Minister of Finance and Others v Gore NO (230/2006) [2006] ZASCA 98 (8 September 2006)

THE SUPREME COURT OF APPEAL

OF SOUTH AFRICA

1. Case no: 230/2006

2. REPORTABLE

In the appeal between:

MINISTER OF FINANCE First Appellant

NATIONAL GOVERNMENT OF THE REPUBLIC OF SOUTH AFRICA Second Appellant

MINISTER OF WELFARE AND POPULATION DEVELOPMENT

Third Appellant

PREMIER OF THE WESTERN CAPE PROVINCE

Fourth Appellant

and

STEPHEN MALCOLM GORE NO Respondent

Before: Cameron JA, Mthiyane JA, Brand JA, Mlambo JA and Malan AJA

Heard: Tuesday 15 and Wednesday 16 August 2006

Judgment: Friday 8 September 2006

Prescription – Prescription Act 68 of 1969 – ‘knowledge’ of ‘the facts from which’ debt arises – must be justified, true belief – unjustified suspicion not constituting knowledge – Vicarious liability – for employee’s deliberate dishonesty – employees’ actions, though fraudulent, closely resembling what they were employed to do – policy considerations supporting imposition of liability for fraud – Causation – company proving that without fraud, tender would have been awarded to it

Neutral citation: Minister of Finance v Gore NO [2006] SCA 97 (RSA)

JUDGMENT

_______________________________________________________

CAMERON and BRAND JJA:

1. The liquidator of a defunct company, 3D-ID Systems (Pty) Ltd, sued the four appellants, who represent national government and the Western Cape provincial administration (the defendants), for damages arising from the fraudulent award of a tender in 1994 for which the company was a bidder. At the trial in the Pretoria High Court, where the defendants’ liability was separated as a first issue from the quantum of any damages, Hartzenberg J found for the plaintiff liquidator. (We refer indifferently to the current liquidator and the predecessor for whom he was substituted as ‘the plaintiff’; and to the defunct company and its close corporation antecedent as 3D-ID.) With the leave of the trial judge, the four defendants ((i) the Minister of Finance, responsible for the State Tender Board; (ii) the national government itself; (iii) the Minister of Welfare and Population Development, responsible for the payment of social pensions and welfare grants after April 1994 (‘national government’); and (iv) the Premier of the Western Cape, executive head of the Western Cape provincial administration (‘the province’)) now appeal.

2. The defendants’ first hurdle is that the appeal has lapsed. The appeal record should originally have been lodged by 6 December 2005, a date extended to 6 February 2006, by when this had still not been done, with no further extension granted. The record was eventually ready only on 31 March, and received the eventual date-stamp of this court’s registrar on 5 May 2006 – a gaping three-month chasm. The State Attorney’s explanation for the omissions that led to this is neither coherent nor entirely plausible and he must unavoidably be censured for ineptitude or inattention (or both). These while pronounced are, however, not so prodigious that condonation should be refused without regard to the merits of the matter, which we therefore turn to consider.

3. The disputed tender was government’s first attempt to employ automated fingerprint identification and verification technology for welfare pay-outs – and was designed to address massive fraud in registrations and payments that was plaguing not only the then Cape Provincial Administration (CPA), but other provinces too. As early as 1992, government identified this as a priority problem. And the CPA’s planned venture was seen as a possible blueprint for other provinces. But this was not to be: before even the closing date, the tender process was poisoned at its very heart by fraud within the CPA. That came about as follows.

4. 3D-ID had an ‘inside track’ on the tender requirements, for the simple reason that it had helped the CPA devise them. As early as 1992 its founder Mr Darryl Pamensky (who had previously supplied government with fingerprint identification pads) and Mr Melchior Rabie (plaintiff’s principal witness) met with Mr Anton Scholtz, a senior official in the CPA welfare department, to discuss possible solutions to the pay-out problem. But the breakthrough came in 1993 when a California corporation, Identicator, produced new fingerprint identification and verification technology that could rapidly compile and accurately search a huge database of fingerprints on a portable or personal computer (PC). 3D-ID formed an exclusive association with Identicator, securing sole South African rights to its innovation. More advanced discussions with the CPA now included ADJ (André) Louw, the CPA deputy director for social security in the welfare directorate, and from August to November 1993 highly successful demonstrations and tests of the product that Identicator specially developed for the CPA were conducted before various groups of officials from both provincial and national government.

5. The tender as eventually advertised was in two parts: the first (Part A) involved supplying only software and equipment; the second (Part B) entailed fully outsourcing the pay-out service. The technology required in either case had to be capable of performing three distinct functions: registering beneficiaries (enrolment); the complex electronic task of identifying new fingerprints as unduplicated by searching the entire database of already captured fingerprints (identification); and, once a ‘clean’ database of unduplicated fingerprints had been established, verifying any particular enrolled beneficiary’s presented fingerprint as identical to that already captured on the system (verification). In addition, the technology had to run on PCs, so that the triple function could be carried out at a large number of paypoints dispersed across the province.

6. The CPA was careful to emphasise to 3D-ID that the tender process had to be both open and authentic; but it seemed clear that Identicator had the only product anywhere in the world with the capacity to eliminate the fraud that was disabling the payments system. Rabie and Pamensky had not expected the tender to include Part B, but obtained substantial capital backing from a Swiss investor residing in South Africa, Mr Hans Dieter Fuchs. With the tender specifications drawn, the State Tender Board at the request of the CPA issued a call for tenders on 11 March 1994.

7. By the closing date of 11 April 1994, a total of thirteen entities had submitted tenders. 3D-ID and three others tendered for Part A as well as Part B; six for only Part A; while three tendered only for Part B. All the entities that tendered for Part B either tendered also for Part A or tendered for Part B in association with a Part A tenderer. One of these was Nisec CC, a corporation its sole member, Mr Michau Huisamen (a Port Elizabeth businessman with no previous experience of information technology), acquired ‘off the shelf’ from an accounting firm just days before the tender invitation. The CPA’s evaluation committee, formally chaired by the CPA’s director of social welfare services, Dr Terblanche, but in effect chaired by Louw (as will become clearer later), recommended that Nisec be awarded Part B of the tender. The CPA accepted this recommendation, and – to Rabie’s consternation and in the face of his protests – on 16 June 1994 the State Tender Board awarded a five-year contract to Nisec.

8. Rabie was convinced that skulduggery underlay the Nisec award. He threw all his efforts into trying to prove this. Before the end of July 1994 3D-ID signalled that it would challenge the award, and in September it launched a review application and sought to interdict the award. 3D-ID was not only refused access to the tender documentation, but officials from both the CPA and central government lodged affidavits vigorously defending the award. Nisec, a respondent in the review, obtained an order obliging 3D-ID to lodge security for its costs. Undaunted (after obtaining the formal record of decision-making in December 1994), 3D-ID brought an Anton Piller application in January 1995 to seize documents and files from the province, the State Tender Board and Nisec that it alleged would prove fraud. But the review, interdict and Anton Piller applications were all futile: and in March 1995 3D-ID was ordered to pay the costs of the latter on a punitive scale. Armed with this and other costs orders, Nisec in March obtained an order provisionally winding up 3D-ID, which was made final in September 1995.

9. At this stage Rabie’s quest to prove irregularity or fraud seemed to have foundered. But he persisted. In September 1995, he laid a complaint and filed an affidavit with the office of the director of the Office for Serious Economic Offences (OSEO), which eventually led to an OSEO investigation. The award was in the meanwhile unravelling. Within a very short time, problems with capacity had started manifesting, and in early 1995 the provincial administration of the new Western Cape province (PAWC) commissioned a major accounting firm to investigate the tender. Louw and Scholtz had in the meanwhile resigned their provincial administration jobs and taken up employment with Nisec on highly remunerative terms. But Nisec’s incapacity to deliver in terms of the tender soon became plain, and PAWC recommended in October 1996 that its contract be terminated. The Western Cape tender board cancelled the contract in December 1996 on the grounds of Nisec’s incapacity and because the award had been improperly obtained. Nisec challenged this – and though its review application failed, a full bench of the Cape High Court in February 1997 found insufficient evidence that improper means had been used to obtain the tender.

10. Time nevertheless vindicated Rabie’s indignant assertions. An OSEO examination of Terblanche’s secretary’s computer hard drive eventually revealed that ten days before the closing date, Louw and Scholtz – fraudulently conspiring with Huisamen and Mr André Scholtz, Scholtz’s brother (a provincial employee in Port Elizabeth) – had put together Nisec’s tender on Friday 1 April 1994 at the CPA offices; that Louw and Scholtz had corruptly negotiated contracts of employment for themselves with Nisec, plus substantial bribes (which Huisamen paid into their wives’ banking accounts); that Louw, left to steer the evaluation committee and to draft submissions to the new provincial executive and to the State Tender Board, had with lies and distortions manipulated the entire process to secure the award to Nisec.

11. Thus armed, the plaintiff issued summons claiming damages from the four defendants. The summons was served in January 1999, nearly five years after the events in issue. When the matter came to trial in November 2004, the defendants sought by last-minute amendment to introduce a plea that the plaintiff’s claims had prescribed; and contended that national government and the province were not vicariously liable for Louw’s and Scholtz’s corrupt conduct; and that 3D-ID would in any event not have secured the contract and had thus suffered no damages. These were the issues that were tried before Hartzenberg J, who found for the plaintiff on all of them. On appeal the defendants persist in their trial defences. In addition, the province contended that public policy demanded that a public body be immunised from liability for the consequences of fraud committed in the course of a tender process. We examine these four defences in turn.

First defence: Did the claim become prescribed?

12. In terms of Prescription Act 68 of 1969, the period of prescription in respect of the debts the plaintiff claimed was three years after they became due. Section 12 (3) provides –

‘A debt shall not be deemed to be due until the creditor has knowledge of the identity of the debtor and of the facts from which the debt arises: Provided that a creditor shall be deemed to have such knowledge if he could have acquired it through exercising reasonable care’.

If s 2(1) of the Limitation of Legal Proceedings (Provincial and Local Authorities) Act 94 of 1970 applied to the province¹ (which the plaintiff for reasons it is not necessary to consider put in issue), the limitation period would be 24 months.

13. The question thus is whether the plaintiff had ‘knowledge’ of ‘the facts from which’ the debt arose before 15 January 1996 (or, if a different period applies to the province, before 15 January 1997). It is well established that the defendants bear the burden of proving when the plaintiff acquired (or should be deemed to have acquired) the knowledge in question. National government contended that Rabie had all the knowledge needed to institute action by at the latest January 1995. The province argued that Rabie had sufficient facts at latest when he lodged his OSEO complaint and affidavit in September 1995. Either contention if sound would render the plaintiff’s claim unenforceable.

14. It was common cause that Rabie’s knowledge before 3D-ID’s winding-up should be imputed to the liquidators. Hartzenberg J held that Rabie’s knowledge after 3D-ID was finally liquidated (at about the time of the OSEO complaint) should also be imputed to the liquidator. The plaintiff formally denied this in its draft replication to the prescription plea, and filed a supporting affidavit from the liquidator. On appeal it challenged Hartzenberg J’s conclusion; but we find no reason to fault it. Rabie, who, more than the liquidator, was the force behind the litigation, had an incentive to convey to him any information he obtained and to report any action he took, despite the secrecy and inhibition surrounding the OSEO affidavit. The liquidator was not called to testify, and in the absence of contrary evidence we consider that if he did not actually have the details of Rabie’s OSEO evidence, he could have acquired them by exercising reasonable care. The matter must in our view be decided on the basis that Rabie conveyed to the liquidator all he knew as he came to know it.

15. Hartzenberg J concentrated on Rabie’s state of mind, and ‘whether the conduct of the defendants was convincing enough to dissuade a prospective plaintiff from instituting action’. He found that Rabie had no more than a suspicion that fraud had been committed, without any ‘witness to substantiate’ it. He found that the stand taken under oath by the province’s officials ‘was so convincingly and emphatically contradictory’ to any suggestion of fraud that the delay could not be faulted. Far from concluding that Rabie could reasonably have acquired knowledge earlier, Hartzenberg J found he had done all in his power to acquire such knowledge, but his vigorous efforts had proved fruitless.

16. These conclusions are hard to fault. The statutory prescription periods are meant to protect defendants from undue delay by litigants who are laggardly in enforcing their rights. To suggest that the plaintiff was dilatory would be inapt, to say the least. It would therefore be most surprising if it were to be non-suited for delay. In our view that is not the law.

17. This court has in a series of decisions emphasised that time begins to run against the creditor when it has the minimum facts that are necessary to institute action. The running of prescription is not postponed until a creditor becomes aware of the full extent of its legal rights,² nor until the creditor has evidence that would enable it to prove a case ‘comfortably’.³ The defendants relied on these authorities to contend that Rabie knew at the latest by the latter half of 1995 that Louw and Scholtz had defrauded 3D-ID out of its tender. They pointed out that Rabie insistently asserted under oath, starting with his replying affidavit in the review (October 1994), and repeated in his Anton Piller (January 1995) and liquidation affidavits (April 1995), that fraud tainted the tender process. The allegations of fraud then made found expression later in the particulars of claim.

18. Rabie certainly did cry fraud soon after 3D-ID lost the tender. But what did he know when he did so? The defendants’ argument seems to us to mistake the nature of ‘knowledge’ that is required to trigger the running of prescriptive time. Mere opinion or supposition is not enough: there must be justified, true belief. Belief on its own is insufficient. Belief that happens to be true (as Rabie had) is also insufficient. For there to be knowledge, the belief must be justified.

19. It is well established in our law that:

1. Knowledge is not confined to the mental state of awareness of facts that is produced by personally witnessing or participating in events, or by being the direct recipient of first-hand evidence about them;

2. It extends to a conviction or belief that is engendered by or inferred from attendant circumstances;

3. On the other hand, mere suspicion not amounting to conviction or belief justifiably inferred from attendant circumstances does not amount to knowledge.⁴

It follows that belief that is without apparent warrant is not knowledge; nor is assertion and unjustified suspicion, however passionately harboured; still less is vehemently controverted allegation or subjective conviction.

20. What Rabie knew in essence was that only 3D-ID’s technology could meet the demanding tender specifications. When 3D-ID did not win the award, he suspected that something must have been amiss in the tender process. His conviction was strengthened by two calls he received: one from an anonymous caller claiming to be within the provincial administration, and the other from a fellow tenderer. Added to this were stray indications he gleaned of misrepresentation and irregularity in the procedure. From this he inferred with passionate certainty that fraud must have taken place; but he lacked a firm evidentiary basis for his belief.

21. The affidavits the defendants invoke to establish the fatal delay abound with assertive contentions such as ‘I contend’, ‘it is clear’, ‘I submit’ and ‘there must have been’. Even Rabie’s OSEO affidavit, which provides the high point for the defendants’ argument, is replete with inferential assertion: ‘the only reasonable inference that can be drawn … is’; ‘based on the facts and information herein recorded … it can reasonably be inferred that …’. All this reveals Rabie’s want of adequate proof.

22. The latter point deserves elaboration. That there must have been fraud was an inference Rabie drew from the facts mentioned earlier, namely 3D-ID’s superior technology, Nisec’s palpable inexperience and attendant indications of processual error and misrepresentation. Counsel for the province contended that Rabie’s long pre-tender collaboration with Louw and Scholtz, and his knowledge of 3D-ID’s decisive technological edge, meant that his conclusion that there must have been fraud was more than merely speculative. But Rabie’s conclusion continued to rest on speculative inference, and he had no direct means of knowing that fraud had in fact been perpetrated. Knowledge of a fact can derive from inference, but belief in the fact becomes knowledge only once justification for the belief exists. This will generally mean that the means of establishing it must exist. This Rabie did not have until much later, and no amount of vehemence on his part could convert his subjective conviction into fact.

23. This follows not only from his want of proof, but from the response his allegations elicited:

1. The provincial administration and its legal representatives went out of their way to confute Rabie’s fraud claims as baseless, frivolous, vexatious, scurrilous and defamatory. The central actors, including the most senior officials in the tender process, went on record to vouch for its propriety, and assured the court that Rabie’s claims were ‘no more than unfounded accusations, without any evidentiary basis’.

2. Rabie was forced to withdraw the allegations in his replying affidavit in the attempt to review the Nisec tender, and likewise to withdraw the averments in his Anton Piller founding affidavit.

3. Repeated recourse to legal action, based on ‘unfounded’ allegations of fraud, not only met with failure but was visited with judicial rebuke in the form of a punitive costs order. The review application failed; the ex parte order 3D-ID obtained in the Anton Piller application was discharged with costs on a punitive scale; Nisec’s application to liquidate 3D-ID on the basis of the unpaid costs orders succeeded, despite strenuous opposition from Rabie.

4. Even when the province itself concluded at the end of 1996 that Nisec’s award had been improperly obtained, the full bench of the Cape high court determined in February 1997 in Nisec’s unsuccessful challenge that impropriety had not been established.

24. Despite the vehemence of his convictions, the response to Rabie’s claims – including the judicial discountenancing of his attempts to vindicate his views in various court actions – was such that they did not constitute justified belief under the statute.⁵ In fact, as plaintiff’s counsel pointed out, with hindsight it is evident that he was groping in the dark. On the one hand, he claimed that Terblanche ‘and possibly Mr Wentzel’ (the chairman of the State Tender Board in the province) knowingly colluded with Scholtz and Louw. This was wrong. On the other, he inferred that ‘there must have been’ fraudulent complicity between Louw, Scholtz and Huisamen. This was right: but he lacked the means to prove it.

25. Rabie acquired the minimum knowledge needed to institute action only at the end of 1998, when OSEO finally released the evidence that showed that the Nisec tender had been prepared on a CPA computer. This was ‘the smoking gun’ that senior counsel in February 1997 advised him to obtain before he contemplated further litigation based on fraud. With this in hand, the plaintiff promptly issued summons. It was not time-barred when it did.

Second defence: Are the defendants vicariously liable for the fraud of Louw and Scholtz?

26. The defendants contended that Louw and Scholtz were acting outside the course and scope of their employment with the administration in perpetrating the fraud, and thus that their employers were not liable for any loss their conduct may have inflicted on the plaintiff. They laid emphasis on a number of egregiously dishonest acts Louw and Scholtz committed that were alien to their responsibility to the provincial administration as stewards of the tender process. These included:

1. the corrupt agreement with Huisamen to secure the award to Nisec even before the tender was advertised;

2. the fact that Louw and Scholtz, with Huisamen and Scholtz’s brother, prepared Nisec’s tender – a prospective competitor – on administration premises;

3. their entering Nisec – a shelf corporation with no experience or capacity in information technology – in the tender race;

4. the manipulation of the entire award process by concealments and distortions and deliberate lies;

5. that Louw and Scholtz secured jobs for themselves with Nisec even before the tender was awarded.

27. These aspects underscore the fact that imposition of vicarious liability on an employer for an employee’s deliberate wrongdoing creates special difficulties, as to both its conceptual basis and the policy justifications underlying it.⁶ The observation of Watermeyer CJ that ‘the dividing line which separates acts within the scope of a servant’s employment from those without is one impossible to draw with certainty’⁷ applies with particular force in these cases. Yet, while the act of an employee who steals from – or defrauds –the employer is the very antithesis of an act in the course and scope of employment, there is no general principle that an employer cannot be responsible for an employee’s intentional wrongful conduct that causes the employer loss.⁸ On the contrary, instances of such liability are by no means rare.⁹ But the difficulties these cases raise make it important to bring to the fore the policy reasons that warrant imposing liability in each case.¹⁰

28. Even though a deliberately dishonest act that, subjectively seen, was committed solely for the employee’s own interests and purposes may fall outside the ambit of conduct that renders the employer liable, it is in our law established that liability may nevertheless follow if, objectively seen, there is a ‘sufficiently close link’ between the self-directed conduct and the employer’s business.¹¹ Applying this, the traditional two-pronged test, Hartzenberg J found that the defendants failed on both the subjective and objective components: a conclusion that seems to us to be clearly correct.

29. However gross the violation of their duties by Louw and Scholtz, it cannot be gainsaid that all their actions that were directed at wrongfully securing the contract for Nisec were nonetheless performed so that the tender would be awarded. The effect of their subjective intentions was thus not wholly self-directed. Indeed, as the trial judge observed, although Nisec suffered from manifest incapacities, Louw and Scholtz could hardly have regarded it as their future lifeline if they thought that it could not perform the contract at all. Louw and Scholtz of course did not testify, but the circumstances point overwhelmingly to the probability that they saw Nisec’s carrying out the tender as a lucrative continuing source of gain for themselves. Their subjective intentions are therefore very far from absolving the defendants from liability.

30. And, as in Minister van Veiligheid en Sekuriteit v Japmoco BK h/a Status Motors,¹² the objective nature of the employees’ actions also points to liability. Though they were defrauding both their employer and 3D-ID (as well as the other tenderers), their actions were tightly aligned to the functions they were employed to perform. To draw the distinction – admittedly fine – that applied in Japmoco, the award of the tender to Nisec was false, but it was not a total fake.¹³ This case seems to us to fall clearly within the line of liability drawn in Japmoco. Even when the full bench considered the evidence surrounding the tender award in early 1997, it rejected the indications of impropriety that the provincial administration proffered: which serves to show how closely the employees’ actions, though fraudulent, resembled what they were employed to do. This closeness of purpose, planning and effect, indicate that all the policy reasons for requiring the employer to bear the burden of its employees’ wrongdoing apply in this case, while no countervailing considerations apply. The defendants cannot escape vicarious liability.

Third defence: Causation – would 3D-ID have been awarded the tender?

31. We turn to the issue of causation. The plaintiff’s case is that, had it not been for the fraudulent and corrupt acts of Louw and Scholtz, the evaluation committee would have recommended the award of, and the State Tender Board would not have awarded the tender to Nisec but to 3D-ID. This the defendants deny. The question of causation, it is often said, is one of fact. But of course, as Lord Hoffmann explained in Fairchild v Glenhaven Funeral Services,¹⁴ this only means that the answer depends on fact. The question itself is formulated by law.

32. In our law the time-honoured way of formulating the question is in the form of the ‘but for’ test. Can it be said that, but for the wrongful act complained of, the loss concerned would not have ensued? Applying this requires the process of inferential reasoning described by Corbett CJ in International Shipping Co (Pty) Ltd v Bentley:¹⁵ What would have happened if the wrongful conduct is mentally eliminated and hypothetically replaced with lawful conduct? A plaintiff who can establish that, in such event, the loss would, on a preponderance of probabilities, not have occurred, recovers his damages in full, because causation is regarded as having been established as a fact. A plaintiff who cannot do so will get nothing. That there is no discount either way stems from the nature of the inferential process: the verdict must go one way or the other even if the scales are tipped only slightly in one direction (see eg Allied Maples Group Ltd v Simmons & Simmons (a firm);¹⁶ Minister van Veiligheid en Sekuriteit v Geldenhuys)¹⁷.

33. With reference to the onus resting on plaintiff, it is sometimes said that the prospect of avoiding the damages through the hypothetical elimination of the wrongful conduct, must be more than 50%. This is often followed by the criticism that the resulting all-or-nothing effect of the approach is unsatisfactory and unfair. A plaintiff who can establish a 51% chance, so it is said, gets everything, while a 49% prospect results in total failure. This however is not how the process of legal reasoning works. The legal mind enquires: what is more likely? The issue is one of persuasion, which is ill reflected in formulaic quantification. The question of percentages does not arise (see to this effect Baroness Hale in Gregg v Scott).¹⁸ Application of the ‘but for’ test is not based on mathematics, pure science or philosophy. It is a matter of common sense based on the practical way in which the ordinary person’s mind works against the background of everyday life experiences. Or, as was pointed out in similar vein by Nugent JA in Minister of Safety and Security v Van Duivenboden:

‘A plaintiff is not required to establish the causal link with certainty, but only to establish that the wrongful conduct was probably a cause of the loss, which calls for a sensible retrospective analysis of what would probably have occurred, based upon the evidence and what can be expected to occur in the ordinary course of human affairs rather than metaphysics.’¹⁹

34. Both the recommendation by the evaluation committee and the State Tender Board’s decision to award the tender, involved an administrative discretion that required the exercise of judgment. Determining what decision they were likely to have reached in the exercise of their discretion, but for the fraudulent conduct of Louw and Scholtz, inevitably requires some measure of second guessing the administrative functionaries. Fortunately we can take guidance from the decision of this court in Minister of Safety and Security v Carmichele²⁰ that, in a situation such as this, the question is objective: how is a reasonable functionary likely to have exercised that discretion? Since the two administrative bodies concerned had to exercise their discretion as part of the state tendering process provided for s 187 of the interim Constitution, which then applied, it must also be accepted, we think, that reasonable bodies in their position would have been guided by the constitutional norms underwritten by that section. They would have applied the values of a fair, public and competitive tender system.

35. Central to the plaintiff’s case that, without the fraud, 3D-ID would have been the successful tenderer, is the contention that 3D-ID’s tender was the only one submitted that actually complied with the specifications of the invitation. In the circumstances, the plaintiff contended, the fact that 3D-ID proved to be the most expensive of all the tenderers would, on the probabilities, not have prevented it from gaining the award. To this the defendants’ answer – as it eventually turned out – amounted to a contradiction of the plaintiff’s central contention in all its parts. (a) First, they denied that 3D-ID’s tender complied with the tender specifications. (b) Second, they maintained that another tenderer, Cash Payment Systems (CPS), was capable of meeting the tender requirements while Nisec (who admittedly did not comply with the requirements) would nevertheless have been able to persuade the evaluation committee that it did. (c) Third, they contended that, if 3D-ID had been the only compliant tenderer, the State Tender Board would have decided, in the circumstances prevailing, not to award the tender at all.

36. We deliberately made reference to the defendants’ case ‘as it eventually turned out’ because this was not how it was originally pleaded or even as it was put to the plaintiff’s witnesses. In fact, the defendants’ conduct of their case understandably reminded Hartzenberg J of trench warfare. Though this does not mean that one or more of the trenches might not afford adequate protection, the implications cannot be ignored in evaluating the merits of the defences.

(a) Did 3D-ID’s tender comply with the tender specifications?

37. The issue whether 3D-ID’s tender complied with the specifications of the invitation must be considered against the background of how the invitation itself came into existence. More specifically, it will be remembered that the invitation resulted from 3D-ID/Identicator demonstrations to the CPA and national government. The purpose of the demonstrations, which included simulated tests and live payouts, was to persuade the institutions that the fingerprint identification technique devised and adapted by Identicator could provide the solution to the problems experienced by the CPA in the payment of social pensions, and particularly the fraud permeating the payout system. It is common cause that eventually the responsible officers of the CPA were persuaded that the 3D-ID/Identicator product was indeed the answer. Likewise it was common cause that 3D-ID and Identicator assisted in preparing the technological specifications for the invitation so as to ensure that the successful tenderer would deliver the same result.

38. This would make it surprising if the 3D-ID tender did not at least comply with the technological requirements specified. Nonetheless, the defendants expended much time and energy at the trial persisting in their denial that it did. On appeal these defences, which failed to impress Hartzenberg J, were not pursued. Before us the focus shifted to the non-technical aspects. The defendants now contended that 3D-ID had failed to fulfil the tender requirements in two respects: (i) first, in that it did not provide for the enrolment of future pension beneficiaries; and (ii) second, because its proposal for security arrangements was deficient.

39. These contentions must be understood against the background of how Part B of the tender invitation – which was eventually awarded – was framed. The fingerprint technology was crucial to this portion of the tender. For that reason, very specific requirements were formulated. So para 11.3 of the invitation required that:

‘All recipients of social pensions and other welfare grants must be enrolled on software capable of registering fingerprints and such software must be able to positively identify and verify recipients . . . Tenderers must submit a detailed implementation plan to enrol existing clients and the cost per head of enrolling the existing clients must also be indicated.’

Para 11.5 provided that:

‘The successful tenderer will be required to use the software of the CPA to render the payout function. The software to be used for the fingerprint identification and verification must be compatible and interface with the software presently used by the CPA.’

Para 12.17 admonished that:

‘The standards set out in paras 11.1 to 11.7 [containing the technical specifications] are the minimum standards acceptable to the CPA. Tenderers that cannot comply with these standards will not be considered.’

40. Non-technical aspects were treated in far less specific terms by the rather laconic pronouncement in para 12.1 that:

‘Tenderers are expected to tender a workable solution [for rendering the service of pension payouts previously performed by the CPA itself].’

(i) Enrolment of future beneficiaries

41. While provision for enrolling existing pensioners was mandatory (para 11.3), the tender invitation did not specifically refer to enrolment of future pension beneficiaries. However, in response to an enquiry by CPS whether ‘there are any specific towns where ongoing registration sites must be located’, Louw wrote to all tenderers that ‘ongoing registration of future beneficiaries must form part of the solution offered’. The ‘solution offered’ by 3D-ID, under the heading ‘Future Applicants’ was this:

‘Any future applicants will be enrolled at the CPA’s present regional offices as the prospective clients have to come in to the offices to complete their application forms. This would be the most effective way to complete this task and 3D-ID Systems will quote the CPA separately (see Annexure D) for the equipment needed to perform this task and pricing.’

42. At the time the CPA had thirteen regional offices. Under 3D-ID’s proposal, future applicants would therefore have to go to one of these, where they would be enrolled by CPA officials using the equipment 3D-ID offered to provide at the quoted price of R1,2m. According to Rabie’s testimony this proposal was based on his understanding that, according to the existing system, pension applications could be made only at a regional office. His reasoning, Rabie said, was that future applicants should be required to make their fingerprint enrolment at the time and place of application. It turned out, however, that Rabie was mistaken in believing that applications could be made only at a regional office. In fact, pension applications could also be made at one of at least 30 ‘service points’ in the Western Province alone.

43. According to its minutes and memoranda, the evaluation committee found no deficiencies in the workable solution proposed by 3D-ID. In fact, it was found to be the only viable implementation plan – apart, supposedly, from that Nisec offered. In answer to 3D-ID’s formal objections to the award to Nisec, the CPA indicated that, although 3D-ID’s compliance with the tender specifications was not disputed, Nisec succeeded because its tender, which was lower, had also complied.

44. In the review and other proceedings that followed, both the CPA and the State Tender Board persisted in this. Affidavits on behalf of the CPA were mostly deposed to by Terblanche, while affidavits on behalf of the State Tender Board were deposed to by Wentzel. Terblanche affirmed that:

‘I do not doubt that the applicant’s [ie 3D-ID’s] product complies with the tender specification. It is, however, not the only product that does so. [Nisec’s] tender was also in accordance with Part B of the tender specification.’

Wentzel stated:

‘What it [the State Tender Board] did accept was that both [Nisec] and [3D-ID] complied with the minimum requirements set out in the tender invitation.’

45. In the pleadings and at the trial, the defendants again omitted to proffer future enrolments as a subject of any concern. This aspect first reared its head at the pre-trial conference when the plaintiff enquired from the province whether it 'now admits that 3D-ID's tender complied with all the requirements of the tender'. The province responded:

'No. Price carried a weight of 50%. 3D-ID's price was the highest and thus did not meet the requirement . . . . Furthermore, in addition to its tender, 3D-ID required the CPA to purchase the enrolment hardware and software in the amount of R1,2m.'

46. Even now the contention was thus not that 3D-ID's proposal regarding future enrolment rendered its tender non-compliant. What the allusion to ‘future enrolment’ obviously contrived to show was that the 3D-ID tender not only was the most expensive, but also carried a further effective cost of over R1m. Rather surprisingly in these circumstances, Terblanche then testified at the trial that 3D-ID’s proposal regarding future enrolment was so deficient that it completely disqualified it as a tenderer.

47. The reasons Terblanche proffered for this contention were essentially twofold. In the first place, he said, the whole purpose of Part B, as underlined by Louw’s response to the CPS query, was to outsource both the payment and the enrolment of all pension beneficiaries, including future applicants. It followed, Terblanche said, that 3D-ID's tender did not amount to a total solution, in that it proposed to 'back source' future enrolments to the CPA. Apart from the fact that the 3D-ID tender would require the CPA to purchase additional equipment, Terblanche added, it also required the CPA to make personnel and office space available at regional offices. His second objection was aimed at 3D-ID’s proposal that enrolments had to take place at regional offices (as opposed to the CPA’s network of existing service points). This, Terblanche said, would require indigent applicants in rural areas to travel large distances at their own expense, which was in conflict with the CPA's policy of bringing its services to the people. The CPA’s objective was accordingly that pensioners should eventually be able to enrol, not only at service points, but at any one of its 650 payout points. When asked in cross-examination why these vital deficiencies in the tender had never been pointed out before, his explanation was that both he and the evaluation committee concentrated only on the technological aspects, and not so much on the ‘workable solution’ proposed.

48. What is important to recognise, we think, is that 3D-ID's tender did not neglect to address the issue of ongoing enrolments. As Rabie explained, 3D-ID decided, on the basis of his understanding of the pension payment process, that the best and most practical solution would be to do registrations and enrolments at the same time and at the same venue. It was therefore not a question of 3D-ID’s being unwilling to do future enrolments or trying to minimise its services under the tender. Instead, 3D-ID's proposal was the considered result of Rabie's conclusion as to what would be in the interest of all concerned. The advantage of his proposal to the prospective pensioners, as Rabie saw it, would be that they could receive their pensions immediately after the registration/enrolment process, as opposed to waiting for a further period of at least one month (which he thought would result from Terblanche’s procedure).

49. We find it unnecessary to enter into the debate between Rabie and Terblanche as to which of the solutions would be the best. Whether the 3D-ID proposal, objectively speaking, constituted the best solution is not the issue. The issue is whether it was a workable solution: for this is what the tender invitation required. Otherwise stated, since considerable flexibility was given to prospective tenderers to propose a workable solution, a tenderer could not be disqualified because some or other CPA official might prefer a different solution. Such a ‘concealed trench’ approach would, in our view, be in conflict with the constitutional norm requiring a fair tender process. 3D-ID's tender contained a motivated plan for future enrolments. Whether or not Terblanche favoured it is of no real consequence. It was undeniably workable and the tender could therefore not be disqualified on that basis.

50. In any event, there was nothing to prevent 3D-ID or the CPA, under the 3D-ID proposal from enrolling new pensioners at the pay points when registering them, if that ultimately proved to be the preferred solution. The software and other technology tendered by 3D-ID were capable of running on personal computers and thus could be used at payout points in the field. In the circumstances the fair solution would be, not to disqualify the tender, but simply to inform the tenderer, who complied with all the mandatory requirements of the tender, that the CPA preferred a workable solution that differed from the one proposed on some relatively minor aspect initially left open in the invitation. The fact that the 3D-ID proposal required the CPA to purchase additional equipment and technology for future enrolments was hardly likely to constitute an insurmountable hurdle. It was ultimately, as indicated by the province at the pre-trial conference, something relevant simply to the evaluation of the price of 3D-ID's tender.

51. Significant, in our view, is that Terblanche's objections in relation to 3D-ID's plan for future enrolments were seemingly of no concern to any of the ultimate decision makers, namely the evaluation committee, the CPA or the State Tender Board. As Terblanche conceded, he was in no position to speak for the evaluation committee because despite being its nominal head he was effectively absent from all its deliberations. In any event, Terblanche's statements at the time clearly did not regard the 3D-ID proposal for future enrolment as disqualifying its tender. His ex post facto explanation for the total absence of reference to a factor that purportedly rendered the whole tender patently unsuitable, ie that he concentrated on the technical aspects, is not convincing.

(ii) Security arrangements

52. As in the case of future enrolments, provision of security equipment and arrangements formed no part of the tender requirements. Nevertheless, it was accepted by all concerned, including Rabie, that, in view of the large amounts of cash to be distributed and the logistical difficulties associated with the execution of Part B, any 'workable solution' would have to incorporate some form of provision for security. In its tender 3D-ID therefore specified what it planned to provide. It set out the complement of security personnel to be employed; the number of vehicles to be used; and other equipment it intended to procure. Reference was also made to the fact that 3D-ID had succeeded in obtaining insurance from Lloyds of London in an amount sufficient to secure its obligations to the CPA for pension money delivered in its care.

53. None of the tender evaluation committee's minutes, worksheets or memoranda contained any criticism of 3D-ID's proposed security system. In their answers to 3D-ID's challenge to the Nisec award, neither the CPA nor the State Tender Board indicated that 3D-ID was disqualified because of some deficiency in its security proposal. On the contrary, as mentioned earlier, they admitted then that 3D-ID's tender accorded with the tender specifications. The sole objection then raised was that 3D-ID's tender was too expensive. What is more, as with future enrolments, the defendants’ pleadings did not raise security deficiencies, nor were they even properly put to the plaintiff's witnesses.

54. Despite this lack of forewarning, the province levelled a two-pronged attack at 3D-ID’s security component. The first line of attack was based on the evidence of Terblanche, to the effect that, in view of the security problems the CPA experienced in the past, the tender would not have been awarded to any tenderer who was not associated with an established security operator. The second objection relied on the expert evidence of Mr Richard Phillips, the general manager of Fidelity Guards Cash Management Services (Pty) Ltd ('Fidelity Guards') who had 28 years experience in the security industry, and who represented Fidelity Guards during the tender process. Phillips criticised the 3D-ID proposal on various technical aspects which, in his view, rendered it practically unworkable.

55. It is difficult to evaluate Phillip’s criticism on its merits. Because of the way in which defendants conducted their case, most of the alleged technical deficiencies Phillips referred to were not put to the plaintiff’s witnesses. As a result, answers to his difficulties could only be suggested to him during cross-examination in the form of hypothetical solutions. Although he expressed doubt about the feasibility of these solutions, he could not say that they were beyond the realms of possibility. Fortunately, in the circumstances, it is not necessary for us to decide the matter on the merits of the technical debate introduced by Phillips.

56. This is because it is clear in our view that the evaluation committee and the State Tender Board did not approach the matter of security at nearly the level of Phillips’s technicality. In fact, both these bodies clearly regarded the matter of security as one of the non-essential elements of the tender. With reference to these non-essential issues their attitude appears to have been that lesser difficulties could be ironed out, even after the award of the tender, as long as the tenderer could render the essential computerised fingerprinting services. The indifference of both to individual tenderers’ security arrangements is illustrated by their response to a complaint by Fidelity Guards – after the award of the tender to Nisec – that its security vehicles and apparatus, which had been prepared at great expense for purposes of the tender, were not even inspected by the evaluation committee. Wentzel answered:

‘Regarding the question of equipment and mobile pay-out vehicles, I wish to elucidate that this was not a requirement of the tender and therefore not a criterion for evaluation on its own.’

57. Phillips also understood that a certain degree of negotiation regarding matters such as security would take place after acceptance of the tender. That this understanding was correct is borne out by the fact that Nisec only furnished details of how it would discharge its security obligations after being awarded the tender. This approach to non-essential elements was sensible. As long as a tenderer complied with the specified requirements, why should it not be allowed to negotiate aspects that were not specified? Why should an otherwise compliant tender be rejected out of hand, merely because the CPA did not agree with some or other aspect of its proposed ‘workable solution’?

58. This, in our view, answers also Terblanche’s assertion that the tender would only be awarded to a tenderer associated with an experienced security operator. Rabie’s stated belief was that he could have persuaded the evaluation committee and, ultimately, the State Tender Board that, although the directors of 3D-ID themselves had no experience in the security industry, they could satisfy the CPA’s security requirements. One possibility he advanced was that 3D-ID could have acquired experienced personnel from existing security firms. An alternative was that it could buy a security business. Despite Terblanche’s insistence to the contrary, we can see no reason in principle why an evaluation committee, acting reasonably, could not have been persuaded by these.

59. But, even if the State Tender Board were to have been as insistent as Terblanche on an association with an established security operator, Rabie testified – and Phillips confirmed – that it was extremely unlikely that the successful tenderer would have had any difficulty in finding one. As Rabie put it, security companies were queuing up to provide that service, particularly since it was known in the industry that the CPA’s plan for privatisation of pension payment was a pilot program for the rest of South Africa. If the evaluation committee therefore took up an intransigent attitude, 3D-ID would in all likelihood have been able to come to an arrangement with an established security company.

(b) The CPS tender and the Nisec tender

60. The second leg of the province’s argument on causation – which national government did not embrace – was that another tenderer, CPS, also complied with the technological requirements and could therefore also have been awarded the tender. In fact, the province contended, because 3D-ID’s tender was nearly double that of CPS, the latter was the most likely candidate.

61. As has by now become a recurring theme, this part of the province’s case was not foreshadowed in its pleadings. On the contrary, when the plaintiff asked at the trial particulars stage, whether it is contended that ‘any other tenderers in fact complied with the technological and other requirements of the tender’, the province answered:

‘According to the evaluation committee, the other tenderers did not comply with the technological and other requirements of the tender.’

This was obviously evasive and ambiguous. But, by not distancing itself from the evaluation committee’s stated view, the province obviously created the impression that it agreed. That was its pleaded case.

62. The viewpoint that no other tenderer did – or was in fact able to – meet the technological requirements was supported by the plaintiff’s expert, Mr Peter Bouwer, who was employed by another competing tenderer, Q-Data. The problem, he explained, lay in the very specific requirement posed by para 11.3 of the invitation, that pension beneficiaries ‘must be enrolled on software capable of registering fingerprints and such software must be able to positively identify and verify recipients’. Though software performing both enrolment and verification was relatively freely available at the time, he said, this could not perform the identification function as well. In preparing Q-Data’s tender, he testified, he was asked, as the technical expert of the company, to find software that could execute all three the required functions. But, he said, although he searched both locally and abroad and despite spending a large amount on the search, the only software that was able to register (or enrol), verify and identify fingerprints, all on the same system, was that of Identicator. Q-Data’s attempts to obtain this technology from Identicator were unsuccessful, because of its commitment to 3D-ID.

63. Contrary to the province’s case as pleaded, it was then put to Bouwer that the software tendered by CPS could in fact meet the requirements of the tender. But Bouwer’s opinion was that it could not. Without entering too deeply into the technical debate that ensued, the difficulty raised by Bouwer was essentially that CPS’s tender relied on two different software systems. While one system was utilised to perform the registration and verification of fingerprints, the identification function was to be carried out by a different system. Moreover, so Bouwer testified, even if the single software system requirement was ignored, CPS’s two software systems did not speak the same computer language and were therefore incompatible. After some technical debate in cross-examination, he conceded, however, that although he did not believe it would work in practice, he could not exclude the theoretical possibility that CPS’s two software systems could be combined to produce the results the tender required, although in a different manner.

64. The thesis that the CPS technology could be harnessed to produce the required results was supported by an expert the province called to testify, Mr Leonard Klopper. Though Klopper admitted that he had never tested his thesis in practice, his view was that it was hypothetically feasible to combine the two CPS software systems in that way. We find it unnecessary to declare the victor in this technical debate. It is not the province’s case that CPS did in fact offer the suggested solution. Since Klopper himself did not even read the CPS tender, he could not comment on any solution it contained. Bouwer’s undisputed evidence, on the other hand, was that apart from the fact that the CPS tender did not offer the solution suggested by Klopper, it could not have done so, because, to his knowledge, CPS was conducting an unsuccessful search for an answer to the technical difficulties posed by the tender. This is borne out by the admission in other proceedings of Mr S Etzebeth, the managing director of CPS, that as far as CPS was concerned, it did not at that stage believe it had the required technology.

65. This, in our view, renders the feasibility of the solution suggested by Klopper entirely irrelevant. The question is not whether the technology referred to in the CPS tender enabled an expert, with or without the benefit of hindsight, to come up with some solution that complied with the requirements of the tender, but whether the tender submitted by CPS in itself offered such solution. After all, that was what the evaluation committee had to evaluate. In this regard it is common cause that the members of the evaluation committee regarded the CPS tender as ‘very poor and largely non-compliant’. Indeed, the recorded view of one of its members, Ms Brenda Faye, a qualified computer technologist, was that CPS’s proposal was ‘abysmal’. In these circumstances, the proposition that, but for the fraud and corruption involved, the tender may have been awarded to CPS, can in our view, be excluded as a matter of near certainty.

The Nisec tender

66. National government also raised the argument that, even if 3D-ID had complied with the specifications, there was another more likely winning candidate. Unlike the province, however, the horse they backed was not CPS but Nisec. When this argument was rather belatedly raised before us for the first time, the reaction was one of surprise, since it had been formally admitted on behalf of the defendants at the trial that, as a fact, the software tendered by Nisec did not comply with the mandatory requirements. It was also common cause, from the outset, that the glowing report Nisec received from the evaluation committee was fraudulently orchestrated by Louw and Scholtz.

67. The argument on behalf of national government, that even if the wrongful conduct of Louw and Scholtz is mentally eliminated and hypothetically replaced by lawful conduct, Nisec would still have won the tender, was founded four-square on the so-called ‘one thousand fingerprint test’ arranged by Louw and Scholtz for the evaluation committee. As it happened, the test was attended only by Terblanche, Louw and Scholtz. Particularly noticeable in their absence were the evaluation committee’s two qualified information technologists. The purpose of the test was to enable Nisec to demonstrate the ability of its software to identify a particular fingerprint against a database of one thousand others. Rabie conceded that it would be unreasonable to insist on a benchmark test of all 175 000 pensioners in the CPA database and that a test involving one thousand fingerprints could thus not be regarded as inappropriate. According to the evidence of Terblanche, the software tendered by Nisec was able to meet the requirements of this test. As we have indicated, the argument that Nisec would have won the award relied entirely on the fact that its software had passed this proficiency test. It matters not, it was argued, that in fact Nisec’s software proved incapable of performing the required functions, because this would have become apparent only after the award.

68. But this ignores the uncontroverted evidence of Bouwer, that it would have been quite readily ascertainable by an expert in the field that the Nisec tender did not comply. What one must postulate, is a reasonably competent and fair evaluation by all the members of the evaluation committee, including its expert members, which would eliminate tenders that were readily identifiable as non-compliant. The fact that these two experts did not attend Nisec’s performance of the one thousand fingerprint test, is telling in itself. In all likelihood it was part of the manipulation orchestrated by Louw and Scholtz. Had these experts been present, as they would have been in a reasonably competent evaluation process, the probabilities indicate that Nisec would have been caught out.

(c) Tender not awarded at all

69. The defendants’ final contention was that, but for the fraud of Louw and Scholtz, the tender would probably not have been awarded at all. This argument was largely based on the evidence of Dr J C Stegmann, a senior employee of the CPA who also served as member of the State Tender Board.

70. The CPA’s request to award the tender to Nisec first came before the State Tender Board on 1 June 1994. It was accompanied by a motivation prepared on behalf of the CPA by Louw. According to the minutes, Stegmann raised a number of concerns with regard to the motivation. The board seemingly adopted these because it informed the CPA that:

‘As a result of the following aspects which were brought to the board’s attention, the board decided not to approve the tender at this stage.’

Then followed the list of Stegmann’s concerns to which the CPA was required to respond.

71. Included amongst these were:

‘(ii) the tender is only in respect of a service to one population group in the current Cape Province and the services to the brown and white population groups are not provided for in the tender;

(iii) it seems that the partitioning of the Cape Province into three provincial governments in the near future has not been taken into consideration and whether the service has been clarified with any provincial government.’

72. The context, as Stegmann explained in evidence, was that before 27 April 1994, the CPA was responsible for payment of social pensions to black beneficiaries only. Pensions of other population groups were managed by the then administrations of the House of Assembly, the House of Representatives and the House of Delegates. Consequently, these were not included in the tender – which Stegmann considered could be regarded as a perpetuation of apartheid. Moreover, he said, the then Cape Province encompassed what after 27 April 1994 became the separate provinces of the Western, Northern and Eastern Cape and his concern was that the tender had not been approved by the latter two provincial governments.

73. The CPA’s response to the two enquiries was prepared by Louw. It read:

‘(ii) Although the tender as published only made provision for the then CPA clients (153 000), tenderers were at the information meeting held on 25 March 1994 requested verbally and in writing to extend the service to all population groups, should it be required by the CPA. This was accepted.’

And:

‘(iii) Although provincial governments are in place for the three new provinces, they presently have no decision making powers. The matter has nevertheless been politically clarified with the Western Cape Provincial Minister of Health and Welfare, Minister Rasool, who in turn clarified the tender with his counterparts in the Eastern and Northern Cape.’

74. Both these answers turned out to be deliberate misrepresentations in furtherance of Louw’s fraudulent manipulation of the process. All that was said about the extension of services at the information meeting of 25 March 1994, which Louw relied on in (ii), was that:

‘tenderers must commit themselves to the extension of this tender should additional pay points and additional clients and pay days be required in future.’

It is common cause that this clearly had nothing to do with the extension to other population groups. The alleged ‘clarification’ of the tender with the provincial governments of the Northern and Eastern Cape, which he relied on in (iii), apparently never happened.

75. At the next meeting of the State Tender Board, held on 14 June 1994, it was formally decided to accept the Nisec tender. It seems, however, that the board was not satisfied with the assurance of alleged (informal) acquiescence by all three provincial governments involved. In consequence, its acceptance was formulated as follows:

‘2. Approval was granted by the Regional Tender Board for the acceptance of the tender from Nisec CC subject to the following conditions:

2.1 That the letter of acceptance only be issued by this office to the successful tenderer, once the premier or the relevant minister of the Western Cape Provincial Government has given his written agreement that the service can be implemented; and

2.2 That the Eastern and Northern Cape Provincial Governments only be incorporated by this office, once the premiers or relevant ministers of these governments have given their written agreements that the service can be implemented in their respective regions.’

76. The next day, 15 June 1994, Mr E Rasool, the then Minister of Health and Welfare in the Western Cape, gave his formal consent to the tender on behalf of his government and on 16 June 1994 the tender was formally awarded to Nisec in respect of the Western Cape region only, with the reservation that services were also to be extended to the Eastern and Northern Cape regions, once ministerial approval by these two governments had been obtained. According to Stegmann’s testimony it became apparent soon thereafter that the services contemplated in the tender could not without more be extended to the other population groups, because the pay-out system for those groups depended on different methods of administration and different data bases. Had this been known to the board, Stegmann testified, the tender would not have been awarded to any entity (irrespective of who complied). In fact, he said, the tender would not even have been invited, essentially because it could be regarded as racially discriminatory.

77. Like so many other arguments of the defendants, these contentions were nowhere to be found in their pleadings or their responses to plaintiff’s requests for pre-trial particulars. Furthermore, they were never put to Rabie or to plaintiff’s expert, Bouwer, who could possibly have commented on whether the various pension payment systems were capable of interfacing with each other or of being integrated into a single system. Moreover, Stegmann’s testimony as to the alleged incompatibility of the systems was not within his personal knowledge and was not confirmed by any person with knowledge of the relevant facts. That Nisec later proved to be unable to extend its services to other population groups is neither here nor there. After all, it soon became patently clear that Nisec was not even able to render its tendered services to the primary target group. In our view the defendants therefore failed to establish the factual basis that is vital to Stegmann’s entire thesis, namely, that it was not technically possible to extend the services to other population groups. What is more, as Stegmann himself acknowledged, his contentions essentially went to whether the tender should have been invited at all. His concerns would thus presumably have been considered and – by inference – rejected by the relevant officials in the provincial and national governments before the tender was invited at all. In any event, it was never the defendants’ case that the tender should not have been invited.

78. It must also be borne in mind that the tender as awarded by the State Tender Board in fact provided services for one population group only. The board did not insist that the award be made conditional on later extension to other population groups or even that it be proven capable of such extension. The minutes of the board meeting seem to suggest that its members regarded the question whether a tender should be awarded despite its differentiation between racial groups, as a political issue that was not for them to decide. That is one of the reasons why great pains were taken to ensure that the tender was not awarded without formal political consent. In effect, the State Tender Board was therefore prepared to award the tender, despite the political risk that accusations of racial discrimination could follow, as long as the politicians were prepared to accept the political risk.

79. Ultimately it is clear that both the CPA and the State Tender Board were desperately keen to award the tender. Enormous pressures were brought to bear upon them to find a solution for the fraud that was rampant with welfare payments, not least because the extent of the fraud had received considerable coverage in the press. Apart from the enormous financial consequences, it therefore also became a political embarrassment. At the same time, the CPA’s own pension program suffered from serious inefficiencies and had all but broken down. Due to a shortage of trained staff and outdated computer equipment, it would soon be unable to perform its pension payment duties. The political consequences of a collapse of these services need hardly be elaborated. The fingerprinting technology tender was presented as the only possible solution to all these problems. An added bonus would be the prestige to the CPA for being first to find a workable solution to what had become a nationwide problem involving losses in hundreds of millions of Rands. It is therefore hardly surprising that Stegmann’s concerns were not shared by the majority of the decision makers within the CPA or the State Tender Board. What is more, it appears that Stegmann himself was only too pleased to be persuaded otherwise. When he was given an opportunity to block the tender award, he did not take it. Instead, he indicated that he was satisfied with very cursory answers to questions that he had posed during the first discussion of the matter on 1 June 1994. It is in fact clear that some of the concerns expressed in the State Tender Board’s letter to the CPA as a result of the discussion, were not addressed at all.

80. Despite the defendants’ arguments to the contrary we are therefore satisfied that, but for the wrongful conduct of Louw and Scholtz, it is more likely than not that 3D-ID, as the only qualifying tenderer, would have received the award, even though its price was substantially higher than all the other tenders. This means that, in our view, the element of causation had been established.

Fourth defence: Wrongfulness – should fraudulent conduct in the tender process be exempt from liability?

81. This brings us to the province’s final contention, namely, that the plaintiff failed to establish a further element of delictual liability, namely wrongfulness. The province invoked the judgments of this court in Olitzki Property Holdings v State Tender Board²¹ and Steenkamp NO v Provincial Tender Board, Eastern Cape,²² arguing that they constitute authority for the general proposition that our law does not extend a delictual claim to an unsuccessful tenderer against a government department for losses suffered in the course of a tender process – including losses inflicted by fraud.

82. But the decisions in Olitzki and Steenkamp must be understood against the well-established principle of our law of delict that negligent conduct causing pure economic loss is not prima facie wrongful. In these circumstances, wrongfulness depends on the existence of a ‘legal duty’. The imposition of such a duty is determined judicially with reference to considerations of public and legal policy, consistent with constitutional norms (see eg Minister of Safety and Security v Van Duivenboden;²³ Gouda Boerdery BK v Transnet;²⁴ Trustees, Two Oceans Aquarium Trust v Kantey & Templer).²⁵ As this court has explained, the imposition of a duty means that the conduct under consideration attracts delictual liability for resulting damages. Conversely, when it is said that the defendant owes the plaintiff no legal duty and that there was thus no wrongfulness, it means that, despite the existence of blameworthy conduct, the defendant enjoys immunity against liability for damages resulting from the conduct.

83. Olitzki decided that the constitutional guarantee of a fair tender system in s 187 of the interim Constitution does not in itself provide the basis for imposing a legal duty to compensate for loss resulting from breach of the guarantee. That case concerned a claim for damages arising from the non-award of a tender resulting from irregular, unreasonable and arbitrary conduct – but fraud was not at issue. In these circumstances it was held that the constitutional injunctions of s 187 did not create a duty to tenderers that on breach could be translated into a claim for damages (paras 25-31).

84. In Steenkamp, where out of pocket expenses incurred because of the negligent award of a tender were at issue, the conclusion was summarised thus (para 46):

‘Weighing up these policy considerations [referred to in paras 24-45] I am satisfied that the existence of an action by tenderers, successful or unsuccessful, for delictual damages that are purely economic in nature and suffered because of a bona fide and negligent failure to comply with the requirements of administrative justice cannot be inferred from the statute in question. Likewise, the same considerations stand in the way of the recognition of a common-law legal duty in these circumstances.’

85. Drawing on these decisions, the province argued that, for the same considerations of policy, this court should refuse to extend Aquilian liability to loss caused by fraud in the tender process. The province conceded that, unlike those cases, the conduct of the defendants’ employees here consisted of deliberate dishonesty and corruption, as opposed to bona fide negligent bungling. However, the province contended that fault and wrongfulness are discrete elements of the Aquilian action – with the consequence that because subjective factors such as the perpetrator’s state of mind and motive pertain to the former element, they are irrelevant in determining the latter. Authority for this proposition was sought in J C van der Walt & J R Midgley, Principles of Delict, 3ed, 71. It followed, the province argued, that Olitzki and Steenkamp applied.

86. But the province’s argument starts from the wrong premise. We do not think that it can be stated as a general rule that, in the context of delictual liability, state of mind has nothing to do with wrongfulness. Clear instances of the contrary are those cases where intent, as opposed to mere negligence, is itself an essential element of wrongfulness. These include intentional interference with contractual rights (see eg Dantex Investment Holdings (Pty) Ltd v Brenner NNO)²⁶ and unlawful competition (see eg Geary & Son v Gove)²⁷. Closer to the mark, in our view, is the following exposition by Boberg, The Law of Delict, Vol 1 (Aquilian Liability) 33, who correctly highlights the significance of the perpetrator’s state of mind in determining wrongfulness:

‘Examination of these crystallized categories of wrongfulness reveals the determining factors. They are: (a) the nature of the defendant’s conduct (was it a positive act or an omission; did it consist of deeds or mere words?); (b) the nature of the defendant’s fault (was the harm suffered by the plaintiff (was it physical harm or mere pecuniary loss?). These criteria do not operate independently but in conjunction with one another. Thus harm of one kind (eg physical) may be actionable whether caused intentionally or negligently, harm of another kind (eg mere pecuniary loss) may be actionable only if caused intentionally (otherwise it is problematical) … . At the root of each of these crystallized categories of wrongfulness lies a value judgment based on considerations of morality and policy – a balancing of interests followed by the law’s decision to protect one kind of interest against one kind of invasion and not another. The decision reflects our society’s prevailing ideas of what is reasonable and proper, what conduct should be condemned and what should not . . . .’

87. In the language of the more recent formulations of the criterion for wrongfulness: in cases of pure economic loss the question will always be whether considerations of public or legal policy dictate that delictual liability should be extended to loss resulting from the conduct at issue. Thus understood, it is hard to think of any reason why the fact that the loss was caused by dishonest (as opposed to bona fide negligent) conduct, should be ignored in deciding the question. We do not say that dishonest conduct will always be wrongful for the purposes of imposing liability, but it is difficult to think of an example where it will not be so.

88. In our view, speaking generally, the fact that a defendant’s conduct was deliberate and dishonest strongly suggests that liability for it should follow in damages, even where a public tender is being awarded. In Olitzki and Steenkamp, the cost to the public purse of imposing liability for lost profit and for out of pocket expenses when officials innocently bungled the process was among the considerations that limited liability. We think the opposite applies where deliberately dishonest conduct is at issue: the cost to the public of exempting a fraudulent perpetrator from liability for fraud would be too high.

89. These considerations would indicate that liability should follow even if the plaintiff’s case were based on dishonesty on the part of the State Tender Board itself. But that is not the case before us, and this constitutes a further problem for the province’s argument. This case does not concern the direct liability of the tender-awarding authority itself: it concerns government’s vicarious liability for its employees’ conduct. The province’s argument is therefore misconceived, since it starts from the wrong premise and therefore inevitably arrives at the wrong conclusion. The plaintiff’s case is that defendants are vicariously liable for the wrongful conduct of Louw and Scholtz. Once we have decided the issue of vicarious liability in favour of the plaintiff, as we have, the only remaining question in the context of wrongfulness is whether Louw and Scholtz, public employees in charge of a tender process, should themselves be exempt from the consequences of their own dishonest conduct. The issue in Olitzki and Steenkamp – whether loss resulting from conduct by the tender-awarding authority itself should be visited with delictual liability – does not arise. For present purposes the question about wrongfulness is no different than if Scholtz and Louw themselves were the defendants.

90. Thus understood the question is: is there any conceivable consideration of public or legal policy that dictates that Louw and Scholtz (and, vicariously, their employer) should enjoy immunity against liability for their fraudulent conduct? We can think of none. The fact that the fraud was committed in the course of a public tender process cannot in our view serve to immunise the wrongdoers (or those vicariously liable for their conduct) from its consequences. And we find no suggestion in Olitzki and Steenkamp that the tender process itself must provide government institutions with a shield that protects them against vicarious liability for the fraudulent conduct of their servants. The wrongfulness issue therefore cannot shield the defendants.

Conclusion and order

91. We conclude that Hartzenberg J correctly determined the issues before him in favour of the plaintiff and against the defendants. The defendants’ prospects on the merits of the appeal are therefore insubstantial and for this reason they should be refused condonation for the late filing of the appeal. There is a matter pertaining to the form of the order granted in the court below which by agreement between the parties we rectify.

1. Condonation is refused with costs, including the costs of two counsel.

2. By agreement between the parties, paragraph 5 of the order of the court below is substituted with the following:

‘The second defendant and the fourth defendant are jointly and severally liable to pay such damages as the plaintiff may prove.’

E CAMERON AND FDJ BRAND

JUDGES OF APPEAL

CONCUR:

MTHIYANE JA

MLAMBO JA

MALAN AJA