Airports Company South Africa v Big Five Duty Free (Pty) Limited and Others (CCT257/17) [2018] ZACC 33; 2019 (2) BCLR 165 (CC); 2019 (5) SA 1 (CC) (27 September 2018)

CONSTITUTIONAL COURT OF SOUTH AFRICA

Case CCT 257/17

In the matter between:

AIRPORTS COMPANY SOUTH AFRICA Applicant

and

BIG FIVE DUTY FREE (PTY) LIMITED First Respondent

DFS FLEMINGO SA (PTY) LIMITED Second Respondent

TOURVEST HOLDINGS (PTY) LIMITED Third Respondent

Neutral citation: Airports Company South Africa v Big Five Duty Free (Pty) Limited and Others [2018] ZACC 33

Coram: Cachalia AJ, Dlodlo AJ, Froneman J, Goliath AJ, Jafta J, Khampepe J, Madlanga J, Petse AJ and Theron J

Judgments: Froneman J (majority): [1] to [64]

Jafta J (concurring): [65] to [83]

Cachalia AJ (dissenting): [84] to [111]

Heard on: 22 May 2018

Decided on: 27 September 2018

Summary: Interpretation of contracts — Judgments in rem — When can settlement agreements be made an order of court

Agreements between private parties — setting aside a court order — Section 217 of the Constitution

ORDER

Application for leave to appeal from the Supreme Court of Appeal (hearing an appeal from the High Court of South Africa, Gauteng Division, Pretoria):

1. Leave to appeal is granted.

2. The appeal is upheld.

3. The order made by the Supreme Court of Appeal is set aside.

4. The order in the High Court under case number 16829/15, dismissing the application of Big Five Duty Free (Pty) Limited with costs, including the costs of two counsel, is reinstated.

5. The first respondent is to pay the costs of the application in this Court and Supreme Court of Appeal, including the costs of two counsel.

JUDGMENT

FRONEMAN J (Dlodlo AJ, Goliath AJ, Khampepe J, Madlanga J, Petse AJ and Theron J concurring):

Introduction

[1] This judgment makes clear two legal propositions. The first is that a judgment in rem may not be set aside by only a settlement agreement between the litigating parties in an appeal against that judgment. For a judgment in rem to be set aside by a settlement agreement, the court hearing the appeal must give its sanction to the agreement being made an order of court on the basis that the setting aside is justified by the merits of the appeal. The second is that the court sanctioning the settlement agreement should give its reasons for doing so.

[2] A judgment in rem determines the objective status of a person or thing.[1] This Court has adopted an objective theory of invalidity regarding the exercise of public power.[2] A judgment that declares a tender invalid, because it is unlawful in contravention of section 217 of the Constitution,[3] is an objective pronouncement on the constitutional validity of an administrative act. That kind of judgment has a public

character that transcends the interests of only the litigating parties. It is a specific kind or example of a judgment in rem.

[3] Our law already recognises that judgments in rem are not subject to mere settlement on appeal. In the context of intellectual property law the Supreme Court of Appeal held, in Marine 3 Technologies, that the judgment appealed against was “one in rem in that it [affected] a public register, . . . notwithstanding the settlement of the matter, [Marine 3 was] constrained to proceed with the appeal”.[4] The implication is clear: unless the appeal court determines that the merits of the appeal accords with the outcome of the

settlement agreement it cannot make the settlement agreement an order of court. And a court must give reasons for its decisions,[5] especially when it gives its approval to a settlement agreement on appeal that sets aside a trial court’s judgment in rem.

[4] This is an application of what this Court held in Eke[6] to be one of the requirements for a settlement agreement to be made an order of court, namely that “its terms must accord with both the Constitution and the law [and] must not be at odds with public policy”.[7]

[5] The dispute in this case revolves around the meaning and effect of a settlement agreement and the import of sanctioning it as an order of court.

Factual background

[6] After a competitive bidding process, the applicant, Airports Company South Africa (ACSA) awarded a contract to the first respondent (Big Five) to operate duty free shops at its international airports for ten years. The second and third respondents (Flemingo and Tourvest) were unsuccessful bidders for the same contract. Flemingo took the award on review. Both ACSA and Big Five opposed the application for review. Phatudi J upheld the review and set aside the tender (Phatudi J order).

[7] Big Five appealed to a Full Court of the High Court of South Africa, Gauteng Division, Pretoria (High Court). ACSA abided the outcome of the appeal. After the appeal had been heard but before judgment, Big Five and Flemingo settled the case. In terms of the settlement, Flemingo abandoned the Phatudi J order and withdrew the review. The Full Court made the settlement an order of court on 20 June 2014.[8] No reasoned judgment accompanied the Full Court order. No further leave was sought to appeal the order.

[8] Seven months later ACSA announced that it was not bound by the Full Court order, that it remained bound by the Phatudi J order setting the contract aside, and that it intended to embark on a bidding process to award the contract afresh.

[9] Big Five launched an application in the High Court for orders declaring that ACSA remains bound by its original award of the contract to Big Five because the Phatudi J order had been overturned by the Full Court order. The High Court, per Hughes J, dismissed the application.[9] Big Five appealed to the Supreme Court of Appeal. It upheld the appeal and declared that ACSA remained bound by its

original award of the contract to Big Five.[10] ACSA seeks leave to appeal to this Court against the Supreme Court of Appeal’s order.

Jurisdiction and leave to appeal

[10] ACSA is an organ of state whose procurement processes involve the exercise of public power, which must in terms of the Constitution

be fair, equitable, transparent, competitive and cost-effective.[11] The Phatudi J order declared the tender award constitutionally invalid.[12] Whether that kind of order has the effect of an order in rem, and to what extent private settlement agreements seeking to set aside orders of that kind may be made orders of court also raise

important constitutional and other legal issues over which this Court has final jurisdiction.[13]

[11] Resolution of these constitutional legal issues will have an impact beyond the present litigation between the parties. Furthermore, the applicant also has reasonable prospects of success on the immediate issue between the parties. It is thus in the interests of justice that leave to appeal must be granted.

Discussion and analysis of the issues

[12] ACSA contends that the settlement agreement does not have the effect of setting aside the Phatudi J order. Big Five says that

it has precisely that effect.

[13] Although all parties now accept that: (1) as laid down in Eke, a settlement agreement between litigating parties can only be made an order of court if it conforms to the Constitution and the law; and (2) private parties cannot, by agreement, set aside a judgment in rem. These legal principles were not expressly articulated in either the Full Court or Supreme Court of Appeal orders. It

is also by now common cause that the Phatudi J order was indeed an order in rem.

[14] The problem then lies in the application of these principles to the settlement agreement and its incorporation as an order of court by the Full Court and the Supreme Court of Appeal. A number of possibilities exist, depending on the interpretation of the

agreement, as well as the interpretation of the courts’ approval of the settlement agreement as an order of court:

(a) One is that the agreement does not have the effect of setting aside the Phatudi J order and that the sanction of this private

agreement does not infringe the Constitution or law in any way. The Phatudi J order remains legally valid and the only effect of the settlement agreement is that the dispute between Big Five and Flemingo has been settled between them as private parties. ACSA’s appeal must succeed if this interpretation is accepted.

(b) Another is that the agreement does have the effect of setting aside the Phatudi J order. Big Five says once that is found it must succeed and the appeal must be dismissed. The issue of whether the Full Court’s sanction of the agreement was a proper one involving consideration and sanctioning the merits of the appeal, namely the setting aside of the Phatudi order, must then be accepted without further ado, as there was no further appeal lodged against the sanctioning of the settlement agreement as an order of court.

(c) If, however, one gets past the procedural obstacle of a failure to lodge an appeal against the Full Court order, one must then

determine whether the Full Court’s and the Supreme Court of Appeal’s sanction of the settlement agreement also involved a determination of the merits of the appeal, namely that the Phatudi J order should be set aside. If the Full Court order amounted to that, then Big Five succeeds. If it is clear that it does not amount to a sanction on the merits, then ACSA succeeds.

(d) But what if the interpretation of the agreement and the intentions of the Full Court are unclear? Are we bound to a definitive interpretation based on the limited information available to us, or is there another remedy available?

Interpretation or remittal?

[15] It is possible to approach the issues as merely ones of interpretation, first, of the settlement agreement and, second, of the court sanction of the agreement as an order of court. The purely interpretative approach may, however, be somewhat problematic.

[16] The Full Court only made an order incorporating the settlement agreement as the order of court, without any reasoned judgment. From

that material it is difficult, if not impossible, to determine what interpretation it gave to the meaning and import of the settlement

agreement, whether it considered its interpretation of the agreement to have any impact on the in rem Phatudi J order and, if so, whether it was satisfied that the appeal should have been upheld and the Phatudi J order set aside.

[17] The Supreme Court of Appeal interpreted the settlement agreement as intended to set aside the Phatudi J order. That Court also found that the Full Court intended the same:

“What was the purpose of the withdrawal and abandonment, coupled with the agreement of settlement, if not to set aside the

order of Phatudi J? It could be none other than to agree that the award to Big Five by ACSA was to stand. There is no other purpose that the parties could have intended to achieve.

The argument that the full court did not, independently of the parties, intend to set aside the Phatudi J order cannot be accepted.

The court order was made to give effect to the agreement between Big Five and Flemingo. Any distinction between the

parties' intention and that of the full court is thus obviously false. If one asks what the parties intended to achieve in

context – the facts known to them, the reasons for the agreement, the clear statement of Flemingo that it withdrew the review application in its entirety and abandoned any right in the order – the only answer could be that the parties did not intend the Phatudi J order to stand. It was a necessary implication of what they expressly stated. There is no other sensible construction of the agreement.”[14]

[18] The difficulty, even with the Supreme Court of Appeal’s fuller, reasoned, judgment, is that in the end the Phatudi J order

was set aside without any reasoned consideration of whether it was correct on the merits. On the face of it this approach is in conflict with the principle that a judgment in rem may not be set aside by a settlement agreement between the litigating parties in an appeal against that judgment, without the court giving its sanction to the agreement on the basis that the setting aside is justified by the merits of the appeal.[15]

[19] One way out of this difficulty would be to uphold the appeal on the grounds that there was insufficient information on the background context to the settlement agreement, as well as the Full Court order sanctioning it, for the Supreme Court of Appeal to come to any definitive conclusion on the interpretation of the agreement and particularly the correctness of the setting aside of the Phatudi

J order by the Full Court. A just and equitable remedy may then be to remit the matter to the Full Court to provide, in a reasoned judgment, its interpretation of the settlement agreement, whether it impacted on the in rem order made by Phatudi J and, if so, whether it was satisfied that the appeal should have been upheld and the Phatudi J order set aside. This would not involve any rehearing of the appeal, because it was already fully argued before the Full Court. Nor would it amount to an appeal against the Full Court order, because it would not pronounce upon the correctness of the order, but only seek reasons for it. This would underline the need for full reasons by an appellate court when making settlement agreements orders of court, especially where they might impact on in rem orders that had been granted in the court from which the appeal comes.

[20] Would remittal be a just and equitable way to resolve the matter?

[21] This Court has on a number of occasions ordered that matters be remitted to a lower court. This is because it was just and

equitable in the circumstances to do so. Mdeyide,[16] Merafong[17] and Qhinga[18] are examples of cases in which remittal was ordered.

[22] In Mdeyide, a case involving the issue whether the plaintiff had the necessary capacity to litigate, this Court held that:

“In light of the above, regrettably, it is necessary to set aside the order of the High Court in its entirety and remit the matter. It is open to an appellate court, in appropriate circumstances, to remit a matter for evidence to be obtained on matters that have been left obscure on the record.”[19]

[23] Florence[20] and the majority judgment in Everfresh[21] are instances where remittal was refused. In Florence, Van der Westhuizen J came to a different conclusion on the outcome than he did in Mdeyide:

“It is tempting to remit the matter to the Land Claims Court to consider further evidence on the current value of the property as well as other relevant factors to reach a decision in light of the guidance provided by this judgment. It is open to an appellate court, in appropriate circumstances, to remit a matter for evidence to be obtained in the hope that, in doing so, the issues will be fully ventilated. This Court, however, will be disinclined to remit, when doing so would be speculative or would involve wasted costs and energy on further legal contests.”[22]

[24] In Everfresh Moseneke DCJ, writing for the majority, held that:

“It is not good enough to ask the court a quo whether the common law needs to be developed. That hypothetical question

arises in every single case involving the common law. The [minority] judgment . . . eschews the merits and requires the High

Court to do that preliminary enquiry. I thus hold that it is not in the interests of justice to remit unless there is a reasonable

prospect that the court to which the matter is remitted is likely to hold that the common law needs to be adapted. Otherwise the remittal may be said to be speculative.” [23]

[25] From these cases it is possible to draw a few conclusions. Firstly, this Court will remit a matter to the lower courts if it

involves a question that would taint the legitimacy of the decision made - where for example one party may possibly not have legal capacity to litigate. Secondly, it will remit a matter to the lower courts if it determines that a legal question that necessitates pronouncement upon was ignored by the lower courts with the effect that this Court does not have the necessary information before it in order to answer the question properly. Thirdly, matters will only be remitted if it makes practical sense to do so. It will not remit matters in a speculative way without ascertaining that there are reasonable prospects of the lower court reaching a different decision to its original one once it has considered the new information. Moreover, the additional costs involved in prolonged litigation must also be considered.

[26] This is a matter where the second factor – in this case the lack of a reasoned definitive determination on the merits of sanctioning the setting aside of the Phatudi J order by the Full Court and the Supreme Court of Appeal – comes into play. To be weighed up against it are a number of practical considerations, including, the important fact that ACSA failed to avail itself of the opportunity to appeal against the Full Court order; that a considerable period of time has passed; that it is uncertain whether all the members of the Full Court are still available and will be able to reconstruct the reasons for their original order; that further remittal will incur further costs; and that courts are often called upon to interpret agreements and orders on the only available information placed on record before them. It is not without some hesitation and regret that I conclude that remittal is not the appropriate course here. What it underscores though, is the necessity for a reasoned judgment when making a settlement agreement an order of court on appeal where the settlement agreement impacts upon a court order in rem already made.

[27] I have had the pleasure and benefit of reading the concurrence of my brother Jafta J (second judgment). It appears to me that

there is little difference between us on the proper interpretation of the agreement.[24] The second judgment then goes on to say that because—

“the award of the tender was in breach of section 217 of the Constitution, [the agreement] could not be implemented. The reversal of Phatudi J’s order if it were to be done competently, could not cure the inconsistency and the invalidity. The invalidity of the tender stemmed from its inconsistency with the Constitution.”[25]

This inconsistency “rendered the order of the Full Court invalid. Consequently, the order of the Full Court did not overturn the order of Phatudi J and as a result Hughes J was right to dismiss Big Five’s application”.[26] This seems to assume that the merits of the Phatudi J order could not have been assessed by the Full Court in determining

whether to make the agreement an order of court. For the reasons set out earlier, I am of the view that it was open for the

Full Court to make that assessment if the agreement had the effect of seeking to set aside the Phatudi J order. But

logically I cannot see how that issue - the Full Court’s assessment on the merits of setting aside the Phatudi J order - can be settled before determining whether the agreement purports to set aside the Phatudi J order.

[28] The reality is that courts are often forced to interpret documents in less than ideal circumstances. This is another one of those occasions. I turn now to that task.

The proper interpretation of the settlement agreement

[29] There is no dispute about the principles of interpretation. The correct approach to the interpretation of documents was summarised by the Supreme Court of Appeal in Endumeni Municipality:

“Interpretation is the process of attributing meaning to the words used in a document, be it legislation, some other statutory

instrument, or contract, having regard to the context provided by reading the particular provision or provisions in the light of the document as a whole and the circumstances attendant upon its coming into existence. Whatever the nature of the document,

consideration must be given to the language used in the light of the ordinary rules of grammar and syntax; the context in which the provision appears; the apparent purpose to which it is directed and the material known to those responsible for its production. Where more than one meaning is possible each possibility must be weighed in the light of all these factors. The process is objective, not subjective. A sensible meaning is to be preferred to one that leads to insensible or unbusinesslike results or undermines the apparent purpose of the document. Judges must be alert to, and guard against, the temptation to substitute what they regard as reasonable, sensible or businesslike for the words actually used. To do so in regard to a statute or statutory instrument is to cross the divide between interpretation and legislation; in a contractual context it is to make a contract for the parties other than the one they in fact made. The ‘inevitable point of departure is the language of the provision itself’, read in context and having regard to the purpose of the provision and the background to the preparation and production of the document.”[27] (Footnotes omitted.)

[30] I begin then with the “inevitable starting point”, the language of the settlement agreement itself. Section 1 of

the agreement sets out the fundamental point that the agreement was in full and final settlement of all “claims which the parties have against each other” arising out of the present course of litigation. Sections 3 and 4 are most relevant to the issues at hand and I quote them in full below. Section 5 concerns representation and indemnification. Finally, section 6 contains several general clauses.

[31] Section 3 deals with the review application and clause 3.1 states:

“The Respondent [Flemingo] abandons the order of Phatudi J granted on 17 May 2012 in the review application. Pursuant hereto and on signature hereof by Respondent it will serve a notice of abandonment of this order in terms of rule of court 41(2). Without

limiting the generality of the terms ‘abandon’ the Respondent in addition waives and abandons all rights, title and

interest in and to this order.”

[32] Nowhere does the agreement say that it has the effect of setting aside the judgment of Phatudi J. Instead of explicit wording to that effect, Big Five relies on two clauses of the settlement agreement:

“3.3 The Respondent [Flemingo] hereby withdraws in its entirety the review application proceedings on the basis of and having the effect that these proceedings were never instituted and/or proceeded with and will serve on signature hereof a notice of withdrawal

reflecting therein that the matter is settled.

. . .

3.8 Upon signature of this agreement by the parties [Flemingo] acknowledges that ACSA is free to and can now implement the award of its tender (as referred to in this paragraph) to [Big Five] without limitation or restriction and without any challenge thereto

whatsoever by [Flemingo].”

[33] However, it is not clear that the wording of either of those clauses has the effect of setting aside the Phatudi J order.

[34] Clause 3.8 states that Flemingo acknowledges that ACSA is free to implement the tender award absent Flemingo’s opposition. At its highest, it is a promise made by Flemingo that it will not interfere with the implementation of the tender by ACSA. More realistically, its function is to spell out the consequences of the promises Flemingo has bound itself to – it will no longer concern itself with the contract, its validity, or its implementation. That reading gives meaning to the verb “acknowledges”.

[35] Ultimately, it does not matter whether clause 3.8 is properly read as Flemingo acknowledging the consequences of its promises or as Flemingo making a wholly new promise to step out of the fray. Neither its acknowledgment nor its promise to step aside is sufficient to set aside the Phatudi J order.

[36] The centre of the interpretative dispute is really clause 3.3. Unfortunately, it is difficult to know what that clause means. The problems with the clause are myriad.

[37] Flemingo claims to withdraw “in its entirety the review application proceedings . . . and will serve on

signature hereof a notice of withdrawal reflecting therein that the matter is settled”. However, withdrawal, in the ordinary legal sense of that word, can have nothing to do with the proceedings before Phatudi J. Those proceedings were complete and Flemingo could not withdraw from proceedings which had already concluded.[28] Clause 3.3 might then refer only to withdrawal from the proceedings before the Full Court. The problem with that interpretation is that the appeal proceedings appear to be covered by section 4 of the agreement.[29]

[38] A less literal interpretation also gives rise to difficulty. We could interpret “withdrawal” from proceedings to

include abandonment – the analogue of withdrawal where proceedings have already concluded. On that interpretation,

clause 3.3 captures Flemingo’s promise to withdraw from the proceedings before the Full Court and to abandon the Phatudi J order. But, as both parties agreed before this Court, the abandonment of a judgment in rem does not have the effect of setting it aside. That must be right. By abandoning a judgment, a party gives up any rights it had by virtue of that judgment – for example, the right to be paid by another party. A party cannot unilaterally affect the rights of others or change an objective fact – like whether or not an administrative act was unlawful – by giving up their own rights. Moreover, abandonment appears to be covered by clause 3.1 of the settlement agreement in which Flemingo explicitly states that it abandons the Phatudi J order.

[39] The meaning of “withdraw” in the clause is opaque. Unfortunately, the problems with clause 3.3 do not end there.

[40] Flemingo purported to withdraw “on the basis of and having the effect that these proceedings were never instituted and/or proceeded with”. From an interpretative standpoint, that statement is highly problematic. The grammar of the sentence is poor but there are two discernible propositions in it:

(a) that the basis for the withdrawal was that the proceedings were never instituted and/or proceeded with; and

(b) that the withdrawal is to have the “effect that these proceedings were never instituted and/or proceeded with”.

The first proposition appears to be meaningless. Clause 3.3 can thus only mean that the withdrawal is to have the effect that these proceedings were never instituted and/or proceeded with.

[41] An immediate problem is that the phrase “these proceedings” is ambiguous. It could mean either “the proceedings before the Full Court”, or “the proceedings before both Phatudi J and those before the Full Court”. While the headings of sections 3 and 4 of the agreement support the latter reading, the consequences of that reading are surprising

at best.

[42] The stipulated consequence makes perfect sense where “these proceedings” refers only to the appeal. Flemingo’s

withdrawal from the proceedings before the Full Court, together with the fact that the matter was then settled by agreement rather than by a judgment, would have the effect that those proceedings were never instituted and/or proceeded with.

[43] On the alternative interpretation, the consequences would be entirely novel – so novel as to border on the absurd. Clause 3.3 would mean that Flemingo’s withdrawal (whatever that means) from the proceedings before Phatudi J had the effect that

those proceedings were never instituted and/or proceeded with. But not even successful appeals have that effect. Where the order of a court is set aside by a higher court, the earlier judgment does not vanish in a puff of smoke.

[44] What then are we to make of clause 3.3? Looking at the agreement as a whole, it seems that its drafters were trying to indicate that Flemingo was doing everything within its power to withdraw from the fray – that Flemingo was to take all possible actions to remove itself and that it acknowledged its role in the tender process had definitively come to an end. Furthermore, the parties were trying to do what they could to unwind the process that had begun with the review application. But it was never in their gift to set aside the Phatudi J order.[30] That could not have been the effect of the settlement agreement without court approval, which had to be done on the basis of an independent assessment of the correctness of the Phatudi J order.

[45] For Big Five to succeed on the interpretative point, “withdraw” must be read to have the effect of setting aside a judgment in rem. In the absence of a clear indication from the Full Court that their order was intended to give “withdraw” that meaning, that is not possible. At least not without resorting to the authority famously referred to by Lord Atkin in Liversidge.[31] This is not Wonderland.[32]

[46] Notwithstanding the heading of section 3, “Review Application”, the best way to make sense of “withdrawal”

and its effects in clause 3.3 is in relation to the appeal proceedings. Paying heed to that heading, it is also necessary

to read “withdrawal” to mean “abandonment” where it makes sense to do so.

[47] It therefore seems that the best interpretation of clause 3.3 is as follows:

(a) Flemingo withdraws its opposition to the appeal and abandons the judgment of Phathudi J in the review application.

(b) Where possible, the terms of this agreement will have the effect that the relevant proceedings were never instituted and/or proceeded with.

(c) Where possible, Flemingo’s withdrawal will have the effect that the relevant proceedings were never instituted and/or proceeded with.[33]

(d) Flemingo promises to serve a notice of withdrawal reflecting that the matter is settled when the agreement is signed.

[48] While that interpretation makes the settlement agreement somewhat repetitious, it is consistent with the parties’ belt and braces approach.

[49] Most importantly in this matter, that interpretation clearly leaves in place the Phatudi J order. On its proper construction, the settlement agreement does not purport to set aside a judgment in rem. That is sufficient to decide the matter and it is not necessary to delve further into the parties’ alternative arguments.

[50] For the sake of completeness, it remains to deal directly with the most plausible arguments made in favour of Big Five’s interpretation.

The reasoning of the Supreme Court of Appeal

[51] The Supreme Court of Appeal reasoned as follows:

“[T]he agreement must be construed in the light of the circumstances attendant upon it – the factual matrix

or context. This is now settled law and I do not propose to repeat the recent authorities that state that an agreement must be construed in context. The context here was that Flemingo and Big Five, amongst others, had made bids for the lease of the duty free shops at three international airports on the terms set out in ACSA's invitation to tender. ACSA awarded the tender to Big Five. One of the unsuccessful bidders, Flemingo, took the award on review. It argued that the award was unlawful. Phatudi J found that it was unlawful for reasons that may or may not be good. Big Five appealed against the order with the leave of this Court. Before the appeal was heard, Flemingo and Big Five agreed that the appeal would not be prosecuted to finality but that Flemingo would abandon the order and withdraw the review proceedings – as if they had never happened.

What was the purpose of the withdrawal and abandonment, coupled with the agreement of settlement, if not to set aside the order of Phatudi J? It could be none other than to agree that the award to Big Five by ACSA was to stand. There is no other purpose that the parties could have intended to achieve.

The argument that the full court did not, independently of the parties, intend to set aside the Phatudi J order cannot be accepted.

The court order was made to give effect to the agreement between Big Five and Flemingo. Any distinction between the

parties' intention and that of the full court is thus obviously false. If one asks what the parties intended to achieve in

context – the facts known to them, the reasons for the agreement, the clear statement of Flemingo that it withdrew the review application in its entirety and abandoned any right in the order – the only answer could be that the parties did not intend the Phatudi J order to stand. It was a necessary implication of what they expressly stated. There is no other sensible construction of the agreement.”[34]

[52] That reasoning places inordinate weight on the intention of the parties and pays little heed to the unrealistic meaning that such

an interpretation would attribute to the actual words of the agreement, particularly clause 3.3.

[53] Of course, the intention of the parties is relevant to the interpretation of the settlement agreement, but ultimately the meaning of the agreement is borne out by the words used. And words do not mean whatever the parties later say they mean: “[w]hile

the intention of the speaker is a vital component of the interpretive enquiry, the basic unit of meaning is the sentence employed”.[35]

[54] One reason to be wary of placing weight upon the intention of Big Five and Flemingo to set aside the Phatudi J order is that we simply do not know their true intention. The agreement itself does not explicitly state that intention. Big Five and Flemingo do not explicitly say that they intend to set aside the judgment below. And there are many reasons to enter a settlement

agreement. Whenever a settlement is entered into, the resolution of the dispute is brought back into the hands of private

parties. No reasons need to be given and whatever factors induced the settlement will usually remain unknown to the public and the court. In the face of uncertainty, it is a mistake to place the entire weight of the interpretative exercise on what the court thinks the parties’ intention might have been.[36]

[55] Even if the parties actually intended to set aside the Phatudi J order, that does not mean that they gave effect to that intention

by entering into the settlement agreement.

[56] We know that Big Five was labouring under a mistaken view about the consequences of abandonment. Previously it argued that, even in the absence of the settlement agreement having been made an order of the Full Court, the review was abandoned by Flemingo and therefore it had no legal effect. In argument before us it was conceded that abandonment alone could not have nullified the Phatudi J order.

[57] It might well be that Big Five and Flemingo were wrongly relying on the effect of abandonment to set aside the earlier judgment. If that is right, then their failure to craft an agreement that actually set aside the Phatudi J order is easily explained – they misunderstood the consequences of withdrawal and abandonment on judgments in rem.

[58] Regardless of their true intention, the words used in the settlement agreement do not have that effect.

The role of Eke in the interpretative arguments

[59] Both ACSA and Big Five attempted to deploy this Court’s reasoning in Eke to their advantage. The relevant part of Eke reads:

“This in no way means that anything agreed to by the parties should be accepted by a court and made an order of court. The order can only be one that is competent and proper. A court must thus not be mechanical in its adoption of the terms of a settlement agreement. For an order to be competent and proper, it must, in the first place ‘relate directly or indirectly to an issue or lis between the parties’. Parties contracting outside of the context of litigation may not approach a court and ask that their agreement be made an order of court.

Secondly, ‘the agreement must not be objectionable, that is, its terms must be capable, both from a legal and a practical point of view, of being included in a court order’. That means, its terms must accord with both the Constitution and the law. Also, they must not be at odds with public policy. Thirdly, the agreement must ‘hold some practical and legitimate advantage.’”[37] (Footnotes omitted.)

[60] Big Five reasoned that if the Full Court had followed the requirements set out in that case, then the court must have been satisfied that the merits of the case warranted setting aside the Phatudi J order. ACSA reasoned that if the Full Court had followed those requirements, they would not, without comment, have made an agreement setting aside the Phatudi J order an order of court.

[61] Both parties then relied on the same second premise: we must presume that the Full Court adhered to the requirements set out in Eke. In that way, Big Five deployed Eke to make more palatable its claim that the Full Court order set aside the Phatudi J order and ACSA deployed Eke to bolster their interpretation of the order.

[62] Those arguments suffer the same defect. The first if/then premise each side prefers predetermines the conclusion at which they arrive. The conclusion of the argument is effectively assumed by the first premise. That is not valid legal reasoning. For that reason, I have eschewed reliance on Eke in determining this case.

Costs

[63] While ACSA is a public functionary, the heart of the matter is an interpretative dispute and Big Five is acting in its commercial

interests. Flemingo and Tourvest took

no active part in the matter before this Court. The Biowatch principle does not apply. Costs are to follow the result.

Order

[64] The following order is made:

JAFTA J:

[65] I have had the advantage of reading the judgment of my colleague Froneman J (first judgment). Having concluded that the Phatudi J order was a judgment in rem which was not subject to, without more a settlement on appeal, the first judgment holds that the dispute revolves around the meaning and effect of a settlement agreement.

[66] However, as I see it, the antecedent question must be whether it was competent for the Full Court to make the settlement agreement

an order of court in the present circumstances. For if it was not, whatever the terms of that order may mean would be irrelevant to the question whether the tender that was competently set aside by Phatudi J was revived. It was only if that tender was properly revived that the relief sought by Big Five could succeed. In other words, the order of the Full Court will have the legal consequences of reviving the tender only if it was itself competently granted.

[67] I am grateful to the first judgment for its exposition of the facts which I will not repeat here except to the extent necessary for

the conclusion I reach. A good place to start the enquiry into whether in the present circumstances it was competent for the Full Court to make the settlement agreement an order of court is the Constitution.

Section 172(1) of the Constitution

[68] Section 172(1) of the Constitution affirms the role played by courts in ensuring that public power is properly exercised.[38] In addition, this provision enhances the capacity of courts to uphold and enforce the Constitution. In mandatory terms, it obliges courts to declare invalid laws or conduct that is not in line with the Constitution. It provides:

“When deciding a constitutional matter within its powers, a court¾

a. must declare that any law or conduct that is inconsistent with the Constitution is invalid to the extent of its inconsistency.”

[69] Therefore, when it was established to the satisfaction of the High Court that ACSA had awarded the tender to Big Five in contravention of section 217 of the Constitution, Phatudi J had no option but to declare the award invalid and set it aside.[39] This much is clear from the text of section 172(1)(a). What is also evident from this section is the fact that the power to declare conduct that is inconsistent with the Constitution invalid is reserved for courts. Even so, it is only courts with the necessary competence which may exercise the power to nullify conduct that is inconsistent with the Constitution.[40]

[70] The question that arises for consideration here is whether, once a competent court has exercised judicial power and declared conduct that is inconsistent with the Constitution invalid, litigants may seek to reverse this declaration by agreement. As the first

judgment rightly observes, the Phatudi J order transcended interests of parties in that litigation and as a matter of law it could not be overturned by parties’ settlement agreement on appeal, without overruling the legal conclusions on which that order rested.[41]

[71] What happened here is this: Following the Phatudi J order, Big Five appealed to the Full Court which heard full argument on 4 June

2014 but reserved judgment. But before the hearing, settlement negotiations between Big Five and Flemingo were at an advanced

stage. The settlement was concluded on 13 June 2014 and a copy of the signed agreement was sent to ACSA on 17 June 2014.

Flemingo wanted the settlement agreement to incorporate claims against it by ACSA. However, this did not materialise because

ACSA was not willing to waive its claims against Flemingo without board approval.

[72] On 19 June 2014, Flemingo delivered at the High Court several notices, some of which were drawn in terms of rule 41(2) of the Uniform Rules of Court.[42] In terms of the rule 41(2) notices, Flemingo abandoned judgments of Tolmay J and Phatudi J. Tolmay J had granted an interim interdict against Big Five and in favour of Flemingo which later succeeded in the review before Phatudi J. In addition, Flemingo filed a notice which purportedly withdrew the review application. This was done even though the matter had reached an appeal stage and the appeal had already been heard.

[73] It is not clear from the papers why Flemingo had to file the notices on matters that were addressed in the settlement agreement.

For example, the agreement records that Flemingo withdraws the review application and the appeal. In it Flemingo also abandons the orders of Tolmay J and Phatudi J. But more significantly, clause one of the agreement reveals the intention of the parties to it. It stipulates:

“This agreement is in full and final settlement and compromise of all or any claims (including for losses, damages and costs) which the parties have against each other arising out of the institution of the interim interdict application, review application by [Flemingo] against [Big Five], ACSA and the Tender Board of ACSA and the appeals to date.”

[74] This is contrary to what was held by the Supreme Court of Appeal in its judgment. That Court concluded that the purpose of the agreement was to set aside the Phatudi J order. It stated:

The argument that the [Full Court] did not, independently of the parties, intend to set aside the Phatudi J order cannot be

accepted. The court order was made to give effect to the agreement between Big Five and Flemingo. Any distinction between the parties’ intention and that of the Full Court is thus obviously false. If one asks what the parties intended to achieve in context – the facts known to them, the reasons for the agreement, the clear statement of Flemingo that it withdrew the review application in its entirety and abandoned any right in the order – the only answer could be that the parties did not intend the Phatudi J order to stand. It was a necessary implication of what they expressly stated. There is no other sensible construction of the agreement.”[43]

[75] In so concluding the Supreme Court of Appeal overlooked a number of things. First, that Court overlooked the facts. In

the founding affidavit filed in the High Court, Big Five had averred:

“Pursuant to the settlement agreement, Flemingo abandoned the interdict and review judgment in their entirety by delivering notices in terms of rule 41(2) on 19 June 2014. The Full Court, mero motu made the settlement agreement an order of court on 20 June 2014. That order, incorporating the settlement agreement is annexed as FA 17.”

[76] It is evident that the parties never asked the Full Court to make their settlement agreement an order of court. Instead, the

Full Court acted of its own accord (mero motu). If the parties had intended to set aside the Phatudi J order, they would have expressly said so in their settlement agreement and they would have asked the Full Court to make that agreement an order of court. They did none of this.

[77] Second, the Supreme Court of Appeal disregarded the question whether it was competent for the Full Court in present circumstances

to make the settlement agreement an order of court. A perusal of the settlement agreement shows that it was not.

[78] Clause 3.8 of the agreement stipulates:

“Upon signature of this agreement by the parties [Flemingo] acknowledges that ACSA is free to and can now implement the award of its tender (as referred to in this paragraph) to [Big Five] without limitation or restriction and without any challenges thereto

whatsoever by [Flemingo].”

[79] This clause implies that as from 13 June 2014 at the time the agreement was signed by Flemingo, ACSA was free to implement the tender that had been set aside. In law this was not competent. Litigants cannot overturn a court order by private treaty. What was suggested in clause 3.8 was inconsistent with the Constitution in a number of respects. As the award of the tender was in breach of section 217 of the Constitution, it could not be implemented. The reversal of the Phatudi J order if it were to be done competently, could not cure the inconsistency and the invalidity. The invalidity of the tender stemmed from its inconsistency with the Constitution.

[80] Section 2 of the Constitution affirms the supremacy of the Constitution and declares that law or conduct inconsistent with it is

invalid. In terms of the principle of objective constitutional invalidity, the tender in question became invalid the moment it was awarded in breach of the Constitution. Phatudi J merely confirmed an existing objective fact.[44]

[81] Moreover, the present settlement agreement did not meet the requirements for making such agreements an order of court. The

Full Court overlooked this and readily made the agreement an order of court. Before the parties’ agreement is made an order of court, the court contemplating to do so must be satisfied that the order it is about to make is “competent and proper” and that it is consistent with the Constitution and the law. This was affirmed in Eke where it was stated:

“This in no way means that anything agreed to by the parties should be accepted by a court and made an order of court. . . . The order can only be one that is competent and proper. A court must thus not be mechanical in its adoption of the terms of a settlement agreement.

That means its terms must accord with both the Constitution and the law.”[45]

[82] As stated, the terms of the settlement agreement we are concerned with here are inconsistent with the Constitution and this rendered the order of the Full Court invalid. Consequently, the order of the Full Court did not overturn the Phatudi J order and as a result Hughes J was right to dismiss Big Five’s application.

[83] For all these reasons I support the order made in the first judgment.

CACHALIA AJ:

[84] The first judgment would overrule the Supreme Court of Appeal’s judgment on the ground that it erred in interpreting the settlement agreement as having the effect of setting aside the Phatudi J order. The second judgment would also uphold the appeal, but on the ground that the Supreme Court of Appeal overlooked the fact that it was not “competent and proper” for the Full Court to have incorporated the agreement in its order. I regret that I am unable to agree with either of the two judgments.

In my view, the Supreme Court of Appeal correctly found that the purpose of the agreement was to set aside the Phatudi J order.

And, with one proviso, which I shall discuss in due course, it also correctly concluded that it was lawful for the Full Court to include an agreement having this effect in its order.

[85] Before considering the issues in this appeal, it is necessary to set out the facts fully to understand the relevant background more

clearly. I repeat some of the facts mentioned in the first two judgments so as to make this judgment easier to read.

[86] Following a competitive bidding process, ACSA awarded the contract to Big Five on 26 August 2009. The contract would endure for ten years. A month later, Flemingo, one of the two unsuccessful bidders, applied to interdict the implementation of the contract pending a review of ACSA’s decision in the High Court. The High Court granted the interim interdict on 11 December 2009. ACSA and Big Five opposed the review. On 17 May 2012, Phatudi J upheld the review and set aside the decision to award the contract to Big Five. He also dismissed an application for leave to appeal by Big Five on 2 October 2012. ACSA did not apply for leave to appeal at this stage.

[87] On 21 November 2012, the Supreme Court of Appeal granted Big Five leave to appeal to the Full Court. The appeal was set down

for 4 June 2014. ACSA was content to abide by whatever the outcome might be.

[88] By May 2014, settlement negotiations between Big Five and Flemingo had advanced rapidly. Flemingo was concerned to avoid any claims against it by ACSA and therefore wanted the latter to be party to the settlement agreement. The draft agreement was sent to ACSA’s Group Executive, who supported it. However, as the approval of ACSA’s Board could not be obtained before 4 June 2014, the appeal proceeded without its concurrence.

[89] The appeal was fully argued and judgment reserved. A week later, before judgment was delivered, Big Five and Flemingo concluded the agreement, without providing for a waiver of claims against Flemingo.

[90] In terms of clause 3.3 of the agreement, Flemingo withdrew “in its entirety the review application proceedings on the basis of and having the effect that these proceedings were never instituted”. And, under clause 3.8, Flemingo accepted that “ACSA is free to and can now implement the award of its tender . . . to [Big Five]”. Clause 6.4 stated that they consented to the agreement being made an order of court.

[91] The agreement was delivered to ACSA on 17 June 2014. On 19 June 2014, Flemingo filed notices in terms of rule 41(2) in which

it formally abandoned the interdict and review judgments. On 20 June 2014, ACSA indicated by email that it would respond formally after it had assessed the legal and financial impact of the agreement. The Full Court incorporated the agreement in its order of 20 June 2014.

[92] Following the Full Court order, Big Five repeatedly, but unsuccessfully, attempted to obtain clarity from ACSA on the implementation of the tender. On 14 January 2015, almost seven months later, ACSA announced its intention to initiate a fresh tender process. In response, Big Five launched the present application on 6 March 2015 to compel ACSA to implement the original award of the tender to Big Five. The application failed in the High Court, but that ruling was reversed on appeal to

the Supreme Court of Appeal. ACSA now appeals to this Court. It is against this background that the issues in dispute in this appeal must be decided. I first consider the status of the Full Court order.

Status of the Full Court order

[93] Once a settlement agreement is made an order of court, it has the same standing and qualities as any other court order.[46] Its effect is to finally determine the status of the rights and obligations between the parties and the issues covered by the dispute.[47] It is enforceable like “any other” court order.[48] This means that it may only be impugned through a “legally cognisable” process such as rescission or appeal.[49]

[94] ACSA abided by the order of the Full Court and made no attempt to rescind or appeal it. This, despite the fact that it was

aware of the order and, importantly, that Big Five believed that its consequence was to revive the tender. Yet, it now contends that it is not bound by it.

[95] ACSA’s argument proceeds as follows: As an organ of state, ACSA has a duty to uphold its constitutional obligations and the rule of law. Once Phatudi J had found the award of the tender to be unconstitutional and invalid, ACSA accepted that it could not proceed with the award of the tender to Big Five. It therefore elected to abide by the appeal, but not a private settlement agreement to which it was not party. This is because the order of constitutional invalidity was an order in rem, binding on everyone, and incapable of being abandoned or set aside by private agreement. So, properly understood, ACSA only abided by the outcome of a reasoned judgment on the merits and not the order of the Full Court that merely recorded the private settlement agreement between Big Five and Flemingo.

[96] But ACSA’s contentions do not bear scrutiny. First, the order of the Full Court was an order in rem binding on everyone, no less than the Phatudi J order. This is because it finally determined the status of the Phatudi J order and the underlying administrative action that gave rise to it. It did so for all concerned, including ACSA. ACSA could not ignore the order.[50] If it felt that it was not bound by the order, it had a duty to either appeal the order or have it rescinded.

[97] Second, the facts are against ACSA. It did not agree to be bound only by a reasoned judgment on appeal, as it would have us

believe. It agreed to abide by the decision of the Full Court without more. It was aware of and encouraged the settlement

negotiations between Big Five and Flemingo and did not demur after being informed that the agreement had been made an order of court. It strung Big Five along for seven months until the latter was compelled to approach the High Court to enforce its contractual rights. It cannot now be heard to contend otherwise.

[98] Third, Big Five contended that the Full Court had not merely recorded the settlement agreement in its court order but had made the

order after hearing full argument by the parties, having regard to the record and both judgments of the High Court. This necessarily implies that the Full Court considered the merits of the dispute.[51] It added that the Full Court “was thus satisfied that the outcome envisaged in the settlement agreement was correct and in the interests of justice”. Put differently, it contended that the Full Court must be taken to have made an order that it was properly and competently entitled to make, which accorded with both the Constitution and the law and was not inimical to public policy.[52]

[99] Yet, in its answering affidavit, ACSA adduced no facts to gainsay this assertion; it merely stated that this was a matter for legal argument. But it could not adopt this stance for the onus rested upon ACSA – not Big Five – to demonstrate that the order was not competently and properly made. ACSA therefore cannot impugn the order of the Full Court. It follows too that I am unable to agree with the second judgment’s holding in this regard.[53]

[100] I do not take issue with the obiter dictum in the first judgment that a court “should” give reasons when sanctioning a settlement agreement on appeal that sets aside a judgment in rem of a court of first instance.[54] However, it does not follow that because reasons are not given, an order incorporating a settlement agreement is not competent. If reasons are not given at the time the order is made, the parties may request reasons afterwards, especially if intended to impugn the order. ACSA could and should have adopted this course had it intended to challenge the order through some legally cognisable

process. It cannot take issue with the absence of a reasoned judgment in these proceedings.

Interpreting the Full Court’s order incorporating the settlement agreement

[101] This brings me to the second issue in this appeal: The proper interpretation of the order of the Full Court incorporating the settlement agreement. In this regard, the key dispute is whether that order had the effect of setting aside the Phatudi J order. If it did, as the Supreme Court of Appeal found it had, the appeal must be dismissed. If on the other hand it did not do so, which means that the order remains extant, as the first judgment finds, the appeal must fail. I proceed to interpret the settlement agreement.

[102] The first judgment cites Endumeni Municipality[55] as summarising the proper approach to interpreting a written agreement.[56] To that, I would add the later dictum of the Supreme Court of Appeal in Bothma-Batho Transport:[57]

“Whilst the starting point remains the words of the document, which are the only relevant medium through which the parties have expressed their contractual intentions, the process of interpretation does not stop at a perceived literal meaning of those words, but considers them in the light of all relevant and admissible context, including the circumstances in which the document came into being. The former distinction between permissible background and surrounding circumstances, never very clear, has fallen away. Interpretation is no longer a process that occurs in stages but is ‘essentially one unitary exercise’.”

(Footnote omitted.)

[103] The Supreme Court of Appeal unanimously held that the purpose of the settlement agreement was none other than to agree that the award to Big Five by ACSA was to stand. There was no other purpose that the parties could have intended to achieve, it said.[58]

[104] However, ACSA contends that textually the settlement agreement does not have that effect. It maintains that the Full Court

order does not expressly uphold the appeal, set aside the Phatudi J order, and settle the merits of the judgment or the validity of the tender award. It goes no further, says ACSA, than to acknowledge that it is free to implement the tender without obligating it to do so.[59]

[105] The first judgment endorses ACSA’s contention and takes issue with the Supreme Court of Appeal’s reasoning as having

placed “inordinate weight on the intention of the parties” instead of examining the meaning of the actual words used. Where the intention of the parties is not explicit from the language, the judgment continues, it is incorrect for a court to place the entire interpretative exercise on what the court imagines the intention of the parties might have been.[60]

[106] On this approach, and regardless of their true intentions, the reasoning continues, ACSA misunderstood the consequences of the true

import of the language used in clause 3.3 because neither the withdrawal nor the abandonment of the Phatudi J order had the

legal effect of setting it aside.[61]

[107] I do not think that the criticism of the Supreme Court of Appeal’s approach to the interpretation of the settlement agreement is warranted. As is clear from Endumeni Municipality and Bothma-Batho Transport above, the point of departure in an interpretive process is always the language of the agreement in what is ultimately a unitary exercise. But it is only the starting point in an exercise to establish the contractual intention of the parties. Equally important in this analysis is the context within which the language is used in the light of the document as a whole, the circumstances attendant upon its coming into existence, the apparent purpose to which it is directed and the material known towards those responsible for its production. Importantly, a sensible meaning is to be preferred to one that leads to “unbusinesslike” results or undermines the apparent purpose of the document.

[108] In my view, the error in the approach of the first judgment is to fixate on the words used in the relevant clauses without weighing

them in the light of the other factors. From this incorrect premise it concludes that the agreement went no further than to settle the dispute between Big Five and Flemingo while leaving the Phatudi J order in place.[62]

[109] But as I have said, this ignores the other factors. First, the dispute was primarily about the award of the tender to Big Five. That was what the Phatudi J order set aside and what the parties to the settlement agreement sought to undo. The agreement made clear in its preamble that its purpose was to settle the review in which the Phatudi J order had been made. To that end, Flemingo expressly abandoned the Phatudi J order and waived “all rights, title and interest in and to this order”. It also withdrew its application for review with the effect that “these proceedings were never instituted”.

Importantly, the parties acknowledged that the effect of their agreement was “that ACSA is free to and can now implement the award of its tender” to Big Five. This could only mean that the purpose of the agreement was to set aside the Phatudi J order so that the original award of the tender could be reinstated. Second, the interpretation preferred by the first judgment, which is aimed at preserving the in rem judgment of Phatudi J, rather than giving effect to the purpose of the document, is insensible and leads to an “unbusinesslike” result, which is not what the parties wanted to achieve by this agreement.

[110] To conclude: having regard to the factors enumerated, which take account of the language used in its proper context, the circumstances attendant upon which the agreement was concluded and made an order of court, the purpose to which it was directed and its commercial efficacy, it is quite clear that the parties intended that their agreement would effectively set aside the Phatudi J order.

It ill behoves ACSA, which is a stranger to the agreement by which it had agreed to abide, to now opportunistically contend that it means something other than what the parties themselves say it means.[63]

[111] I would accordingly dismiss the appeal with costs, including the costs of two counsel.

For the Applicant: G Marcus SC, B Morris, J Grant and A Armstrong instructed by Norton Rose Fulbright South Africa Inc.

For the First Respondent: W Trengove SC, S Budlender and A Msimang instructed by CRD Attorneys.

[1] Tshabalala v Johannesburg City Council 1962 (4) SA 367 (T) at 368H.

[2] Shoprite Checkers (Pty) Ltd v MEC for Economic Development, Eastern Cape [2015] ZACC 23; 2015 (6) SA 125 (CC); 2015 (9) BCLR 1052 (CC) at para 31; Ferreira v Levin N.O.; Vryenhoek v Powell N.O. [1995] ZACC 13; 1996 (1) SA 984 (CC); 1996 (1) BCLR 1 (CC) (Ferreira) at para 26.

[3] Section 217 of the Constitution states: “(1) When an organ of state in the national, provincial or local sphere of government, or any other institution identified in national legislation, contracts for goods or services, it must do so in accordance with a system which is fair, equitable, transparent, competitive and cost-effective. (2) Subsection (1) does not prevent the organs of state or institutions referred to in that subsection from implementing a

procurement policy providing for— (a) categories of preference in the allocation of contracts; and (b) the protection or advancement of persons, or categories of persons, disadvantaged by unfair discrimination. (3) National legislation must prescribe a framework within which the policy referred to in subsection (2) must be implemented.”

[3] Section 217 of the Constitution states:

“(1) When an organ of state in the national, provincial or local sphere of government, or any other institution identified in national legislation, contracts for goods or services, it must do so in accordance with a system which is fair, equitable, transparent, competitive and cost-effective.

(2) Subsection (1) does not prevent the organs of state or institutions referred to in that subsection from implementing a

procurement policy providing for—

(a) categories of preference in the allocation of contracts; and

(b) the protection or advancement of persons, or categories of persons, disadvantaged by unfair discrimination.

(3) National legislation must prescribe a framework within which the policy referred to in subsection (2) must be implemented.”

[4] Marine 3 Technologies Holdings (Pty) Ltd v Afrigroup Investments (Pty) Ltd [2014] ZASCA 208; 2015 (2) SA 387 (SCA) (Marine 3 Technologies) at para 6. The terms of the settlement agreement in Marine 3 Technologies acknowledged that “[t]his is a judgment in the public interest that cannot be abandoned by Afri-Group” at para 5. The agreement therefore merely recorded that Afri-Group saw no reason why the Supreme Court of Appeal should not find in favour of the appellant and that it would not oppose the appeal. The parties did not purport to determine the outcome of the appeal. The Supreme Court of Appeal reached a decision on the merits, taking into account the parties’ settled position. This approach was also applied by the Supreme Court of Appeal in The GAP Inc v Salt of the Earth Creations (Pty) Ltd [2012] ZASCA 68; 2012 (5) SA 259 (SCA) at para 2.

[5] Stuttafords Stores v Salt of the Earth Creations [2010] ZACC 14; 2011 (1) SA 267 (CC); 2010 (11) BCLR 1134 (CC) at paras 10-2.

[6] Eke v Parsons [2015] ZACC 30; 2016 (3) SA 37 (CC); 2015 (11) BCLR 1319 (CC).

[7] Id at para 26.

[8] Big Five Duty Free (Pty) Ltd v DFS Flemingo SA (Pty) Ltd, unreported order in the High Court under case no A887/2012 (Full Court order).

[9] Big Five Duty Free (Pty) Limited v Airports Company South Africa Limited [2016] ZAGPPHC 688 at para 57.1.

[10] Big Five Duty Free (Pty) Limited v Airports Company South Africa Limited [2017] ZASCA 110; [2017] 4 All SA 295 (SCA) (SCA judgment) at para 29.

[11] Section 217(1) of the Constitution.

[12] D F S Flemingo SA (Pty) Ltd v Airports Company South Africa Ltd [2012] ZGPPHC 66 at paras 21-2. That kind of order in itself raises a sufficient constitutional issue within this Court’s

jurisdiction: AllPay Consolidated Investment Holdings (Pty) Ltd v Chief Executive Officer, South African Social Security Agency [2013] ZACC 42; 2014 (1) SA 604 (CC); 2014 (1) BCLR 1 (CC) at para 4. See also Department of Transport v Tasima (Pty) Limited; Tasima (Pty) Limited v Road Traffic Management Corporation [2018] ZACC 21 at para 37; and Trencon Construction (Pty) Ltd v Industrial Development Corporation of South Africa Ltd [2015] ZACC 22; 2015 (5) SA 245 (CC); 2015 (10) BCLR 1199 (CC) at para 1.

[13] Sections 165(5), 167(3)(b)(i) and (ii) of the Constitution establishes this Court’s jurisdiction in this matter.

See further Eke above n 6 at para 1.

[14] SCA judgment above n 10 at paras 26-7.

[15] See [1] above.

[16] Road Accident Fund v Mdeyide (Minister of Transport intervening) [2007] ZACC 7; 2008 (1) SA 535 (CC); 2007 (7) BCLR 805 (CC) (Mdeyide) at para 43.

[17] Merafong City Local Municipality v Anglogold Ashanti Limited [2016] ZACC 35; 2017 (2) SA 211 (CC); 2017 (2) BCLR 182 (CC) (Merafong) at para 68.

[18] Qhinga v S [2011] ZACC 18; 2011 (2) SACR 378 (CC); 2011 (9) BCLR 980 (CC) at para 35.

[19] Mdeyide above n 16 at para 43.

[20] Florence v Government of the Republic of South Africa [2014] ZACC 22; 2014 (6) SA 456 (CC); 2014 (10) BCLR 1137 (CC).

[21] Everfresh Market Virginia (Pty) Ltd v Shoprite Checkers (Pty) Ltd [2011] ZACC 30; 2012 (1) SA 256 (CC); 2012 (3) BCLR 219 (CC) (Everfresh).

[22] Florence above n 20 at para 91.

[23] Everfresh above n 21 at para 79.

[24] [73]-[78] below.

[25] [79] below.

[26] [82] below.

[27] Natal Joint Municipal Pension Fund v Endumeni Municipality [2012] ZASCA 13; 2012 (4) SA 593 (SCA) (Endumeni Municipality) at para 18. Approach to interpretation cited with approval in the judgment of Cameron J in National Credit Regulator v Opperman [2012] ZACC 29; 2013 (2) SA 1 (CC); 2013 (2) BCLR 170 (CC) at fn 105 and see also see para 93-94. Later cited with approval and discussed in the majority judgment of Cameron J in AMCU v Chamber of Mines of South Africa [2017] ZACC 3; 2017 (3) SA 242 (CC); 2017 (6) BCLR 700 (CC) at fn 28.

[28] It is quite clear from the Uniform Rules of Court that withdrawal is open to the parties before a judgment is given, but not

subsequently: “41 Withdrawal, Settlement, Discontinuance, Postponement and Abandonment: (1)(a) A person instituting any proceedings may at any time before the matter has been set down and thereafter by consent of the parties or leave of the court withdraw such proceedings, in any of which events he shall deliver a notice of withdrawal and may embody in such notice a consent to pay costs; and the taxing master shall tax such costs on the request of the other party. . . . (2) Any party in whose favour any decision or judgment has been given, may abandon such decision or judgment either in whole or in part by delivering notice thereof and such judgment or decision abandoned in part shall have effect subject to such abandonment.

The provisions of subrule (1) relating to costs shall mutatis mutandis apply in the case of a notice delivered in terms of this

subrule. (3) If in any proceedings a settlement or an agreement to postpone or withdraw is reached, it shall be the duty of the attorney for the plaintiff or applicant immediately to inform the registrar accordingly. (4) Unless such proceedings have been withdrawn, any party to a settlement which has been reduced to writing and signed by the

parties or their legal representatives but which has not been carried out, may apply for judgment in terms thereof on at least

five days' notice to all interested parties.”

[28] It is quite clear from the Uniform Rules of Court that withdrawal is open to the parties before a judgment is given, but not

subsequently:

“41 Withdrawal, Settlement, Discontinuance, Postponement and Abandonment:

(1)(a) A person instituting any proceedings may at any time before the matter has been set down and thereafter by consent of the parties or leave of the court withdraw such proceedings, in any of which events he shall deliver a notice of withdrawal and may embody in such notice a consent to pay costs; and the taxing master shall tax such costs on the request of the other party.

(2) Any party in whose favour any decision or judgment has been given, may abandon such decision or judgment either in whole or in part by delivering notice thereof and such judgment or decision abandoned in part shall have effect subject to such abandonment.

The provisions of subrule (1) relating to costs shall mutatis mutandis apply in the case of a notice delivered in terms of this

subrule.

(3) If in any proceedings a settlement or an agreement to postpone or withdraw is reached, it shall be the duty of the attorney for the plaintiff or applicant immediately to inform the registrar accordingly.

(4) Unless such proceedings have been withdrawn, any party to a settlement which has been reduced to writing and signed by the

parties or their legal representatives but which has not been carried out, may apply for judgment in terms thereof on at least

five days' notice to all interested parties.”

[29] Section 4 of the agreement is entitled “[t]he Appeal Proceedings” and clause 4.1 states: “The Respondent withdraws its opposition to the appeal and any rights it may acquire and have acquired in respect of the appeal against the Appellant. The Appellant thereafter withdraws its appeal against Respondent. Appellant and Respondent agree to bear and pay their own legal costs and disbursements.”

[29] Section 4 of the agreement is entitled “[t]he Appeal Proceedings” and clause 4.1 states:

“The Respondent withdraws its opposition to the appeal and any rights it may acquire and have acquired in respect of the appeal against the Appellant. The Appellant thereafter withdraws its appeal against Respondent. Appellant and Respondent agree to bear and pay their own legal costs and disbursements.”

[30] See [1] on the proposition that parties can’t set aside judgments in rem by agreement.

[31] Liversidge v Anderson [1942] AC 206; [1941] UKHL 1l; [1941] 3 All ER 338.

[32]“‘When I use a word’, Humpty Dumpty said, in rather a scornful tone, ‘it means just what I choose it to mean - neither more nor less’. ‘The question is’, said Alice, “whether you can make words mean so many different things’. ‘The question is’, said Humpty Dumpty, ‘which is to be master - that’s all’”. L Carroll Through the Looking-Glass (Macmillan and Co, London 1872) chapter 6, quoted in Liversidge, above n 32 at 245.

[33] But see above n 30.

[34] SCA judgment above n 10 at paras 25-7.

[35] Commissioner for the South African Revenue Service v The Executors of Estate of Late Sidney Ellerine [2018] ZASCA 39 at para 31.

[36] See above n 27.

[37] Eke above n 6 at paras 25-6.

[38] Mpumalanga Department of Education v Hoërskool Ermelo [2009] ZACC 32; 2010 (2) SA 415 (CC); 2010 (3) BCLR 177 (CC) at para 96.

[39] Mvumvu v Minister of Transport [2011] ZACC 1; 2011 (2) SA 473 (CC); 2011 (5) BCLR 488 (CC) at para 40.

[40]Justice Alliance of South Africa v President of Republic of South Africa [2011] ZACC 23; 2011 (5) SA 388 (CC); 2011 (10) BCLR 1017 (CC) at para 33; Fose v Minister of Safety and Security [1997] ZACC 6; 1997 (3) SA 786 (CC); 1997 (7) BCLR 851 (CC) at para 96.

[41] See above [3].

[42] See rule 41(2) above n 28.

[43] SCA judgment above n 10 at paras 26-7.

[44] Ferreira above n 2 at para 28.

[45] Eke above n 6 at paras 25-6.

[46] Eke above n 6 at para 29.

[47] Id at para 31.

[48] Id at para 29. See also Moraitis Investments (Pty) Ltd v Montic Dairy (Pty) Ltd [2017] ZASCA 54; 2017 (5) 508 (SCA) at paras 10 and 16.

[49] Provincial Government: North West Province v Tsoga Developers CC [2016] ZACC 9; 2016 JDR 0553 (CC); 2016 (5) BCLR 687 (CC) at para 52.

[50] Department of Transport v Tasima (Pty) Ltd [2016] ZACC 39; 2017 (2) SA 622 (CC); 2017 (1) BCLR 1 (CC) at paras 177-99.

[51] To the extent that the SCA judgment above n 10 holds that an appeal court making a settlement agreement an order of court having the effect of setting aside judgment in rem need not have regard to the underlying merits of the dispute, I respectfully disagree.

[52] Eke above n 6 at paras 25-6.

[53] See [82] above.

[54] See [1]. However at [3] it is said that a court “must” give reasons when a court on appeal approves a settlement agreement that sets aside a judgment in rem. I do not read this paragraph to mean that an order not accompanied by reasons would be incompetent.

[55] Endumeni Municipality above n 18.

[56] See [29].

[57] Bothma-Batho Transport (Edms) Bpk v S Bothma & Seun Transport (Edms) Bpk [2013] ZASCA 176; 2014 (2) SA 494 (SCA) (Bothma-Batho Transport) at para 13.

[58] SCA judgment above n 10 at para 26.

[59] The relevant clauses of the settlement agreement are set out at [30] to [34].

[60] See [51] to [54] above.

[61] See [55] to [58].

[62] See [44] to [48].

[63] Compare Aussenkehr Farms (Pty) Ltd v Trio Transport [2002] ZASCA 28; 2002 (4) SA 483 (SCA) at para 25.