, whether of his or her own accord or on application filed within one month of the decision, refer the decision to the court for reconsideration and, if necessary, variation.’ 3 (Emphasis added.) 2 Section 17(2)(d) of the SC Act reads: ‘The judges considering an application referred to in paragraph (b) may dispose of the application without the hearing of oral argument, but may, if they are of the opinion that the circumstances so require, order that it be argued before them at a time and place appointed, and may, whether or not they have so ordered, grant or refuse the application or refer it to the court for consideration.’ 3 The section was amended by the Judicial Matters Amendment Act 15 of 2023, which came into operation on 3 April 2024. Before its amendment, the section read: ‘The decision of the majority of the judges considering an application referred to in paragraph (b), or the decision of the court, as the case may be, to grant or refuse the application shall be final: Provided that the President of the Supreme Court of Appeal may in exceptional circumstances, whether of his or her own accord or on application filed within one month of the decision, refer the decision to the court for reconsideration and, if necessary, variation.’ 6

[9]This Court’s jurisprudence on s 17(2)(f) is anchored in Motsoeneng v South African Broadcasting Corporation (Motsoeneng), 4 and Bidvest Protea Coin Security v Mabena (Bidvest). 5 The essence of these decisions is that it is the court to which the President refers the decision of the two judges for reconsideration that must determine whether exceptional circumstances exist, not the President.

[10]Counsel for FirstRand did not refer to these authorities in their heads of argument. However, once their attention was drawn to the authorities, counsel abandoned the argument that the President determines the issue. Once that issue was abandoned, the debate was whether exceptional circumstances warranted reconsideration of the order refusing leave to appeal. It is to that issue I turn. As explained in Luphondo v The State (Luphondo) 6 a party seeking a reconsideration must establish that there are exceptional circumstances, such that there is a likelihood of a grave failure of justice or bringing the administration of justice into disrepute if no reconsideration occurs. That is the threshold FirstRand has to meet.

[11]

[11]FirstRand premised its case for reconsideration on the High Court’s misstatements about s 10(4) of the Act, which forms part of s 10, headed ‘Unregistrable trade marks’. Section 10 prohibits the registration of certain marks. Section 10(4) prohibits the registration of: ‘[a] mark in relation to which the applicant for registration has no bona fide intention of using it as a trade mark, either himself or through any person permitted or to be permitted by him to use the mark as contemplated by section 38.’(Emphasis added.) 4 Motsoeneng v South African Broadcasting Corporation Soc Ltd and Others [2024] ZASCA 80; 2025 (4) SA 122 (SCA). 5 Bidvest Protea Coin Security (Pty) Ltd v Mabena [2025] ZASCA 23; 2025 (3) SA 362 (SCA). 6 Luphondo v The State [2026] ZASCA 24; 2026 (1) SACR 613 (SCA); [2026] 2 All SA 238 (SCA). 7

[12]The High Court misstated the law regarding: (a) the nature of an enquiry under s 10(4); and (b) the onus. As to (a), the High Court stated that it was required to determine whether Abu Dhabi Bank’s trade mark was confusingly or closely similar to FirstRand’s registered trade mark. The High Court also said that the enquiry concerned whether the registration of Abu Dhabi Bank’s trade mark would infringe FirstRand’s rights. Both of these statements are incorrect. Neither the similarity of the marks nor the infringement of rights is the focus of s 10(4), but rather of s 10(14).

[13]Regarding the onus, the High Court stated that ‘[t]he law requires that the alleged infringement relied upon for seeking protection in terms of the provisions of the Act be established by the proprietor of the registered trade mark’. To the extent the high court suggested that the onus lies on the opponent, it erred. These are opposition proceedings, and the onus rests on the applicant for registration to satisfy the court that there is no bar to registration. 7

[14]There are two possible reasons for the High Court’s error concerning the onus. First, the High Court seemed to have conflated the onus requirement in opposition proceedings under s 10(4) with that in revocation proceedings under s 27 of the Act. The latter provides for the removal of a trade mark from the register on the ground of non-use, and under s 27(3), the onus of proving relevant use of the trade mark rests on the mark’s proprietor. Second, Abu Dhabi Bank’s answering affidavit stated that, to succeed under s 10(4), ‘[FirstRand] must demonstrate that, at the time of filing the applications, there is no real intention to use the trade mark on the part of [Abu Dhabi Bank]. I do not believe that [FirstRand opponents] have met their onus’. 7 National Brands Ltd v Cape Cookies CC and Another [2023] ZASCA 93; 2024 (2) SA 296 (SCA); [2023] 3 All SA 363 (SCA); 2023 BIP 7 (SCA) para 6. 8

[15]In this Court, FirstRand submitted that s 10(4) would be rendered nugatory by the High Court’s findings that: (a) the party opposing the registration of a trade mark bears the onus of showing the applicant’s lack of bona fide intention to use; and (b) a trade mark applicant can rely on the mere fact of filing a trade mark application as proof of its intention to use the mark.

[16]FirstRand contended that if these findings were to stand, they would constitute a grave failure of justice or bring the administration of justice into disrepute. That would arise, so it was argued, because there would be uncertainty in the law of trade marks about where the onus lies in opposition proceedings, and the nature of the evidence required to establish the intention to use the mark to be registered. I address these submissions in turn.

[17]The fact that a court has misstated the law is ordinarily not, in itself, a ground for reversing a lower court’s judgment. This is because an appeal lies against the lower court’s substantive order, not the reasons for judgment. 8 Thus, whether or not a court of appeal agrees with a lower court’s reasoning would be of no consequence if the result remains the same. 9

[18]In the present case, despite its error regarding the onus in s 10(4) opposition proceedings, the High Court went on to consider whether Bank of Abu Dhabi had adduced sufficient evidence of its intended use. That underscores that the misstatement of the law was merely per incuriam, which does not create precedent. This is how High Court judges and trade mark legal practitioners will understand the High Court’s pronouncements. It cannot be said that this error brings the administration of justice into disrepute, within the meaning of s 17(2)(f) of the SC Act. 8 ABSA Bank Ltd v Mkhize and Two Similar Cases [2013] ZASCA 139; 2014 (5) SA 16 (SCA) para 64. 9 Western Johannesburg Rent Board and Another v Ursula Mansions (Pty) Ltd 1948 (3) SA 353 (A) at 355. 9

[19]It would have been entirely different had the High Court deliberately departed from settled law or the binding authority of this Court. Had that been the case, FirstRand would have had an arguable case that the administration of justice was being brought into disrepute within the meaning of s 17(2)(f). I therefore conclude that the High Court’s error as to who bears the onus does not constitute an exceptional circumstance.

[20]FirstRand alleged that, at the time it sought registration of its marks, Abu Dhabi Bank had no bona fide intention to use the marks as trade marks. In this regard, FirstRand relied on the fact that Abu Dhabi Bank neither held a banking licence from the Registrar of Banks in South Africa nor had applied for one. Consequently, it had no authorisation from the Financial Services Board (the FSB) to provide financial services and related services in South Africa. For these reasons, FirstRand contended that Abu Dhabi Bank’s trade mark applications offended s 10(4) because they concerned marks in relation to which Abu Dhabi Bank had no bona fide intention to use them as trade marks.

[21]For its part, Abu Dhabi Bank offered the following explanation. It endeavoured to secure protection for its trade marks as the first step in expanding into the South African market. This was informed by its ‘own internal processes, timelines and/or strategic decisions’. It took a business decision to first submit the trade mark applications in South Africa and secure their registration, before seeking a banking licence. It would reconsider its entry into South Africa only if it is unable to secure registration of its trade marks. It regarded the registration of its trade marks as fundamental, which explained why it identified it as the first step. It decided against applying for a banking licence to operate a bank, only for its trade mark applications to be rejected, with all the attendant financial risks. 10

[22]In its replying affidavit, FirstRand took issue with Abu Dhabi Bank’s understanding of s 10(4). As mentioned, in its answering affidavit, Abu Dhabi Bank suggested that FirstRand bore the onus of establishing that it had no intention of using its trade marks. Beyond that, FirstRand did not address Abu Dhabi Bank’s averments that securing its trade marks was the first step into the South African market and that, once that was achieved, it intended to apply for a banking licence. In its judgment, the High Court accepted Abu Dhabi Banks’ explanation as ‘objectively sound’ and applied the well-known Plascon-Evans principle. 10

[23]The High Court concluded that Abu Dhabi Bank’s version of its intention to use the trade marks had to prevail, as it did not consist of bald or uncreditworthy denials, nor of assertions so far-fetched or clearly untenable that the court was justified in rejecting them merely on the papers. The High Court deemed FirstRand’s allegations about Abu Dhabi Bank’s intended use of its trade marks speculative. Furthermore, the High Court noted that there is no requirement in South Africa to obtain authorisation or a licence to render the services in respect of which the trade mark relates, in order to register a trade mark.

[24]The High Court also referred to s 36 of the Act, which provides that, once issued with a certificate of registration, the proprietor of the trade mark must commence trading within five years of the certificate’s issue. It reasoned that the issues relating to the banking licence and FSB authorisation to render financial services in the country could be addressed within that period. The High Court 10 In accordance with the well-known principle enunciated in Plascon-Evans Paints Ltd v Van Riebeeck Paints (Pty) Ltd 1984 (3) SA 623 (A); [1984] 2 All SA 366 (A) at 634E-635C, which is this: in motion proceedings where disputes of fact arise on the affidavits, a final order can be granted only if the facts averred in the applicant's affidavit, which have been admitted by the respondent together with the facts alleged by the respondent, justify such an order. This is so, unless the respondent’s version consists of bald or uncreditworthy denials, raises fictitious disputes of fact, is palpably implausible, far-fetched or so clearly untenable that the court is justified in rejecting them merely on the papers. 11 also accepted Abu Dhabi Bank’s submission that it is inconceivable that an organisation as large as that bank would expend money and time in applying for trade mark registration in South Africa if it had no intention of doing business in the country. These factors, it reasoned, ‘exclude the likelihood that the application was made for ulterior purposes’. The High Court accordingly dismissed FirstRand’s opposition to Abu Dhabi Bank’s trade mark registration with costs.

[25]Abu Dhabi Bank bore the onus of establishing a bona fide intention to use a mark, as envisaged in s 10(4). This Court in Etraction v Tyrecor (Etraction) 11 held that, in determining whether the onus has been met, the facts should be examined ‘as they appear from the course of events leading up to the application for registration . . .’.

[26]Some dicta have suggested that the opponent bears an evidential burden of proving that the applicant does not intend to use the trade marks it seeks to register. For example, in Groot Constantia Trust v DGB 12 the High Court said: ‘The interpretation of the provisions of Section 10(4) of the Act stands, in my view, on a different platform. As to whether a mark in relation to which the applicant for registration has no bona fide intention of using it as a trademark either by himself or through a third party, is a matter which requires, to my mind, some special set of circumstances of which the Opponent carries evidential burden to disprove lack of bona fides on the part of the respondent. It cannot in my view, therefore, reasonably be expected of the respondent to carry the evidential yoke in this instance to demonstrate its bona fide intention of using the mark, once it has discharged the onus that its registration would not be offensive. To do so, would be raising too high the bar in this regard, so to speak.’ 13 11 Etraction (Pty) Ltd v Tyrecor (Pty) Ltd [2015] ZASCA 78; 2015 BIP 253 (SCA) para 33. 12 Groot Constantia Trust v DGB (Proprietary) Limited [2015] ZAGPPHC 1086; 2015 JDR 2039 (GP); 2015 BIP 330 (GP). 13 Para 22. 12

[27]The same line of reasoning was recently followed by the English Court of Appeal in Lidl v Tesco: 14 ‘A person is presumed to have acted in good faith unless the contrary is proved; but where the objective circumstances relied upon by the party challenging the validity of the registration give rise to a prima facie case of bad faith, the evidential burden shifts to the applicant for registration to explain its intentions at the time of making the application. . .’.

[28]The effect of these dicta is that the intention to use the trade marks is presumed merely by the filing of the application for registration. I disagree. If the intention is presumed by the mere making of the application, leaving it to the opposing party to discharge an evidential burden of bad faith, it may be said that the onus de facto rests with the opponent, as the presumption favours the applicant, without any statutory basis for it. To that extent, these dicta are at odds with what is acknowledged to be the true incidence of the onus.

[29]As I see it, an opponent in trade mark registration bears no evidential burden. It is entitled to simply ask the applicant for registration to demonstrate its intention to use the trade marks. It is then up to the applicant to present acceptable evidence of an intent to use the marks. If the opponent does not accept the evidence, then it bears the evidential burden to adduce evidence of an ulterior motive in seeking registration of the marks.

[30]As this Court pointed out in Etraction, when the bona fide use of a mark is in issue, it must be determined ‘whether there was a genuine intention to use the mark in the course of trade in respect of those goods or whether the registration was intended for an ulterior purpose’. Whether the onus has been discharged is a question of fact. 14 Lidl Great Britain Ltd and Lidl Stiftung & Co. KG v Tesco Stores Limited and Tesco PLC [2024] EWCA Civ 262 para 173. 13

[31]What is important here is the intention at the time of seeking registration of the marks. This Court, in Victoria’s Secrets Inc v Edgars Stores Ltd 15 referred to the English Court of Appeal’s case in In re Ducker’s Trade Mark, 16 which concerned the English equivalent of our requirement for ‘intention to use’. It held that the concept required, among other things: (a) ‘some definite and present intention to deal, in certain goods or descriptions of goods’; (b) a ‘resolve to use [the mark] in the immediate future’; and (c) ‘a settled purpose which has been reached at the time when the mark is to be registered’. Whether the requirement has been met in a particular case must be determined on the specific facts of that case, as mentioned.

[32]Thus, in the present case, once FirstRand opposed its application for registration, Abu Dhabi Bank bore a duty to adduce evidence that the application to register its marks was made with a genuine intention to use them. Abu Dhabi Bank explained that the decision to first secure its trade marks was strategic and business-related. It must be considered, among other factors, that Abu Dhabi Bank offers banking and financial services across several jurisdictions on five continents, including Asia, the Americas and Africa. Its international branch network spans 20 countries.

[33]Although it currently does not hold a banking licence or FSB authorisation in South Africa, it is common cause that, prior to the filing of the trade mark registration application, Abu Dhabi Bank was already using its trade marks for banking and financial services in those jurisdictions. Ordinarily, on its own, this does not constitute a present and definite intention to use the marks. It is, however, an important consideration. This is especially so when viewed alongside other 15 Victoria’s Secrets Inc v Edgars Stores Ltd 1994 (3) SA 739 (A); [1994] 4 All SA 307 (A). 16 In re Ducker's Trade Mark [1929] 1 Ch 113 (CA) ([1928] 45 RPC 105) at 121. 14 factors, such as the absence of an overtly ulterior motive in seeking registration of the marks.

[34]FirstRand predicated its opposition on the fact that Abu Dhabi Bank does not possess a banking licence nor authorisation from the FSB to offer financial services in South Africa. This, on its own, does not suggest an ulterior motive as envisaged in Etraction. The High Court correctly held that Abu Dhabi Bank is not required by the Act or any law to hold a banking licence or an FSB authorisation, or to have applied for either, in order to register its marks.

[35]There is no reason to doubt that, should its trade marks be registered, Abu Dhabi Bank would comply with all necessary legal requirements to operate in South Africa, including applying for a banking licence from the Registrar of Banks and authorisation from the FSB. If those were approved, there is no reason to doubt that it would use the trade marks in relation to banking and financial services and expand its operations into South Africa. On these considerations, it is difficult to fault the High Court’s findings.

[36]Even if the High Court’s findings were wrong, FirstRand faces an even more formidable obstacle to establish that they were so gross as to result in a grave failure of justice or bring the administration of justice into disrepute. These requirements set an exceptionally high bar, consistent with our superior courts’ approach to the issue.

[37]In Luphondo, this Court undertook an extensive survey of how both this Court and the Constitutional Court have approached the issue. It also considered the position in Canada. The Court observed that in all the cases where it granted reconsideration in terms of s 17(2)(f), or where the Constitutional Court reviewed 15 its earlier decision refusing leave, the administration of justice would have fallen into disrepute, had there been no judicial intervention.

[38]In the present case, no such circumstances exist. FirstRand’s complaints fall far short of the higher standard required to bring the administration of justice into disrepute. They amount to no more than a ‘mere difference of opinion’. Reading the High Court’s judgment, an informed and reasonable member of the public might disagree with it. But aware of all the relevant circumstances of the case, they are unlikely to undermine the public’s confidence in our legal system or its integrity.

[39]FirstRand’s application does not meet the threshold under s 17(2)(f). Consequently, this Court lacks jurisdiction to reconsider the two judges’ order refusing leave to appeal. There is no likelihood that a grave failure of justice will occur or that the administration of justice would be brought into disrepute if that order is not reconsidered. Costs must follow the result. I consider that the employment of two counsel was warranted, given the issues involved and the importance of the matter to the parties. This is reflected in FirstRand’s employment of three counsel.

[40]In conclusion, it is necessary to reiterate the purpose of s 17(2)(f). In Avnit v First Rand Bank (Avnit) 17 Mpati P made a pointed observation about this: ‘[T]he section is not intended to afford disappointed litigants a further attempt to procure relief that has already been refused. It is intended to enable the President of this Court to deal with a situation where otherwise injustice might result. An application that merely rehearses the arguments that have already been made, considered and rejected will not succeed, unless it is strongly arguable that justice will be denied unless the possibility of an appeal can be pursued . . .’. 18 17 Avnit v First Rand Bank Ltd [2014] ZASCA 132. 18 Ibid para 6. 16

[41]The second judgment appears to treat reliance on this passage as an equivocation on my part as to whether Avnit is authority for the proposition that it is the President who decides the presence of exceptional circumstances. Far from it. The second judgment misunderstands this passage and the reason I rely on it. I rely on the underlying purpose of s 17(2)(f), about which there is no disagreement, not for the proposition that it is the President who determines the presence of exceptional circumstances. The interpretive issue about the repository of the power to decide the presence of exceptional circumstances is conceptually distinct from the provision’s purpose. As I demonstrate later, Mpati P does not address the interpretive issue in the above passage or anywhere in the judgment.

[42]The special narrow ‘safety net’ provided by s 17(2)(f) should not be abused. It should be reserved for truly exceptional circumstances to avoid injustice. This Court, in Luphondo, provided clear guidelines on what should be considered before an application under s 17(2)(f) is launched. There should be ‘something out of the ordinary, markedly unusual, rare or different, and to which the general rule does not apply’, as stated in Liesching II. 19

[43]Parties and their legal representatives should therefore carefully examine the facts of each case to determine whether the contemplated application in terms of s 17(2)(f) falls within the narrow purview of exceptional circumstances. Unmeritorious applications for reconsideration may well amount to an abuse of this process. Where it becomes clear that this injunction has not been heeded, such conduct, where applicable and appropriate, will be met with a punitive costs order. 19 S v Liesching and Others [2018] ZACC 25; 2019 (1) SACR 178 (CC); 2019 (4) SA 219 (CC); 2018 (11) BCLR 1349 (CC) para 40. 17 The second judgment

[44]The above should be the end of the matter. However, in the second judgment, my Colleague, Mbatha ADP, considers whether the power to determine the presence of exceptional circumstances lies with the President or the court. This is unnecessary. As noted, this issue has been settled in Motsoeneng and Bidvest and was recently affirmed by a unanimous court in Luphondo. All three are binding unanimous five-panel benches of this Court, and we are obliged to follow them under the principle of horizontal stare decisis.

[45]More importantly, the issue of the repository of the power under s 17(2)(f) does not arise in this case. As noted, counsel for FirstRand initially contended in their heads of argument that the power lies with the President, but abandoned that contention once Motsoeneng and Bidvest were brought to their attention. They accepted that it is the court, not the President, that determines whether exceptional circumstances exist.

[46]Once a party abandons an issue, as FirstRand did in the present case, it is off the table as a point in dispute between the parties. It is then generally not open to a court to determine the matter on the abandoned issue, as doing so would be in the abstract. This is exactly what the second judgment does. This runs counter to this Court’s and the Constitutional Court’s repeated caution against addressing issues not necessary for the disposal of cases. 20 In Albutt v Centre for the Study of Violence and Reconciliation 21 the Constitutional Court emphasised that: ‘[S]ound judicial policy requires us to decide only that which is demanded by the facts of the case and is necessary for its proper disposal. . . At times it may be tempting, as in the present case, to go beyond that which is strictly necessary for a proper disposition of the case. Judicial 20 Fischer and Another v Ramahlele and Others [2014] ZASCA 88; 2014 (4) SA 614 (SCA); [2014] 3 All SA 395 (SCA) paras 13 and 14; Bliss Brands (Pty) Ltd v Advertising Regulatory Board NPC and Others [2023] ZACC 19; 2023 (10) BCLR 1153 (CC) para 1. 21 Albutt v Centre for the Study of Violence and Reconciliation [2010] ZACC 4; 2010 (3) SA 293 (CC); 2010 (5) BCLR 391 (CC) 18 wisdom requires us to resist the temptation and to wait for an occasion when both the facts and the proper disposition of the case require an issue to be confronted. This is not the occasion to do so. There may well be cases, and they are very rare, when it may be necessary to decide [for example] an ancillary issue in the public interest. 22

[47]The debate around the repository of power between the President and the court will require an appropriate case, where the court seized with it will have the benefit of pointed submissions for and against the divergent approaches identified in Bidvest. This is not the occasion. None of the issues raised in the second judgment was debated with counsel, nor was it put to FirstRand’s counsel that their abandonment of their initial argument was in error.

[48]The second judgment holds that it is not bound by FirstRand’s abandonment of the power repository issue, based on the principle that a court is not bound by an incorrect concession on a point of law. That principle does not apply here. Counsel for FirstRand abandoned their power repository argument on the basis of the binding authority of Motsoeneng and Bidvest. Thus, the ‘concession’, if ever it was, was correctly made. A concession made on the basis of binding authority cannot be said to be erroneous.

[49]Despite the binding authority of Motsoeneng, Bidvest, and Luphondo, the second judgment adds to what I consider an unnecessary proliferation of judgments that run contrary to them. 23 This creates uncertainty in this Court’s jurisprudence, which is exacerbated by the fact that some members of this Court 22 Ibid para 82. 23 See, for example, the minority judgments in S v Lorenzi [2025] ZASCA 58; 2025 JDR 2015 (SCA); Schoeman v Director of Public Prosecutions 2025 (2) SACR 561 (SCA); Fisher v Silverbirch Estate Homeowners' Association (NPC) and Others [2026] ZASCA 69; three-panel judgments in 4 Seasons Logistics CC v Kgotse [2026] ZASCA 9; [2026] 1 All SA 415; Matsi and Another v The South African Legal Practice Council (Gauteng Province) [2026] ZASCA 12; 2026 JDR 0649 (SCA);Lutzkie v Commissioner for the South African Revenue Service [2026] ZASCA 11; 2026 JDR 0648. 19 have, in one case, supported Motsoeneng and Bidvest, while in another, declined to follow them.

[50]For example, Mbatha ADP, the author of the second judgment here, was part of the panel that decided Bidvest. She later endorsed Motsoeneng and Bidvest as a member of the panel in Lekeka v S (Lekeka), 24 which made the correct observation that ‘as a panel of three judges of this Court, we remain bound by . . . Motsoeneng and Bidvest . . .’ 25 In a similar vein, in 4 Seasons v Kgotse 26 (4 Seasons) (Petse, Mbha and Dlodlo AJJA) pronounced Motsoeneng and Bidvest to be ‘clearly wrong’. Yet the same panel had earlier endorsed and followed these judgments in RAF v Lewis. 27 In an earlier judgment, Nel v S, 28 Dlodlo AJA also supported Motsoeneng and Bidvest.

[51]This is undesirable. Litigants are entitled to certainty that their s 17(2)(f) applications will be adjudicated on a predictable, authoritative and principled basis, rather than the changing and evolving views of members of this Court.

[52]The second judgment holds that it is not bound by Motsoeneng and Bidvest as it considers them to have been wrongly decided, because according to the second judgment, they: (a) failed to follow Avnit v First Rand Bank (Avnit) 29 (the Avnit effect); (b) are incompatible with constitutional values (constitutional values); and (c) create an asymmetry (the asymmetry issue). The second judgment holds that, in any event, Motsoeneng and Bidvest have been overruled by 4 Seasons (the panel size issue). Although in my view it is neither necessary nor appropriate to traverse the issue around the repository of power under s 17(2)(f) 24 Lekeka v S [2025] ZASCA 182. 25 Ibid para 17. 26 4 Seasons Logistics CC v Kgotse [2026] ZASCA 9; [2026] 1 All SA 415. 27 Road Accident Fund v Lewis [2025] ZASCA 174; [2026] 1 All SA 70 (SCA). 28 Nel v S [2025] ZASCA 89. 29 Avnit v First Rand Bank Ltd [2014] ZASCA 132. 20 in this case, the issues in the second judgment identified above, warrant comment. I consider each, in turn. The Avnit effect

[53]The second judgment makes much of the decisions before Motsoeneng, regarding the exercise of the President’s discretion under s 17(2)(f), including Avnit. But those decisions did not consider the jurisdictional issue. It was not an issue before the court, such that those judgments were of no assistance in determining the merits of Motsoeneng and Bidvest when the issue did arise and was determined.

[54]Avnit requires elucidation. What is binding in a judgment is the ratio decidendi, which is the principle extracted from the case. 30 This ordinarily depends on the issue to be determined in the case. As to how the ratio decidendi is determined, Schreiner JA laid down the following test in Pretoria City Council v Levinson (Levinson): 31 ‘[W]here a single judgment is in question, the reasons given in the judgment, properly interpreted, do constitute the ratio decidendi, originating or following a legal rule, provided (a) that they do not appear from the judgment itself to have been merely subsidiary reasons for following the main principle or principles, (b) that they were not merely a course of reasoning of the facts . . . and (c) (which may cover (a)) that they were necessary for the decision, not in the sense that it could not be reached along other lines, but in the sense that along the lines actually followed in the judgment the result would have been different but for the reasons.’ 30 Collect v Priest 1931 AD 290. 31 Pretoria City Council v Levinson 1949 (3) SA 305 (A) at 317; see also Makhanya v University of Zululand [2009] ZASCA 69; 2010 (1) SA 62 (SCA); [2009] 8 BLLR 721 (SCA); [2009] 4 All SA 146 (SCA); (2009) 30 ILJ 1539 (SCA) para 81. 21

[55]This test was affirmed in True Motives 84 (Pty) Ltd v Mahdi, 32 where Cameron JA observed that only the ratio of the decision of a court is binding, and not what might have been said in passing. He explained: 33 ‘According to Schreiner JA’s approach, the reasons given creating or following a legal rule are binding on this court provided they were not merely subsidiary to the main principle, that they were not merely linked to the incidental facts . . . and that they were necessary for the decision in the sense that along the lines that the court actually followed the results would have been different, but for the reasons.’

[56]Applying Schreiner JA’s distinction to Avnit, it must first be determined what the issue in that case was. There, the issue concerning s 17(2)(f) was two- fold. First, to determine what constitutes exceptional circumstances to trigger the President’s power in s 17(2)(f). Second, to discern the purpose of the provision. A different issue arose in Motsoeneng, Bidvest and subsequent cases, namely, as between the President and the court to which the matter is referred, who is the repository of the power to determine whether exceptional circumstances exist?

[57]This question was not addressed in Avnit. It is thus not surprising that Avnit does not determine that the power to decide whether exceptional circumstances exist resides with the President, rather than the court to which the application is referred. In considering the purpose of the provision, the President observed: ‘In the context of s 17(2)(f) the President will need to be satisfied that the circumstances are truly exceptional before referring the considered view of two judges of this court to the court for reconsideration. . .’. 34

[58]This statement is no more than a recitation of the provisions of s 17(2)(f) and does not constitute the ratio decidendi of what was before the President in 32 True Motives 84 (Pty) Ltd v Mahdi and Another [2009] ZASCA 4; 2009 (4) SA 153 (SCA) para 105; 2009 (7) BCLR 712 (SCA); [2009] 2 All SA 548 (SCA). 33 Ibid para 105. 34 Avnit para 33. 22 that case. It is not a considered judgment on the issue in dispute in the present case, as the issue did not arise. The interpretative question regarding the repository of power in s 17(2)(f) only arose in Motsoeneng and was expanded upon in Bidvest and in the majority judgment in Schoeman v Director of Public Prosecutions (Schoeman). 35 The latter two judgments, in particular, gave focused attention to the interpretive issue of the repository of the power to determine whether exceptional circumstances exist.

[59]Viewed in this light, Avnit did not lay down any precedent on the interpretive question. It is therefore incorrect to hold, as the second judgment does, that Avnit was binding on the subsequent panels and that they could depart from it only if they found it clearly wrong. Those judgments did not have to declare Avnit clearly wrong because that case did not lay down any precedent on the interpretive issue considered in those judgments. But to be clear, to the extent Avnit can be read to mean that it is the President who determines the existence of exceptional circumstances, it is clearly wrong, and has been overturned for all the reasons articulated in Bidvest and in the majority judgment in Schoeman. The constitutional issue

[60]The second judgment holds that it is entitled to disregard Motsoeneng and Bidvest as binding authorities of this Court because it considers Bidvest, in particular, to conflict with constitutional values. I disagree. Bidvest postulated two interpretations of s 17(2)(f), both of which are consistent with the constitution. The issue is which of the two accords more with the language, purpose and context of the provision. The fact that one does not agree with the interpretation preferred in Bidvest does not render such an interpretation inconsistent with constitutional values. The notion that some constitutional harm 35 Schoeman v Director of Public Prosecutions 2025 (2) SACR 561 (SCA). 23 is done to litigants whose application is not referred to the court does not seem to engage any constitutional issue because it is premised on a flawed premise that the court is reviewing the President’s decision. The asymmetry issue

[61]This ground is premised on a passage in the Constitutional Court’s minority judgment in Godloza, which held that the President of the Supreme Court of Appeal cannot have the power to determine whether exceptional circumstances are present for purposes of refusing reconsideration, but not for purposes of granting it.

[62]The short answer is this: a litigant whose application for leave to appeal has been refused because there are no exceptional circumstances is visited with finality. If there are exceptional circumstances, there is a special statutory exception to finality, and a reason for such a litigant to be treated differently. The only question is who must finally decide whether there are exceptional circumstances. The panel size issue

[63]This issue has been authoritatively settled by this Court in Luphondo. There, it was held that the judgments of this Court have the same authority, irrespective of the number of judges on the panel. However, in the event of a difference in opinion between a smaller bench and the larger bench, the binding authority is that of the larger bench. The implication is therefore that a smaller bench cannot competently overrule a judgment by a larger bench.

[64]Therefore, on the basis of stare decisis, we are bound by that decision. It is therefore not open to the second judgment to hold otherwise, unless it can demonstrate that the judgment is ‘clearly wrong’. It is not clear from the second 24 judgment as to the basis on which it considers itself not bound by Luphondo. It seems to hold that Motsoeneng and Bidvest were competently overruled by 4 Season, despite the latter judgment being rendered by a three-panel bench, as opposed to Motsoeneng and Bidvest, which were both given by five-panel benches. The second judgment grounds this conclusion on two bases. The first is s 13 of the SC Act, and, second, a United Kingdom judgment in Young v Bristol Aeroplane Company [1944] 1 KB 718 (Bristol Aeroplane). 36 I consider these below. Section 13 of the SC Act

[65]The provision is titled: ‘Manner of arriving at decisions by Supreme Court of Appeal’. Section 13(1) provides that proceedings of this Court must ordinarily be presided over by five judges, subject to the proviso that the President may: (a) direct that an appeal in a criminal or civil matter be heard before a court consisting of three judges (s 13(1)(a)); or (b) given its importance, direct that an appeal be heard by a larger number of judges, as she may determine (s 13(1)(b)). In terms of s 13(2)(a), the judgment of the majority of the judges presiding at proceedings before this Court shall be the judgment of the court.

[66]The effect of these provisions is that the judgment of any properly constituted bench of this Court is binding authority. It is immaterial whether the bench comprised three, five or more judges. Thus, a unanimous judgment of three judges carries the same authority as that of five or more judges. Where there is no unanimity, the majority of the bench is the authority. Thus, in a three-panel bench, the majority judgment of two is the authority, whereas in a five-panel bench, the majority judgment of three or four is the authority. 36 Young v Bristol Aeroplane Company [1944] 1 KB 718. 25

[67]Thus, properly understood, s 13 concerns the status of a judgment of this Court. Whether unanimous or by a majority, a judgment of this Court enjoys the same status and authority, irrespective of the number of judges on the panel. This uncontroversial. This is the same point made in Bristol Aeroplane, relied upon in the second judgment. Bristol Aeroplane is well-known for its strict adherence to the doctrine of stare decisis. It affirmed the rule that the Court of Appeal is bound to follow its own previous decisions, except in three tightly defined circumstances, namely: (a) where there are conflicting decisions in past Court of Appeal cases; (b) a previous Court of Appeal decision is inconsistent with a later House of Lords decision; and (c) the previous decision was made per incuriam (made carelessly or by mistake because a relevant statute or rule was overlooked). The judgment says nothing about the authority of different panel sizes in the event of a conflict between them. It therefore does not support the proposition in the second judgment that a smaller panel can overrule a judgment of a larger one.

[68]In any event, the UK Supreme Court’s modern approach to bench-size authority runs contrary to what the second judgment suggests the position in the United Kingdom to be. The issue is governed by paragraph 3.3.1 of the Court’s Practice Direction 3, which requires a party seeking permission to appeal to state whether they ask the Supreme Court ‘to depart from one of its own decisions or from one made by the House of Lords.’ Once this is stated, a larger panel than the one that decided the earlier case is convened, in anticipation of a possible overruling of that decision. As explained by the Supreme Court in Attorney General v Crosland: 37 ‘It not infrequently happens that a party to an appeal to the Supreme Court wishes the court to depart from an earlier decision of the Supreme Court or of the House of Lords. In such a case 37 Attorney General v Crosland [2022] 2 All ER 401; [2022] 1 WLR 367, [2021] UKSC 58. 26 the practice is for the appeal to be heard by an enlarged panel of seven or more justices, precisely to clothe it with that greater authority . . .’. 38 (Emphasis added.)

[69]In Luphondo, this Court undertook a survey the UK Supreme Court authorities which reveals that where that Court and its predecessor, the House of Lords, overruled its own decisions, it was either by an equally constituted bench or by an enlarged bench. 39 In Rock Advertising Limited v MWB Business Exchange Centres Limited, 40 the UK Appeal Court declined to resolve a contentious issue which had given rise to conflicting judgments on the basis that the issue should be resolved by an enlarged panel. Writing for the Court, Lord Sumption explained: ‘[T]he reality is that any decision on this point is likely to involve a re-examination of the decision in Foakes v Beer. It is probably ripe for re-examination. But if it is to be overruled or its effect substantially modified, it should be before an enlarged panel of the court . . .’. 41 (Emphasis added.)

[70]The upshot of the above recent authorities is that Bristol Aeroplane, decided eight decades ago, is no authority for the proposition advanced in the second judgment. In addition to the UK Supreme Court, this Court, in Luphondo, undertook an exhaustive comparative analysis of other jurisdictions, none of which had a smaller bench overturning a larger bench. 42 The second judgment cites no case, either from this Court or from any comparable jurisdiction, in which a court of equivalent status to ours has overruled its earlier decision with fewer judges than the earlier decision. This is unsurprising, as there is none. 38 Ibid para 46. 39 Luphondo paras 49-52. 40 Rock Advertising Limited v MWB Business Exchange Centres Limited [2018] 4 All ER 21; [2019] AC 119. 41 Ibid para 18. 42 Luphondo paras 47-48. 27

[71]The second judgment holds that there are conflicting views on the interpretation of s 17(2)(f), as a result which subsequent panels are entitled ‘to choose between the two.’ But there are not ‘two’ judgments to choose from in respect of s 17(2)(f). Motsoeneng and Bidvest are the binding authorities in this Court for all the reasons explained in Luphondo. These authorities can only be deviated from in very circumscribed circumstances envisaged by this Court in Bloemfontein Town Council v Richter 43 and Patmar v Limpopo Development Tribunal, 44 the bottom line of which is that they should be ‘clearly wrong’.

[72]Recently, in Fisher v Silverbirch Estate Homeowners’ Association 45 the majority of this Court affirmed the doctrine of precedent and the binding authority of Motsoeneng and Bidvest. The majority stressed the importance of consistency and legal certainty in the application of reconsideration proceedings under s 17(2)(f). In the course of his judgment, Goosen JA offered a substantive and trenchant critique of 4 Seasons, with which I agree.

[73]In In the Matter of an Application by Rosaleen Dalton for Judicial Review, 46 writing for a unanimous Court, Lord Reed, the President of the UK Supreme Court, sounded the following caution about an appellate court easily departing from its own decisions: ‘The court will not overrule a previous decision simply because the justices would decide the case differently today. . . This principle is vitally important to the operation and reputation of a court which does not sit en banc, and whose composition consequently varies from one case to another. In such circumstances, the principle is essential to counter the risk that the outcome of cases might otherwise depend, or at least might appear to depend, on who happened to be sitting . . . .[I]f a tenable view taken by a majority in the first appeal could be overruled by a 43 Bloemfontein Town Council v Richter 1938 AD 195 at 232. 44 Patmar Explorations (Pty) Ltd and Others v Limpopo Development Tribunal and Others [2018] ZASCA 19; 2018 (4) SA 107 (SCA) paras 3 and 4. 45 Fisher v Silverbirch Estate Homeowners' Association (NPC) and Others [2026] ZASCA 69. 46 In the Matter of an Application by Rosaleen Dalton for Judicial Review [2023] UKSC 36; [2023] 3 WLR 671 para 47. 28 majority preferring another tenable view in a second appeal, then the original tenable view could be restored by a majority preferring it in a third appeal, and finality of decision would be utterly lost.’

[74]This very issue regarding the repository of power under s 17(2)(f) to determine the existence of exceptional circumstances could well be resolved by a five-judge bench, or, at the discretion of the President, acting in terms of s 13(1)(b) of the SC Act, by an enlarged panel of seven, nine or even 11 judges. On the reasoning of the second judgment, a subsequent three-judge bench, or a majority thereof (two), would be entitled to overturn the judgment of such an enlarged panel. The absurdity of this is self-evident. This would result in ‘intolerable legal uncertainty’, to borrow from Brand JA in Potgieter v Potgieter, 47 albeit in a different context. Further aspects of the second judgment

[75]The second judgment, without much reasoning, reaches certain conclusions and criticises Motsoeneng on the basis that it concluded that the court was empowered to consider the question of exceptional circumstances. But if Motsoeneng had accepted that this issue had already been determined, the court would have been precluded from revisiting the issue. This is to reason in a circle. Of course, if the President’s decision on exceptional circumstances is ‘final’, then there is nothing for the court to which the matter is referred to decide on that issue. But it is this very premise that has led to different views in our Court. Thus, to assume the premise, does not demonstrate that Motsoeneng is wrong. It assumes what has to be reasoned over.

[76]Similar arguments, based on this assumption, are directed at Bidvest. One new argument is that there is no review of the President’s decision to dismiss an 47 Potgieter v Potgieter NO [2011] ZASCA 181; 2012 (1) 637 (SCA) para 34. 29 application for reconsideration, and thus, there cannot be a review of the decision to refer a matter for reconsideration. This interpretation misconstrues the power vested in the President under s 17(1)(f). The power of referral is an exception to the finality that attaches to a decision on petition to two judges of the Court. The repository of the power to determine whether there are exceptional circumstances that warrant reconsideration is precisely the question that has yielded divergent views in this Court. But the second judgment rests on a misconception. The court to which the matter is referred is not reviewing the President’s decision. It is determining whether the jurisdictional facts exist for reconsideration. All the President does is refer the matter to the Court to decide whether the jurisdictional prerequisites are met. If the President does not refer, finality holds. There is nothing to ‘review’ in either case.

[77]It is also said in the second judgment that Bidvest ‘strips the President of the substantive judicial function’ and renders the President an administrative functionary. This is not so. The President plainly exercises judicial power at a prima facie level to decide whether to refer, and the question is simply whether the jurisdictional facts are in place for the Court to assume the power to reconsider. Motsoeneng and its progeny hold that the Court must do so.

[78]An important consideration is that when the President considers an application in terms of s 17(2)(f), she or he does not have the benefit of either the record of the proceedings or the heads of argument. The court has the benefit of all of these, and, importantly, of full oral argument. As I see it, the essence of s 17(2)(f) is this: the President makes a prima facie determination that there is a likelihood of grave injustice or of the administration of justice being brought into disrepute if the order refusing leave is not reconsidered. She directs the court to carefully consider that likelihood and to make a final determination on the prima facie determination. The court, upon consideration of the full record, the heads of 30 argument and oral submissions, may or may not confirm that prima facie determination. If it confirms it, it would likely reconsider the order refusing leave to appeal. If it does not, the order refusing leave to appeal is final.

[79]Section 17(2)(f) is not unique in this regard. There are many analogies in our law. For example, under s 304(4) of the Criminal Procedure Act 51 of 1977, a senior magistrate may submit a record of proceedings to the high court for review if it appears that the proceedings in a particular case are not in accordance with justice. Such a magistrate decides nothing substantive, other than forming a prima facie view that the proceedings might not be in accordance with justice. It is the high court to which the proceedings are referred that makes a firm determination whether they accord with justice.

[80]The second judgment’s interpretation could never be a sensible construction of s 17(2)(f). If the Legislature had intended the President to be the repository of the power to determine the presence of exceptional circumstances, it would also have granted the President the power to reconsider the order refusing leave. The second judgment places considerable weight on the words ‘of her own accord’ in s 17(2)(f) as indicating that the power to determine the presence of exceptional circumstances lies with the President. But this is simply a safeguard should a grave matter come to the President’s attention outside an application.

[81]Thus, nothing turns on the words ‘whether of his or her own accord’. The key issue in the provision is the likelihood of the two jurisdictional factors in the section. Whether the President considers the matter of her own accord or on application has no bearing on the interpretive question about the repository of the power to determine the existence of exceptional circumstances. 31

[82]The second judgment holds that the President has already made a binding determination that exceptional circumstances exist. But in the same breath, it holds that the two judges did not err, and therefore, their order should not be reconsidered. The second judgment does not explain what becomes of the President’s binding determination that exceptional circumstances are present. It must be borne in mind that exceptional circumstances entail the likelihood of a grave injustice or the administration of justice being brought into disrepute. If the President had determined that exceptional circumstances are present, it implies the likelihood of either of the two jurisdictional factors if the order refusing leave is not reconsidered.

[83]Thus, if there are exceptional circumstances (as determined by the President on the approach adopted in the second judgment), the court would logically be inclined to reconsider the order refusing leave to appeal to prevent a likelihood of grave injustice or of the administration of justice being brought into disrepute. This must be so unless there are weighty considerations that militate against reconsideration despite the presence of exceptional circumstances. A court coming to such a conclusion should clearly articulate what those considerations are.

[84]The second judgment gives no reasoned basis why, if exceptional circumstances are present (as determined by the President), it concludes that the order granting leave should nevertheless not be reconsidered. It simply says that the two judges did not err in refusing leave to appeal. If the two judges did not err, the President’s determination (on the approach adopted in the second judgment) that exceptional circumstances are present must be wrong and that determination is (impermissibly) reviewed by the court. 32

[85]4 Seasons exposes this very weakness. Having found that the President had already determined that exceptional circumstances were present, the Court in that case held that ‘the critical question’ was whether there were exceptional circumstances. It concluded, ‘after anxious consideration [that] no exceptional circumstances . . . are discernible.’ 48 This is an irretrievable internal contradiction. Order

[86]The following order is made: 1 The application is struck from the roll. 2 The applicants shall pay the costs incurred by the respondent in opposing the application for reconsideration, including the costs of two counsel. ___________________________ T MAKGOKA JUDGE OF APPEAL Mbatha ADP (dissenting): Introduction

[87]I have read the judgment of my Brother, Makgoka JA, which deals, amongst other things, with the application of s 17(2)(f). It affirms the position adopted in Motsoeneng, followed by Bidvest, in that the threshold question of exceptional circumstances (the term employed by the first judgment to encapsulate a grave failure of justice or bringing the administration of justice into disrepute) should be determined by this Court. This is the first issue where I do not find common ground with him. 48 4 Seasons fn 26 para 63. 33

[88]On reflection and consideration of subsequent judgments which followed Bidvest, I disagree with the conclusion reached regarding s 17(2)(f). It is important that I should highlight how the provisions of s 17(2)(f) evolved. In Avnit, this Court stated that the origin of the section no doubt lies in the situation that arose in Van der Walt v Metcash Trading Co Ltd (Van der Walt). 49 In Van der Walt this Court on successive days in August 2001 made contrary orders in two cases which were materially identical. They were made in response to petitions to appeal against orders of the High Court in summary judgment applications. As a result, s 17(2)(f) was conceptualised out of this acknowledgement of fallibility. Avnit was decided in chambers by the then President of this Court, Mpati P established from the outset that the power to determine the existence of exceptional circumstances resided with the President.

[89]Avnit makes it clear that it is the President who has to be satisfied that exceptional circumstances exist. It states that: ‘In the context of s 17(2)(f) the President will need to be satisfied that the circumstances are truly exceptional before referring the considered view of two judges of this court to the court for reconsideration. I emphasise that the section is not intended to afford disappointed litigants a further attempt to procure relief that has already been refused. It is intended to enable the President of this Court to deal with a situation where otherwise injustice might result. An application that merely rehearses the arguments that have already been made, considered and rejected will not succeed, unless it is strongly arguable that justice will be denied unless the possibility of an appeal can be pursued. A case such as Van der Walt may, but not necessarily will, warrant the exercise of the power. In such a case the President may hold the view that the grant of leave to appeal in the other case was inappropriate.’ 50 (Emphasis added.) 49 Avnit fn 17 para 3. 50 Van der Walt v Metcash Trading Limited [2002] ZACC 4; 2002 (4) SA 317; 2002 (5) BCLR 454 (CC). 34 The evolution of s17(2)(f)

[90]This trend of considering s 17(2)(f) applications continued for several years. In the year 2014 to 2019 five cases on s 17(2)(f) were decided by the then President of this Court. I mention S v Gwababa, 51 S v Malele, 52 Hough v Sisilana (Hough), 53 S v Ntlanyeni, 54 and S v Mathekola. 55 All the aforementioned judgments followed the Avnit decision and were considered in chambers by the President.

[91]In S v Mathekola, which was referred to court for determination by the President, presided over by Lewis JA, the court did not delve into the question whether exceptional circumstances existed or not. It proceeded directly to determine the matter under s 17(2)(d). In all the cases that were referred for reconsideration in the period 2014 to 2019, there was no consideration of whether exceptional circumstances existed or not.

[92]The same trend continued post 2019. I point out that other applications for reconsideration were dismissed outright by the President in chambers. Amongst others, I highlight the decision of Maya P in Hough, where she dismissed the application for reconsideration and gave reasons for her order. None of the s 17(2)(f) applications that were dismissed by the President, were ever referred to the court to re-determine the existence of exceptional circumstances. This historical approach to the interpretation of s 17(2)(f) is consistent with an 51 S v Gwababa [2016] ZASCA 200; 2016 JDR 2291 (SCA). In the reconsideration hearing that followed, in S v Gwababa [2018] ZASCA 152; 2018 JDR 1951 (SCA), this Court described the case before it, para 1, as a case arising ‘from a decision of the then Acting President of this court, (Maya P) in terms of s 17(2)(f) of the Superior Courts Act 10 of 2013 (the Act) to refer a decision of two judges of this court, dismissing an application for leave to appeal, for reconsideration by this court.’ This Court then proceeded to reconsider whether leave to appeal should be granted without further consideration of s 17(2)(f). 52 S v Malele [2016] ZASCA 115; 2016 JDR 1686 (SCA). 53 Hough v Sisilana [2018] ZASCA 4; 2018 JDR 1283 (SCA). 54 S v Ntlanyeni [2016] ZASCA; 2016 (1) SACR 581 (SCA). 55 S v Mathekola [2016] ZASCA 106. This reference can be found in the reconsideration hearing S v Mathekola [2017] ZASCA 100; 2017 JDR 1414 (SCA). 35 interpretation that the power to determine exceptional circumstances belongs exclusively to the President.

[93]The change to this established practice, that the President or Acting President of this Court decides on the existence of exceptional circumstances, was introduced by the judgment in Motsoeneng. In Motsoeneng, this Court found that the court effectively steps into the shoes of the two judges, who refused leave or special leave to appeal. 56 A proposition which I agree with. It held further that this must involve no more than satisfying the requirements for special leave to appeal, which is a point of departure for me. In its finding on the existence of exceptional circumstances it provided as follows: ‘Given that there are no reasonable prospects of success in the contemplated appeal, much less special circumstances, the application hardly meets the higher “exceptional circumstances” threshold set by s 17(2)(f). It must accordingly fail’. 57

[94]The effect of the jurisdictional fact approach is that it has the effect of stepping into the shoes of the President and stepping into the shoes of the two judges, but doing so with a different threshold. It established that the first step of what the court should do, before the application for leave or special leave to appeal is reconsidered, is to determine whether exceptional circumstances exist.

[95]In its reasoning Motsoeneng found as follows: ‘The necessary prerequisite for the exercise of the President’s discretion is the existence of ‘exceptional circumstances’. If the circumstances are not truly exceptional, that is the end of the matter. The application under subsection 2(f) must fail and falls to be dismissed. If, however, exceptional circumstances are found to be present, it would not follow, without more, that the decision refusing leave to appeal must be referred to the court for reconsideration. The President may, in the exercise of her discretion, nonetheless decline to do so. If the President refers the decision of the two judges for 56 See Motsoeneng v South African Broadcasting Corporation Soc Ltd and Others [2024] ZASCA 80; 2025 (4) SA 122 (SCA) para 14. 57 Ibid para 20. 36 reconsideration, the court effectively steps into the shoes of the two judges. Upon reconsideration, it may grant or refuse the application and, if the former, vary the order of the two judges dismissing the application to one granting leave either to this Court or the relevant high court.’ 58 (Emphasis added.)

[96]In my view, the dictum quoted above discloses three critical findings regarding s 17(2)(f). First, that the President is vested with a discretion to grant or refuse the application, though that discretion may only be exercised where exceptional circumstances exist. Second, by refusing the application, the President has exercised that discretion and determined the existence of exceptional circumstances and that once the discretion is exercised that is the end of the matter. Third, where the President has exercised the discretion by referring the matter to the Court for reconsideration, the Court seized with the reconsideration steps into the shoes of the initial two judges who considered the application for leave to appeal. In terms of section 17(2)(b) of the SC Act, the initial two judges are enjoined to determine whether the appeal enjoys reasonable prospects of success, or whether compelling reasons exist warranting the granting of leave to appeal. Those should then be the two issues before the reconsideration Court. The question of exceptional circumstances falls outside the ambit of its inquiry – it is required to determine whether the applicant seeking leave has established reasonable prospects of success or compelling reasons. In the absence of either, leave to appeal ought to be refused. Put another way, any consideration of matters beyond those contemplated by s 17(2)(b) read with s 17(1)(a) would be inconsistent with the Court standing in the shoes of the two judges. 59 58 Motsoeneng para 14. 59 In the event of an initial application for special leave to appeal, the same would apply, with the changes required by the context. The only issues before the reconsideration court, would be the threshold for special leave to appeal. Regarding the applicability of s17(2)(f) to initial applications for special leave to appeal, s17(3) of the SC Act determines that ‘An application for special leave to appeal under section 16(1)(b) may be granted by the Supreme Court of Appeal on application filed with the registrar of that court within one month after the decision sought to be appealed against, or such longer period as may on good cause be allowed, and the provisions of subsection (2)(c) to (f) shall apply with the changes required by the context.’ 37

[97]Notwithstanding what is stated in the passage quoted above, the Court in Motsoeneng held that it was obliged to consider whether the applicant had demonstrated the existence of exceptional circumstances warranting the reconsideration of the granting of leave to appeal. It did so by holding that: ‘Regrettably, the parties misconceived the true nature of the enquiry. In the heads of argument filed on behalf of both, the sole focus was wrongly on the correctness of the judgment of Khan AJ, and whether or not there were reasonable prospects of success in the contemplated appeal against that judgment. Counsel appeared not to appreciate that the requirement of the existence of exceptional circumstances is a jurisdictional fact that had to first be met, and that, absent exceptional circumstances, the s 17(2)(f) application was not out of the starting stalls. At the bar, counsel sought to rehash the arguments that had already been advanced before the high court and before the two judges of this Court, who dismissed the application for leave to appeal. But those contentions had been considered and found to be wanting in a detailed judgment of the high court. In dismissing the petition, the two judges of this Court had self-evidently taken the view that there were no reasonable prospects of an appeal against that judgment succeeding.’ 60 (Emphasis added.)

[98]Implicit in this approach is a finding that the President had not herself determined the question of exceptional circumstances. Although the Court did not expressly say so, its conclusion that it was empowered to consider the issue of exceptional circumstances admits of no other conclusion. Had it accepted that the President had already determined that issue, the Court would have been precluded from revisiting it.

[99]Significantly, Motsoeneng did not find that Avnit was wrongly decided. Avnit stands today as an unreversed judgment of this Court and has never been found to be clearly wrong. As the court of first resort on the interpretation of s 17(2)(f), Avnit remains the governing authority. Motsoeneng itself referred extensively to Avnit but, in departing from its underlying framework without 60 Motsoeneng fn 4 para 19 38 expressly overruling it, created precisely the kind of jurisprudential inconsistency now manifest in the conflicting decisions of this Court. The jurisdictional fact interpretation per Bidvest

[100]On the other hand, the Bidvest judgment identified two possible interpretations of s 17(2)(f), the ‘exclusivity interpretation’ and the ‘jurisdictional fact interpretation’. It advocated for the jurisdictional fact interpretation where the court has to decide whether there are exceptional circumstances. The flaw in this interpretation is that the court can only decide on the existence of the exceptional circumstances in matters referred to it by the President. It is silent on matters in which the President dismissed the application for reconsideration in chambers. These interpretations impact on the right to equality before the law. Section 9(1) of the Constitution provides that all persons are equal before the law and has the right to equal protection and benefit of the law.

[101]As the minority in S v Godloza 61 (Godloza) correctly observed: ‘The approach in Bidvest and that of the majority in Schoeman also seems to leave out of account how the court will determine whether exceptional circumstances are present when, as in this matter, the President of the Supreme Court of Appeal refuses reconsideration, because the matter will not be referred to the court. Surely the President of the Supreme Court of Appeal cannot have the power to determine whether exceptional circumstances are present for purposes of refusing reconsideration, but not for purposes of granting it?' I endorse this reasoning. It reveals a decisive logical gap in the Bidvest framework.

[102]The interpretation given to s 17(2)(f) in Bidvest and applied in the first judgment is, with respect, incorrect, as it fails to cater for the scenario where the President dismisses the reconsideration application outright. No court reviews the 61 S v Godloza and Another [2025] ZACC 24; 2026 JDR 0431 (CC) para 145. 39 decision to dismiss the s 17(2)(f) application by the President. This is a clear indication that these powers are to be exercised only by the President of this Court. It is for this reason that I find that the jurisdictional fact interpretation finds no application to s 17(2)(f). The interpretation given in the Lorenzi and Schoeman minority judgments emphasised that what is referred for reconsideration is not the exercise by the President of her discretion, but the refusal by the two judges, who refused to grant leave or special leave to appeal. Most significantly they advocate that it is only the President who has powers to determine whether there are exceptional circumstances or not. I endorse this view as it shows that the referral to the President was not intended to be a mere bureaucratic process.

[103]When one considers the trite principles of interpretation it is clear that the President alone is empowered to determine the existence of exceptional circumstances. There is no explanation from Bidvest or Motsoeneng as to what the President’s functions are when the matter serves before the President in terms of s 17(2)(f). Bidvest merely advocates that it could not have been intended that the President sits as a court on his or her own account. Were the President merely to act as a post box or halfway-house, I have no doubt that the language of the provision would have explicitly stated so.

[104]The review of the President’s decision to refer the application for leave or special leave to appeal for reconsideration leads to an injustice to litigants. The advocated approach followed in the first judgment obviously prejudices the applicant, through this double scrutiny. The majority in Schoeman, which followed the approach in Bidvest, had this to say: ‘. . . This Court does not review or reconsider the President's decision in any way. This Court alone decides whether there are exceptional circumstances. On the jurisdictional fact interpretation, this Court alone enjoys the competence to do so, and the President does not. If this Court should find that there are exceptional circumstances. . . the President's discretionary 40 judgment that the decision on petition should be reconsidered stands. That decision is not reviewed or subject to correction by this Court. It is not subject to any kind of judicial validation by this Court.’ 62

[105]As much as it states that it does not review the decision of the President, that is what happens when the court re-determines the existence of exceptional circumstances. The President of this Court is a judicial officer, and nothing precludes her or him from applying her or his mind to judicial matters. It would have been a different case, had the provision referred to the Registrar of this Court who performs administrative duties and quasi-judicial functions.

[106]The first judgment, in following Bidvest, effectively strips the President of the substantive judicial function that s 17(2)(f) vests in her or him. The President of this Court is not an administrative officer. The President is a judicial officer exercising the judicial function of the Supreme Court of Appeal. It would be anomalous to treat the President's referral as a mere bureaucratic step — a conduit or post-box — devoid of any substantive legal consequence. Had the legislature intended the President's function to be merely procedural, it would have been simple enough to vest the referral power in the Registrar or another administrative official.

[107]The jurisdictional fact interpretation reads into the statute a function — judicial reassessment of the threshold — for which the provision makes no express provision. The exclusivity interpretation, by contrast, follows naturally from the words used. When the President refers a matter in circumstances where a grave failure of justice would otherwise result, she has already determined that those circumstances exist. What the court is asked to do thereafter is decide the substance of the application for leave or special leave to appeal. Nothing more. 62 Schoeman fn 35 para 68. 41

[108]The exclusivity interpretation, by contrast, achieves certainty, consistency and constitutional conformity. It gives full effect to the President's statutory powers. It avoids duplicating judicial functions. It treats all litigants equally, regardless of whether the President refers the matter or dismisses it in chambers. And it is in harmony with the language of the provision, read as a whole.

[109]The reliance on Bidvest in the first judgment is at cross-purposes with the language of s 17(2)(f). The same sentiment is expressed by Petse AJA in 4 Seasons 63 where he stated that: ‘The passage quoted from Schoeman . . . begs the question as to why then, if this is how the language of the proviso should be construed, the legislature did not explicitly say so. The upshot of this passage implies that in referring the matter to the Court for reconsideration the President require[s] the Court, “if [the Court] should find that there are exceptional circumstances” to reconsider the decision of the two judges who refused or granted leave to appeal. . . To my mind this is, with respect, a convoluted reasoning that is subversive of the manner in which meaning to the words used in a statutory instrument is ascribed.’ The text and structure of s 17(2)(f)

[110]It is important that I quote the current wording of s17(2)(f) as it appears in the SC Act. Section 17(2)(f) provides as follows: ‘The decision of the majority of the judges considering an application [for leave to appeal], or the decision of the court, as the case may be, to grant or refuse the application shall be final: Provided that the President of the Supreme Court of Appeal may, in circumstances where a grave failure of justice would otherwise result or the administration of justice may be brought into disrepute, whether of his or her own accord or on application filed within one month of the decision, refer the decision to the court for reconsideration and, if necessary, variation.’ (Emphasis added) The Bidvest judgment and the first judgment do not provide an explanation for the President’s approach to the phrase ‘whether of his or her own accord’. The failure to do so reveals a discrepancy in their reasoning. This phrase cannot be 63 4 Seasons fn para 46. 42 read disjunctively from the rest of the provision. It has to be interpreted conjunctively with the rest of the provision.

[111]The only logical interpretation to give to the phrase ‘whether of his or her own accord’ is that if she independently discovers, or is alerted to, an inconsistency or if she is of the view that a grave failure of justice would otherwise result or the administration of justice may be brought into disrepute by any of the decisions of this Court to grant or refuse an application for leave or special leave to appeal, she has the freedom to independently refer that decision for reconsideration. This further supports the notion that she independently determines the existence of exceptional circumstances. In such cases she initiates the reconsideration of a dismissed application for leave to appeal, without a formal application from the parties. This discretion enables the President to act proactively to prevent an injustice, particularly where exceptional circumstances exist. This is a parallel process to the more common option of an ‘application filed within one month of the decision’.

[112]The provision allows the President to correct errors in the refusal (or the granting) 64 of leave to appeal without necessitating a party to lodge a formal application. The President can only follow this route, provided that she is satisfied that the administration of justice would be brought into disrepute or a grave injustice would result. This is an extraordinary remedy, which can only be exercised by the President. The inclusion of this phrase resonates with the primary rationale for the inclusion of the phrase ‘where the administration of justice may 64 See the reference to this possibility in Mazizini Community v Minister of Rural Development and Land Reform [2020] ZASCA 57; [2020] 3 All SA 318 (SCA); 2020 JDR 0913 (SCA) para 21. The registrar of this Court had directed that the application be withdrawn, because it did not cite all the necessary parties. This was never done, nor was the original application withdrawn. ‘Leave to appeal was therefore granted, clearly in error, in the absence of affidavits from the parties that had to be cited but were not. The registrar advised the attorneys for the party that was left out that there was no statutory provision for rescission of an order granting leave to appeal, but only a reconsideration in terms of s 17(2)(f). It was however decided, for the sake of expedience, rather to allow the appeal to proceed.’ 43 be brought into disrepute’, which was included in the new wording of s 17(2)(f). 65 It allows for reconsideration if a serious failure arises primarily from the errors that arise within the court system. The President does not form a prima facie view that exceptional circumstances might be present. In fact, he or she finally determines the issue. The questions on precedent

[113]The recent judgments in 4 Seasons, Matsi and Lutzkie referred to in the first judgment, as being decisions of three judge panels, and as therefore having no binding effect on Motsoeneng and Bidvest are noted. I deal with this contention in more detail below. The first judgment, however, fails to appreciate that Motsoeneng did not find that Avnit was wrongly decided. Avnit stands till today and is, in terms of the doctrine of stare decisis, the decision that should be followed. It was the first case since the introduction of s 17(2)(f) and has not been found to be ‘clearly wrong’. 66 In fact, Motsoeneng refers to the Avnit judgment quite extensively. 67

[114]The judgments in 4 Seasons, Matsi and Lutzkie also follow Avnit rather than the process advocated in Bidvest. The latter judgment duplicated the assessment of exceptional circumstances and rendered the President’s assessment of exceptional circumstances meaningless. The Bidvest view, as found by the minority judgments in Lorenzi and Schoeman, is flawed, as it is in conflict with Liesching II and Avnit. On that score, I also find that Bidvest did not set aside Avnit. 65 Section 28 of the Judicial Matters Amendment Act 15 of 2023 66 Bloemfontein Town Council v Richter 1938 AD 195 at 232 where the following is stated: ‘The ordinary rule is that this Court is bound by its own decisions and unless a decision has been arrived at on some manifest oversight or misunderstanding that is there has been something in the nature of a palpable mistake a subsequently constituted Court has no right to prefer its own reasoning to that of its predecessors - such preference, if allowed, would produce endless uncertainty and confusion. The maxim "stare decisis" should, therefore, be more rigidly applied in this the highest Court in the land, than in all others.' 67 See Motsoeneng para 18. See also para 3 above. 44

[115]I acknowledge that the doctrine of stare decisis requires that this Court to follow its own previous decisions and those of the Constitutional Court to ensure legal certainty and consistency. However, this Court may depart from them if convinced that the previous decision was clearly wrong, based on a fundamental misconception of the law. 68 Therefore, courts may depart from the stare decisis if it conflicts with constitutional values. The practical and constitutional consequences

[116]A decision which is in conflict with constitutional values, like Bidvest and the first judgment, is clearly wrong. Therefore, it would follow that those judgments can be departed from. The process advocated by Bidvest is constitutionally wrong as it does not cater for a situation where the President dismisses a s 17(2)(f) application.

[117]I affirm the view that one may depart from the doctrine of stare decisis, as long as the interpretation advocated for is in line with s 39(2) of the Constitution. This requires that ‘[w]hen interpreting any legislation, and when developing the common law or customary law, every court, tribunal or forum must promote the spirit, purport and objects of the Bill of Rights’. Of crucial importance is that the interpretation which I advocate for, promotes the right in s 9(1) of the Constitution, which provides that all persons have the right to equal protection and benefit of the law. The proliferation of a number of decisions from a three- judge panel and several dissenting judgments taking a different view, in particular from Bidvest, indicates the unsoundness of that approach. This on its own appears to be at odds with the rule of law, which is a founding value of the Constitution. On this basis the interpretation I propose should triumph, because it impacts on the rights of litigants. 68 Patmar fn 44 para 4. 45

[118]Promoting the spirit, purport and object of the Bill of Rights invokes equality as a value and as a right. The approach to be adopted is the one set out in Veldman v Director of Public Prosecutions 69 where Ngcobo J, in a separate concurring judgment, stated as follow: ‘The “spirit, purport and objects of the Bill of Rights” is to be gleaned from the rights guaranteed in the Bill of Rights which is 'the cornerstone' of our constitutional democracy.’

[119]In Investigating Directorate: Serious Economic Offences and Others v Hyundai Motor Distributors (Pty) Ltd, 70 Langa DP said that interpreting legislation to promote the spirit, purport and objects of the Bill of Rights means: ‘. . . that all statutes must be interpreted through the prism of the Bill of Rights. All law-making authority must be exercised in accordance with the Constitution. The Constitution is located in a history which involves a transition from a society based on division, injustice and exclusion from the democratic process to one which respects the dignity of all citizens, and includes all in the process of governance. . . The purport and objects of the Constitution find expression in section 1

[71]which lays out the fundamental values which the Constitution is designed to achieve. The Constitution requires that judicial officers read legislation, where possible, in ways which give effect to its fundamental values. Consistently with this, when the constitutionality of legislation is in issue, they are under a duty to examine the objects and purport of an Act and to read the provisions of the legislation, so far as is possible, in conformity with the Constitution. . . . Accordingly, judicial officers must prefer interpretations of legislation that fall within constitutional bounds over those that do not, provided that such an interpretation can be reasonably ascribed to the section.’ 69 Veldman v Director of Public Prosecutions, Witwatersrand Local Division 2007 (3) SA 210 (CC) para 74. Ngcobo J writes that whereas in the past the principle of parliamentary supremacy had meant that fundamental rights were frequently eroded or excluded by legislation, the advent of a constitutional democracy had guaranteed and protected fundamental rights. 70 Investigating Directorate: Serious Economic Offences and Others v Hyundai Motor Distributors (Pty) Ltd and Others In re: Hyundai Motor Distributors (Pty) Ltd and Others v Smit NO and Others [2000] ZACC 12; 2001 (1) SA 545 (CC); 2000 (2) SACR 349 (CC); 2000 (10) BCLR 1079 (CC)paras 21- 23. 71 Section 1 of the Constitution reads as follows: The Republic of South Africa is one, sovereign, democratic state founded on the following values: (a) Human dignity, the achievement of equality and the advancement of human rights and freedoms. (b) Non-racialism and non-sexism. (c) Supremacy of the constitution and the rule of law. (d) Universal adult suffrage, a national common voters roll, regular elections and a multi-party system of democratic government, to ensure accountability, responsiveness and openness. 46

[120]This position was also affirmed in Bato Star Fishing (Pty) Ltd v Minister of Environmental Affairs and Tourism. 72 It deals with the meaning of ‘spirit, purport and object’, connected to equality as follows: ‘The Constitution is now the supreme law in our country. It is therefore the starting point in interpreting any legislation. Indeed, every court “must promote the spirit, purport and objects of the Bill of Rights” when interpreting any legislation. That is the command of s 39(2). Implicit in this command are two propositions: first, the interpretation that is placed upon a statute must, where possible, be one that would advance at least an identifiable value enshrined in the Bill of Rights; and second, the statute must be reasonably capable of such interpretation. This flows from the fact that the Bill of Rights “is a cornerstone of [our constitutional] democracy.” It “affirms the democratic values of human dignity, equality and freedom.” In interpreting . . . we must promote the values of our constitutional democracy. But what are these values? . . . This society is to be built on the foundation of the values entrenched in the very first provision of the Constitution. These values include human dignity, the achievement of equality, and the advancement of human rights and freedoms. The achievement of equality is one of the fundamental goals that we have fashioned for ourselves in the Constitution.’ .’ (Citations omitted.)

[121]In Cipla Medpro (Pty) Ltd v Aventis Pharma SA and Related Appeal, 73 this Court also emphasised that s 39(2) of the Constitution calls upon the courts to 'promote the spirit, purport and objects of the Bill of Rights' when interpreting legislation, but that would not open the door to changing the clear meaning of a statute. If the clear meaning of a statute or statutory provision would be in conflict with the Bill of Rights, then the remedy would be to strike it down. Similarly, in Holeni v Land and Agricultural Development Bank of South Africa, 74 this Court held that it appeared that the spirit, purport and objects of the Bill of Rights were best served by permitting the bank to conduct its affairs within the confines of 72 Bato Star Fishing (Pty) Ltd v Minister of Environmental Affairs and Tourism and Others [2004] ZACC 15; 2004 (4) SA 490 (CC); 2004 (7) BCLR 687 (CC) paras 72- 74. 73 Cipla Medpro (Pty) Ltd v Aventis Pharma SA and Related Appeal [2012] ZASCA 108; 2013 (4) SA 579 (SCA). 74 Holeni v Land and Agricultural Development Bank of South Africa [2009] ZASCA 9; 2009 (4) SA 437 (SCA); [2009] 3 All SA 22. 47 its empowering statute and to restrict the meaning of State in s 11(b) of the Act as referred to above. The panel size question

[122]The first judgment holds that 4 Seasons, Matsi and Lutzkie, all decisions of three-judge panels, could not competently overrule Motsoeneng and Bidvest, which are unanimous five-panel judgments. With respect, this reasoning misapprehends the constitutional and statutory character of this Court.

[123]This requires a consideration of the provisions of s 13 of the SC Act, which I repeat below: ‘13 Manner of arriving at decisions by Supreme Court of Appeal (1) Proceedings of the Supreme Court of Appeal must ordinarily be presided over by five judges, but the President of the Supreme Court of Appeal may - (a) direct that an appeal in a criminal or civil matter be heard before a court consisting of three judges; or (b) whenever it appears to him or her that any matter should in view of its importance be heard before a court consisting of a larger number of judges, direct that the matter be heard before a court consisting of so many judges as he or she may determine. (2) (a) The judgment of the majority of the judges presiding at proceedings before the Supreme Court of Appeal shall be the judgment of the court. (b) Where there is no judgment to which a majority of such judges agree, the hearing must be adjourned and commenced de novo before a new court constituted in such manner as the President of the Supreme Court of Appeal may determine. (3) If, at any stage after the hearing of an appeal has commenced, a judge of the Supreme Court of Appeal is absent or unable to perform his or her functions, or if a vacancy among the members of the court arises - (a) the hearing must, where the remaining judges constitute a majority of the judges before whom the hearing was commenced, proceed before the remaining judges, and the decision of a majority of the remaining judges who are in agreement shall, if that majority is also a majority of the judges before whom the hearing was commenced, be the decision of the court…’ 48

[124]Section 13(1) of the SC Act provides that proceedings of this Court must ordinarily be presided over by five judges, but that the President may direct that an appeal be heard before a court of three judges, or that any matter of importance be heard before a court of seven or more judges. The critical point is that the President's power to constitute a smaller or larger panel is a matter of practice and expediency — not of hierarchy or authority. The SC Act does not vest any greater juridical authority in a five-judge panel than in a three-judge panel. When the President directs that three judges hear a matter, those three judges constitute this Court in the same manner as five would. Their decision is the decision of this Court, with equal force and authority.

[125]The leading authority in the United Kingdom on this question is instructive. In Young v Bristol Aeroplane Company Ltd (Bristol Aeroplane Company), 75 Lord Greene MR affirmed that the Court of Appeal is 'one court', regardless of the number of judges sitting. A full court of five judges has no greater powers or jurisdiction than a division of three. This was affirmed in Williams v Fawcett, 76 where Sir Donaldson MR stated that it is well established that a five-judge Court of Appeal has no more authority than a three-judge Court of Appeal. I am enjoined by s 39(1) of the Constitution to consider such foreign law in the interpretation of legislation, and I find no reason why the same principle should not apply here.

[126]This Court is constitutionally and statutorily 'one court'. The number of judges sitting, whether three, five or seven, does not alter its jurisdiction, authority or the binding force of its decisions. I reiterate that any differentiation in panel size is practical rather than structural or hierarchical. A three-judge panel of this Court deciding a point of law creates binding precedent of equal weight to 75 Young v Bristol Aeroplane Company, Limited [1944] 1 KB 718. 76 Williams v Fawcett [1985] 1 All ER 787. 49 that of a five-judge panel. The first judgment's assertion that 4 Seasons, Matsi and Lutzkie were incompetent to overrule Motsoeneng and Bidvest solely because of the smaller panel size is, therefore, with respect, incorrect.

[127]Moreover, this Court is entitled to depart from its own prior decisions where those decisions are clearly wrong or would perpetuate a conflict with the Constitution. A decision which, as Bidvest does, creates a procedure that discriminates between litigants depending on whether their matter is referred or dismissed by the President, and which effectively disempowers the President in a manner inconsistent with the express statutory language, and conflicts with constitutional values. An interpretation that advances equality before the law and gives meaningful effect to the President's constitutionally embedded judicial function must prevail over one that does not, regardless of the size of the bench that previously endorsed the latter. The challenges identified in the first judgment

[128]I am constrained to identify the specific respects in which the first judgment fails to answer the dissent adequately, and where the first judgment’s own reasoning is vulnerable. The first judgment opens with the observation that a proliferation of disparate views on s 17(2)(f) would not redound to the benefit of this Court, and that the duty to uphold binding authority lies at the heart of our system of precedent. These propositions command agreement. They do not, however, resolve the anterior question whether Motsoeneng and Bidvest were correctly decided. That question, whether a court sitting in terms of s 17(2)(f) is constitutionally obliged to determine afresh whether exceptional circumstances exist, is a live question of statutory interpretation under a constitutional imperative. The obligation to follow precedent is not absolute in those circumstances. 50

[129]In Avnit, the President of this Court, sitting as a single judge in the exercise of the power conferred by s 17(2)(f), addressed the threshold question directly. He observed that the President must be satisfied that the circumstances are truly exceptional before referring a decision for reconsideration. That observation reflects the structure of s 17(2)(f): the legislature reposed the threshold judgment in the President, who acts as a gatekeeper. Once the President has referred the matter, the referral court's function is to reconsider and, if necessary, vary the decision, not to re-examine whether the referral was justified.

[130]The first judgment holds that the statement in Avnit is no more than a recitation of the statutory text and does not constitute a ratio decidendi on the repository of power under s 17(2)(f). That characterisation is too narrow. Schreiner JA’s test in Levinson, which the first judgment endorses, asks whether the reasons given in a judgment were necessary for the decision along the lines actually followed. In Avnit, the President was called upon to exercise the power under s 17(2)(f). He could do so only if satisfied that the exceptional circumstances threshold was met. His conclusion that the threshold assessment vests in the President was the operative legal basis for exercising the power. It was necessary for the decision in that case. It is, accordingly, binding.

[131]The first judgment's reliance on panel size to discount 4 Seasons, Matsi and Lutzkie rests on a premise that the SC Act does not support. Section 13(1) empowers the President to constitute a panel of three, five or more judges as the nature of the matter requires. The provision is one of administrative allocation, not a matter of juridical hierarchy. The Act nowhere provides that decisions of a larger panel bind those of a smaller one, or that a five-judge panel has greater authority than a three-judge panel. When the President directs that three judges hear a matter, those three judges are this Court, and their decision is the decision of this Court with the full force that it entails. 51

[132]This is not merely a structural point. It is confirmed by s 13(2)(a) of the SC Act, which provides that the judgment of the majority of the judges presiding shall be the judgment of the court. The provision does not qualify the authority of that judgment by reference to the number constituting the majority. A majority of three is as much the judgment of the court as a majority of five. To hold otherwise would be to read into the statute a hierarchical distinction that the legislature did not enact and that the constitutional character of this Court, as a single collegiate body, does not permit.

[133]It follows that the question is not one of panel size but of precedential conflict. Where two decisions of this Court are in conflict, this Court must determine which, if either, was correctly decided. The first judgment, having correctly identified the conflict, was required to resolve it on the merits — by asking which interpretation of s 17(2)(f) is correct — and not by applying a hierarchical metric that the law does not recognise.

[134]The first judgment reaffirms the principle, settled in Luphondo that a three- member bench cannot overrule a five-member bench. That principle is not in dispute. What is in question is its application here. Luphondo addressed the bench-size principle in the context of conflicting s 17(2)(f) authorities. It did not engage with the antecedent question whether Motsoeneng and Bidvest were correctly decided in the light of Avnit and the constitutional framework. It does not foreclose the possibility that a five-member bench, such as this one, may revisit the question, which is what this Court is empowered to do. The Luphondo rule, if applied mechanically, would mean that any five-judge panel could immunise its reasoning from correction by any three-judge panel, regardless of the correctness of the reasoning, an outcome that is constitutionally problematic. 52

[135]Bidvest is clearly wrong for the following cumulative reasons. First, it creates an asymmetry of procedure that s 17(2)(f) does not contemplate: litigants whose applications are referred by the President receive independent judicial scrutiny of the threshold by the full court, while those whose applications are dismissed in chambers receive none. This produces unequal treatment in a process which fundamental purpose is to provide a safety valve against injustice. Secondly, it treats the President's referral as devoid of substantive content, contrary to the plain language of the provision and the constitutional status of the President as a judicial officer. Thirdly, it departs from Avnit without finding it clearly wrong, and without expressly overruling it, thereby introducing precisely the kind of juridical uncertainty that stare decisis is designed to prevent. Each of these errors, individually, is a ground for departure. Collectively, they compel it.

[136]More fundamentally, the first judgment does not address the constitutional point: that s 13(1) of the SC Act confers no greater juridical authority on a five- judge bench than on a three-judge bench, and that the power to constitute panels of different sizes is a matter of practice rather than hierarchy. If the statute itself draws no distinction in authority between panels of different sizes, the Luphondo rule — which the first judgment treats as settled — may itself require constitutional scrutiny. The first judgment does not engage with this.

[137]A court is not obliged to follow precedent that is incompatible with constitutional values or that was decided per incuriam. The qualification is narrow and must be applied with great caution. Where a binding authority fails to account for a constitutional provision directly applicable to its subject matter, a subsequent court is not precluded from so finding and, where appropriate, from departing from that authority on that ground. Whether the present case calls for a departure in the full sense is a matter the Court need not resolve, given that the 53 alternative approach set out below also yields a different result from the first judgment. The mischaracterisations in the first judgment

[138]The first judgment’s engagement with the second judgment requires that I clarify some of the issues raised therein. As a result, the mischaracterisations are addressed in turn. The first judgment holds that Mpati P’s observations in Avnit were ‘no more than a regurgitation of the provisions of s17(2)(f)’ and therefore did not constitute a ratio decidendi on the repository of the power to determine exceptional circumstances. As a result, Avnit, ‘did not lay down any precedent on the interpretative question’, so later judgments were free to decide it afresh without declaring Avnit clearly wrong.

[139]The aforementioned is a mischaracterisation, and it is structural. It rests on the premise that Avnit decided only on what constitutes exceptional circumstances and that the question ‘who decides’ simply did not arise. That premise misscribes what Mpati P was doing. He was exercising the s 17(2)(f) power as a single judge and not engaged in an academic exercise. The operative question before him was whether to refer or not. President Mpati answered it by satisfying himself that the circumstances were not truly exceptional, and on that footing he declined to refer. The proposition that the President must be so satisfied was the legal premise on which the disposition turned. In Avnit, had Mpati P not held that the President’s satisfaction as to exceptional circumstances was the gateway to the power, he could not have declined to refer on the ground that the threshold was not met. That reasoning was essential to the integrity and outcome of the judgment. It was not subsidiary to any other principle.

[140]The aforementioned is in line with the first judgment’s authority in Pretoria City Council, where Schreiner JA asked whether the reasons given were 54 necessary for the decision, in the sense that along the lines actually followed the result would have been different but for those reasons. On the first judgment’s own test Avnit qualifies as a ratio decidendi.

[141]The first judgment’s description of Avnit at the level of generality, by stating that Avnit addressed ‘what’ and ‘why’ and never ‘who’ is misplaced. The locus of power and the content of the threshold are inseparable. In Avnit, the main issue is that it is the President who must be satisfied. The second judgment’s contention is that Motsoeneng departed from Avnit’s framework without ever finding it clearly wrong, which is wholly left unanswered by the first judgment. It is treated as irrelevant. The first judgment surprisingly invokes Avnit approvingly at para 41 as authority for the disposition. At the same time, the first judgment refers to it as ‘a recitation’ when the second judgment relies on the adjacent reasoning in the same judgment. Either Avnit carries precedential weight on the operation of s 17(2)(f) or it does not. It cannot be both ways.

[142]The second mischaracterisation of the second judgment rests on the following premise: it treats ‘exceptional circumstances’ (the threshold for referral) and ‘reasonable prospects of success or compelling reasons’ (the threshold for leave) as though they were a single enquiry. They are not, as clearly set out by the second judgment. The second judgment plainly set out above, that on referral, the court ‘steps into the shoes of the two judges’. The two judges, under s 17(2)(b) read with s 17(1)(a), decide one thing and one thing only: whether the appeal enjoys prospects of success or whether there are compelling reasons for it to be heard. That, says the second judgment, is the question for reconsideration by the court. The question of exceptional circumstances ‘falls outside the ambit of its inquiry’. That is so, because the President has already resolved the exceptional circumstances question by referring the matter to the court. The court is only seized with the merits question. There’s no logical short- 55 circuit between the two, because these are different questions with different content.

[143]Once that distinction is restored, the alleged contradiction dissipates. The second judgment’s position is coherent in that the President determines on a conclusive basis for the purpose of referral, that the matter passes the exceptional circumstances threshold. That the reconsidering court, standing in the shoes of the two judges, only asks whether there are reasonable prospects or compelling reasons for granting leave or special leave to appeal. Finding none, it refuses leave or special leave. To refuse leave on the merits does not deny that the matter was exceptional enough to be looked at. It is to conclude that, having looked, the application must fail. As for the relationship between the court and the President, it must be remembered that the matter does not get to the court if the President does not refer it to court. The first judgment’s error is to assume that a finding of exceptional circumstances entails a successful appeal. It cannot be, because a matter may be genuinely exceptional, and yet the application may still lack reasonable prospects of success on its own facts. The existence of exceptional circumstances justifies the second consideration, but the merits determine the outcome of the second consideration. Therefore, the first judgment is mistaken in imputing a contradiction to the second judgment. It is conceded that the second judgment could have spelt this position out more fully earlier on, but a gap in exposition is not a contradiction in logic.

[144]The first judgment also mischaracterised the constitutional values and asymmetry argument raised in the second judgment. It states that Bidvest postulates two interpretations, which are both consistent with the Constitution. And that mere disagreement with the preferred one does not render it unconstitutional. It concluded that the notion of constitutional harm to non- referred litigants rests on the ‘flawed premise that the court is reviewing the 56 President’s decision’. This view misses the second judgment’s point. Its equality argument, grounded in s 9 of the Constitution and further expanded in the second judgment above, is directed at a structural asymmetry produced by the jurisdictional fact interpretation. On that interpretation, a litigant whose application is referred by the President receives independent judicial scrutiny of the threshold by the court. Whereas a litigant whose application is dismissed by the President in chambers receives none, because nothing is referred to the court. Two litigants, identically situated as to the gravity of their complaints, are treated differently depending solely on the President’s determination. That is the inequality. It has nothing to do with whether the court ‘reviews’ the President’s decision. In that respect, the first judgment engages a target the second judgment did not set up.

[145]The Godloza minority judgment, quoted in the second judgment, captured the same asymmetry from the opposite direction: the President cannot sensibly have the power to determine exceptional circumstances for the purpose of refusing reconsideration but not for the purpose of granting it. The first judgment’s conclusion that finality simply attaches where there are no exceptional circumstances, and a special exception applies where there are, merely restates the disputed conclusion rather than meeting the objection. It assumes that the President’s dismissal and the President’s referral occupy different doctrinal determinations. The second judgment point is that the same determination, whether exceptional circumstances exist, is being made in both, and only one of them is subjected to the court’s scrutiny on the first judgment’s model. This asymmetry is not explained in the first judgment.

[146]In as much as noting the views expressed on the constitutionality of Bidvest, the second judgment’s pitched contention, drawing on s 39(2) of the Constitution, is the comparative one. Where a provision admits of two readings, 57 the court must prefer a reading that advances an identifiable constitutional value, in this case, equality before the law, provided that reading is reasonably available. The exclusivity principle treats all litigants alike, whether referred or dismissed in chambers. The jurisdictional interpretation does not. This argument was left untouched by the first judgment, save to state that both readings are consistent with the Constitution. Consistency is the floor. Section 39(2) asks for the better - aligned reading above the floor.

[147]On the panel size question, the first judgment mischaracterises the use of Bristol Aeroplane Company judgment. The second judgment invoked the said judgment for a narrower and entirely orthodox proposition that the court of appeal is ‘one court’, whose authority does not vary with the number of judges sitting. A proposition which is also accepted by the first judgment. The second judgment’s conclusion about overruling is then built on s 13 of the SC Act, not on Bristol Aeroplane Company. By attributing the overruling proposition to the English authority, the first judgment mischaracterised the argument made in the second judgment. In fact, it ignored the statutory argument based on s 13(2)(a), which makes the majority judgment ‘the judgment of the court’ without qualification as to numbers, substantially unaddressed.

[148]The first judgment also failed to confront the contention made by the second judgment that the Luphondo decision may itself require constitutional scrutiny, because if s 13 confers no greater juridical authority on a larger bench, the issue remains unsettled. A ruling that immunises a five member bench’s reasoning from correction by a later three member bench, is with respect flawed, as this hierarchy is not created by the SC Act.

[149]The first judgment espouses certainty by its reliance on Motsoeneng and Bidvest. But in my view, the divergence started when these judgments departed 58 from Avnit, without overruling it. The fact that this led to a proliferation of dissenting judgments on s 17(2)(f), is itself evidence of unsoundness of those decisions. In fact, the first judgment never grapples with the possibility that fidelity to precedent points to Avnit rather than Bidvest. More importantly than certainty is correctness. To allow an incorrect judgment to stand does not serve the interests of justice.

[150]In conclusion, the second judgment’s interpretation of s 17(2)(f) is faithful to the text, it gives content to ‘whether of his or her own accord’ and to the President’s status as a judicial officer. Moreover, it is faithful to the first decision on the section in Avnit, it treats referred and chambers dismissed litigants alike and it is coherent once the threshold and the merits are kept distinct. The value of the second judgment shows that the correct destination has not been reached and the traveller’s route is still open. The correct approach and outcome

[151]In interpreting s 17(2)(f), the court to which the President has referred the matter for reconsideration must step into the shoes of the two judges who refused leave or special leave to appeal and consider whether their decision should stand. The court does not independently assess whether exceptional circumstances exist; that question has already been resolved by the President's decision to refer. The court's function is to determine whether, on the merits, leave or special leave to appeal ought to be granted.

[152]Approached on that basis, and taking the merits as set out in the first judgment, I am satisfied that the two judges who refused leave to appeal did not err. The high court's misstatement of the onus under s 10(4) was per incuriam and did not create a binding precedent. The high court correctly applied the Plascon- Evans principle to Abu Dhabi Bank's version, which was neither bald nor 59 untenable. The absence of a banking licence or FSB authorisation does not, without more, establish an absence of bona fide intention to use the marks, particularly where an international banking institution of the respondent's stature explains a sequential, commercially rational approach to market entry.

[153]Accordingly, there is no basis to vary the two judges' order refusing leave to appeal. For those reasons, and not for the reasons relating to exceptional circumstances as expressed in the first judgment, I would also dismiss the application. However, consistent with the approach adopted in Avnit and S v Ramabanta, 77 and in line with the principle of finality, the proper order is one of dismissal, not striking from the roll.

[154]A dismissal order confirms the finality of the two judges' refusal and does not leave the prior order in procedural limbo. Striking the application from the roll would be inconsistent with the purpose of s 17(2)(f), which is to provide a final safety valve, not an interlocutory pause. Finality in litigation, while not absolute, demands that when reconsideration is sought and denied, the denial be decisive. Post Scriptum

[155]I strongly disagree with the assertions made in paragraphs 44 to 48 of the first judgment. The question of the repository of power under s 17(2)(f) was fully canvassed at the hearing. Counsel for FirstRand, on being made aware of the decision in Motsoeneng and Bidvest, accepted the position as pointed out to him. At the same time, I pointed out to counsel that there are different views on the subject matter and requested the parties to deal with s17(2)(f) despite the concession made. Procedural fairness in dealing with the legal question was applied. Most importantly, the issue did not require new facts or evidence. It was 77 Minister of Police and Another v Ramabanta [2025] ZASCA 95; 2026 (1) SA 100 (SCA). 60 apparent to everyone that some of the panel members accepted counsel's concession.

[156]I did not go on a frolic of my own as asserted in the first judgment. The issue raised with counsel arises from a question of law. My view is that the concession was made in error. It is trite that a court is not bound by a party's wrongful concession, especially on a legal question. I did not create new factual issues or make gratuitous findings, whereby the issues were not before the Court, as cautioned by this Court in National Director of Public Prosecutions v Zuma. 78

[157]The Constitutional Court has uniformly held, as stated in Seebed CC t/a Siyabonga Convenience Centre v Engen Petroleum Ltd: 79 ‘In Matatiele, this Court held that “it is trite that this Court is not bound by a legal concession if it finds the concession to be legally incorrect”.’ It further clarified that while legal concessions are generally accepted, they may be rejected if found to have been improperly made. That is what informed my decision to pen the dissenting judgment. Order

[158]I would make the following order: The application is dismissed with costs, including the costs of two counsel. ________________________ YT MBATHA ACTING DEPUTY PRESIDENT 78 National Director of Public Prosecutions v Zuma 2009 (2) SCA 277 (SCA); 2009(1) SACR 361 (SCA);2009 (4) BCLR 393 (SCA); [2009] 2 All SA 243. 79 Seebed CC t/a Siyabonga Convenience Centre v Engen Petroleum Ltd [2022] ZACC 28; 2023 (12) BCLR 1535 (CC) para 49. 61 Appearances: For applicants: P Ginsburg SC (with G Marriott and T Mpulo- Merafe) Instructed by: Adams & Adams Attorneys, Pretoria Honey Inc., Bloemfontein For respondent: R Michau SC (with T K Mamabolo) Instructed by: Kisch IP, Johannesburg Pieter Skein Attorneys, Bloemfontein.