Rustenburg Platinum Mines Ltd (Rustenburg Section) v Commission for Conciliation, Mediation and Arbitration and Others (598/2005) [2006] ZASCA 175 (26 September 2006)

THE SUPREME COURT OF APPEAL

OF SOUTH AFRICA

1. Case no: 598/05

2. REPORTABLE

In the appeal between:

RUSTENBURG PLATINUM MINES LTD (RUSTENBURG SECTION) Appellant

and

COMMISSION FOR CONCILIATION, MEDIATION AND ARBITRATION First Respondent

T J MOROPA NO Second Respondent

Z SIDUMO Third Respondent

Before: Harms JA, Cameron JA, Cloete JA, Lewis JA, Maya JA

Heard: Monday 11 September 2006

Judgment: Tuesday 26 September 2006

Labour Relations Act 66 of 1995 – Commission for Conciliation, Mediation and Arbitration – powers of arbitrators regarding dismissals – decision on workplace remedy lies primarily with employer – arbitrators must exercise caution before interfering – Labour Appeal Court – proper test for review of CCMA arbitrators’ decisions restated

Neutral citation: Rustenburg Platinum Mines Ltd v CCMA [2006] SCA 115 (RSA)

JUDGMENT

_______________________________________________________

CAMERON JA:

Introduction

1. In June 2000 the appellant (the mine) dismissed the third respondent, Mr Zingisile Sidumo, from his job as a grade II patrolman in its protection services department. Sidumo successfully challenged his dismissal under the compulsory arbitration provisions of the Labour Relations Act 66 of 1995 (the LRA) administered by the first respondent, the Commission for Conciliation, Mediation and Arbitration (the CCMA). The second respondent, Mr T Moropa, a CCMA commissioner, found the dismissal procedurally fair, but substantively unfair. He reinstated Sidumo with three months’ compensation, subject to a written warning valid for six months. The Labour Court and the Labour Appeal Court declined to intervene on review. With special leave granted by this court, the mine now appeals.

Internal disciplinary hearing and appeal

2. Sidumo started with the mine in December 1985. At the time of his dismissal he had served for more than fourteen years without disciplinary infraction. He was employed at the Waterval redressing section, a high security facility that provided benefaction services, separating high grade precious metals from lower grade concentrate. His main duty was access control and to protect the mine’s precious metal product. Because of a high theft problem, which was causing significant daily losses, the mine instituted detailed search procedures for all persons leaving the Waterval plant. This entailed an individual private search of each person in a cubicle, with close physical inspection, plus a metal detector scan. The compulsory procedures were distributed and made known to all, including Sidumo, who in August 1999 signed an acknowledgement that they had been read and explained to him.

3. Losses continued. In response, the mine mounted video surveillance of employee performance at various points, including that of Sidumo over three separate days while he was on duty in April 2000. This revealed that, of 24 searches in the three-day period, he conducted only one properly in accordance with obligatory search procedures, which required him ‘to search everyone leaving the redressing section according to the search procedure which is displayed’.¹ On eight persons he conducted no search at all. On fifteen the search was ‘improper’ (ie, not in accordance with the works instruction). The video revealed that Sidumo allowed some persons to sign the search register without conducting any search at all.

4. Sidumo was charged with ‘offences’ at an internal disciplinary inquiry. A senior superintendent, Mr Page, found him guilty of misconduct in the form of negligence and failure to follow established search procedures. He found that Sidumo was an experienced patrolman who had been placed in a high risk area to safeguard the company’s most valuable product at the redressing section. The misconduct had ‘created potential production losses / theft’ at the redressing section. In mitigation, Page accepted that ‘nothing went out during your shift, as far as you know’, and took into account that Sidumo had a clean disciplinary record and nearly 15 years’ service. Despite this, he found that the misconduct went to the heart of Sidumo’s capacity as a protection services member, and that ‘the trustworthy position between [him] and the company has been broken, which made [a] future relationship intolerable’.

5. Sidumo appealed. The supervisor presiding over the internal appeal, Mr Denner, emphasised that since Sidumo had not been charged with dishonesty, the fact that no actual losses were proved was irrelevant: the charge was negligently failing to follow established procedures. ‘An absolute fact is [that] through your wrongdoings, the company could or may have sustained losses by means of theft (of the prime product)’, ‘which we know impacts financially on the viability of the company’. It was because of his seniority that Sidumo had been employed in a position of trust, ‘which you abused through your negligence and non-adherence to work instructions’. ‘I have considered alternatives to the dismissal’, the supervisor concluded, ‘but found none to be appropriate’.

6. In view of the later proceedings, the findings of both internal tribunals that (a) Sidumo was guilty of negligence and (b) that though no actual losses had been proved, potential losses or theft may have occurred on his shifts, are significant.

CCMA hearing and determination

7. Sidumo then referred his dismissal to the CCMA, applying for reinstatement or compensation. Under the LRA,² a commissioner is empowered to arbitrate a dismissal dispute in which the employee has alleged that the reason for the dismissal is related to his or her conduct or capacity (s 191(5)(a)). The employee bears the burden of proving that he or she was dismissed (s 192(1)). Once that is proved, the employer bears the burden of proving that the dismissal was fair (s 192(2)). A dismissal is unfair if the employer fails to prove, on balance, that the reason for the dismissal is a fair reason relating to the employee’s conduct or capacity (s 188(1)(a)(i)). A commissioner, in considering whether or not the reason is a fair reason, must take into account any relevant code of good practice issued under the LRA (s 138(6) and s 188(2)). The ‘Code of Good Practice: Dismissal’ is such a code (set out in the footnote to para 17 below). A dismissal is unfair if it is not effected for ‘a fair reason’. This is determined by the facts of the case and the appropriateness of dismissal as a penalty. The LRA permits the commissioner to conduct the arbitration in a manner the commissioner considers appropriate in order to determine the dispute fairly and quickly, but he must deal with the substantial merits of the dispute with the minimum of legal formalities (s 138(1)). The commissioner may make any appropriate award in terms of the LRA, including an award that gives effect to the provisions and primary objects of the statute (s 138(9)). The express purpose of the LRA, as set out in s 1, is to advance economic development, social justice, labour peace and democratisation of the workplace by fulfilling the statute’s primary objects. These include promoting ‘the effective resolution of labour disputes’ (s 1(d)(iv)).

8. At the CCMA hearing the mine’s assistant chief chemist, Mr Williams, explained why the redressing station, where ore was extracted from platinum group metals (platinum, vanadium, rhodium, iridium, ruthenium), was a ‘high risk area’. The mine’s losses could not be attributed to ore quality or equipment problems and therefore had to be due to theft. Surveillance equipment was accordingly installed, which resulted in disciplinary action. This restored production from September 2000.

9. One of Sidumo’s superiors in protection services, Mr Botes, who conducted the surveillance, and laid the charge against him, explained that the prevention of theft was Sidumo’s main responsibility: because of the high value of the product, everyone leaving the redressing section had to be searched. During the surveillance period (though not on Sidumo’s watch) one thief had for instance been caught with product worth R44 000. Sidumo was in what Botes called a ‘high integrity position of high trust’. He insisted that Sidumo was well aware of how to conduct searches – this was his main duty, which he was required to perform all the time. The surveillance revealed that in cases where he performed no search at all, ‘the person just walked out’. Although the work should normally have been performed by a senior patrolman, Sidumo had been posted there because of his experience. Botes conceded that the mine’s disciplinary code entailed that ‘discipline and corrective measures are put in place to put [the employee] on the right track again’, but pointed out that the decision regarding dismissal was not his to make.

10. Testifying before the commissioner, Sidumo denied that he had received training, and claimed to have objected to being posted to the security point. The commissioner’s findings implicitly rejected both these defences, since he held that Sidumo was ‘clearly’ guilty of misconduct. Because of inaudible passages, the transcript is not wholly clear, but Sidumo appears to have admitted that he knew that compulsory searching was required at the redressing plant.

11. The commissioner found the procedure the mine followed in dismissing Sidumo to be fair, but held that dismissal was inappropriate. His reasons were: (a) the mine had suffered no losses; (b) the violation of the rule was ‘unintentional or “a mistake” as argued by the employee’; (c) the ‘level of honesty of the employee is something to consider’; and (d) ‘the type of offence committed by the employee does not go into the heart of [the] relationship, which is trust’.

Review proceedings: Labour Court and Labour Appeal Court

12. The mine applied to the labour court to review the commissioner’s award on the basis that there was direct and largely unchallenged evidence that the mine’s yield had been low since May 1998; that over the surveillance period February to May 2000, the metallic yield created a revenue loss of R500 000 per day; that precious metals were discovered on persons during the surveillance; that the mine had experienced theft over the previous 3 to 4 years; that Sidumo was employed to prevent theft and had conducted only a single proper search while under surveillance. The mine therefore contended that the award was not justifiable in relation to the reasons given for it, in that no rational link existed between the evidence before the commissioner and the factual conclusions that were crucial to his award. The finding that the misconduct did not go to the heart of the relationship was irrational. The commissioner had therefore been so grossly careless as to have committed misconduct; his failure to apply his mind meant that no fair hearing had occurred and that a gross irregularity had been committed, while the absence of rational connection entailed that he had exceeded his powers.

13. In his affidavit opposing the mine’s application, Sidumo elaborated the defences the commissioner had rejected. He claimed that as a grade 2 patrolman he was ‘expected to search randomly not to search as it is alleged’, that he had ‘no knowledge of how a search was done at the said redressing station’ and that ‘according to the rules of the [mine] I was not supposed to have been posted there at all’. He also claimed ‘that search according to my grade was not compulsory hence I did not conduct it in the manner expected at Redressing Station’.

14. These claims are at variance with the commissioner’s findings that it was undisputed that Sidumo ‘knew how to conduct a search and he agreed to have signed a document to affirm his knowledge of the procedure to be followed in terms of the company policy’; that Sidumo had ‘contravened the rule or standard regulating conduct in or of relevance to the workplace’ and that by his ‘own admission’ while on duty during the surveillance period he ‘was fully responsible [for] access control’; and that his failure ‘to perform a full search’ constituted misconduct.

15. The Labour Court declined the mine’s review application. Revelas J considered it a case of poor performance rather than misconduct. She found that the employee had a clean service record of almost 15 years, did not commit a violent crime or assault, did not steal, and did not commit an offence that ‘unequivocally demanded dismissal as opposed to any other sanction’. She asked whether the commissioner’s preference for ‘corrective or progressive discipline’ for Sidumo induced ‘a sense of shock’, and concluded No. At best for the mine, she held, there was poor performance or laziness, which was ‘not the type of misconduct which justifies dismissal without prior warning for a first offence after 15 years of service’. There was not ‘an iota of evidence’ that theft had occurred during Sidumo’s shifts. That Sidumo was doing work a more senior employee usually performed was also ‘very significant’. In the absence of dishonesty, employees who do not perform their duties properly ‘should not automatically incur the harsh sanction of dismissal on the basis of strict liability, even if they work in a gold mine’.

16. The LAC (Zondo JP, Mogoeng JA and Comrie AJA concurring) was more critical of the commissioner’s approach. It expressly rejected three of his findings. (a) The finding that the mine suffered no losses as a result of Sidumo’s malperformance had no basis and was ‘indeed wholly unjustifiable’. (b) It was unclear what the commissioner meant by saying that Sidumo’s conduct was a ‘mistake’ or ‘unintentional’: this might have referred to the fact that the charge related to negligence, not intentional conduct – but even in that case, this was not decisive but ‘would have had to be taken into account in the light of all the circumstances’. (c) Lastly, Zondo JP remarked, ‘Quite frankly, how the third factor, namely dishonesty, came into the picture at all, is baffling. No dishonesty by [Sidumo] was alleged.’ The LAC made mention of the commissioner’s fourth finding, namely (d) that the offence ‘does not go into the heart of [the] relationship, which is trust’, but made no express finding regarding the commissioner’s reliance on it (although the LAC’s observation that the misconduct was ‘indeed serious’ may mean it rejected this ground).

17. Despite finding at least three of the commissioner’s grounds for reinstating Sidumo wanting, the LAC declined to intervene. It held that had the three bad reasons been the sole basis of the award, it would have been unjustifiable. But there were other reasons – the commissioner also relied on the Code of Good Practice (Schedule 8 to the LRA),³ which provides that it is not appropriate to dismiss an employee for a first offence unless the misconduct is serious and of such gravity as to make a continued employment relationship intolerable. The commissioner had further relied on Sidumo’s clean record over 14 years and suggested that ‘graduated disciplinary measures such as counselling and warning’ would be appropriate. The LAC considered that the mine’s review papers had failed to challenge the commissioner’s reliance on these factors. For the mine to contend for the first time in argument that long service was irrelevant to the breach of a core function was impermissible. The LAC concluded:

‘That [Sidumo] had a clean record and a long service period is capable of sustaining the finding by the commissioner that the sanction of dismissal was too harsh. Whether or not it would have been enough to sustain the finding had it been challenged in the founding affidavit is another matter. However, I must say that, although the misconduct of [Sidumo] is indeed serious, I am not sure that I would not have been in doubt about whether I should interfere with the finding of the [commissioner]. And in case of doubt, the court should not interfere.’

The test for review of CCMA arbitrations

18. Section 145(1) of the LRA provides for ‘review of arbitration awards’ by the labour court on ground of ‘a defect’. Subject to the court’s power to grant condonation (s145(1A)), the application must be brought within six weeks of the award (or, in cases of corruption, six weeks from discovery of the offence). In terms of s 145(2), ‘defect’ means –

‘(a) that the commissioner –

(i) committed misconduct in relation to the duties of the commissioner as an arbitrator;

(ii) committed a gross irregularity in the conduct of the arbitration proceedings; or

(iii) exceeded the commissioner’s powers; or

(b) that an award has been improperly obtained’.

Until 2002, s 158(1)(g) empowered the labour court, ‘despite s 145’, to review the performance of any function provided for in the LRA ‘on any grounds that are permissible in law’.⁴ In 2002, ‘despite’ was replaced with ‘subject to’.

19. The LRA received assent on 29 November 1995 and came into operation on 11 November 1996. On both those dates, the interim Constitution (Act 200 of 1993) was in force, which provided that the fundamental right to administrative justice (s 24) entitled every person to –

‘(d) administrative action which is justifiable in relation to the reasons given for it where any of his or her rights is affected or threatened’.

20. In Carephone (Pty) Ltd v Marcus NO,⁵ the LAC reconciled the provisions of s 145(2) and s 158(1)(g) with each other, and with the administrative justice provisions of the interim Constitution, by holding that the test for review of a CCMA arbitrator’s decision went beyond mere questions of procedural impropriety, or irrationality only as evidence of procedural impropriety: the question was whether there was ‘a rational objective basis justifying the connection’ the commissioner made between the material properly available and the decision (paras 31 and 37).

21. Despite some initial dissent,⁶ the LAC accepted after the decision of the Constitutional Court in Pharmaceutical Manufacturers of SA: In re Ex parte President of the Republic of South Africa,⁷ and after the enactment of the Promotion of Administrative Justice Act 3 of 2000 (PAJA), that the Carephone test was applicable to the review of CCMA decisions. PAJA, which enacted grounds of review considerably more extensive than those set out in s 145(2) of the LRA,⁸ came into force on 30 November 2000. Of present moment is s 6(2)(f)(ii), which empowers a court to review an administrative action if the action itself is ‘not rationally connected to’ –

‘(cc) the information before the administrator; or

(dd) the reasons given for it by the administrator’.

22. In Shoprite Checkers (Pty) Ltd v Ramdaw NO⁹ the LAC considered the possible effect of PAJA’s enactment on s 145(2). The LAC accepted ‘the possibility that the PAJA may well be applicable to arbitration awards issued by the CCMA’ (para 33), but found it unnecessary to decide the issue.

23. In my view, PAJA by necessary implication extended the grounds of review available to parties to CCMA arbitrations. In interpreting the LRA, and the impact on it of the later enactment of PAJA, the Constitution obliges us to promote the spirit, purport and objects of the Bill of Rights.¹⁰ This means that, without losing sight of the specific constitutional objectives of the LRA, and the constitutional values it embodies, we must in interpreting it give appropriate recognition to the right to administrative justice under the final Constitution and the legislation that gives effect to it.

24. It follows that the overriding factor in determining the impact of PAJA on the LRA is the constitutional setting in which PAJA was enacted. Parliament enacted PAJA because of a constitutional obligation to give legislative effect to the right to just administrative action embodied in the final Constitution.¹¹ That obligation did not exempt from its ambit previous parliamentary enactments that conferred rights of administrative review. It extended to all of them. This is so even though the LRA is a specialised statute. According to its preamble, it was enacted to give effect to s 27 of the interim Constitution (though unlike PAJA it is not the product of an express imperative obliging Parliament to legislate). It did so by regulating various matters within the labour relations field. Both the Constitution, which required Parliament to give general legislative effect to the right to administrative justice, and the legislation so enacted, superseded the LRA’s specialised enactment within the field.

25. As the Constitutional Court has stated,¹² the provisions of s 6 reveal a clear purpose to codify the grounds of review of administrative action as defined in PAJA. The Constitution required PAJA to cover the field, and it purports to do so.¹³ Notable in this respect is that ‘court’ in terms of s 1 of PAJA includes ‘a High Court or another court of similar status’ – which plainly encompasses the labour courts. There can be no doubt that a CCMA commissioner’s arbitral decision constitutes administrative action. In my view, s 6’s codificatory purpose subsumed the grounds of review in s 145(2), and PAJA’s constitutional purpose must be taken to override that provision’s preceding, more constricted, formulation.

26. A slightly different path leads to the same conclusion. At the time the LRA was enacted, the interim Constitution required that administrative action be ‘justifiable in relation to the reasons given for it’. For the reasons set out in Carephone, this right suffused the interpretation of s 145(2). When the administrative justice provisions of the Constitution, as embodied in PAJA, superseded those of the interim Constitution, it could not have been intended that parties to CCMA arbitrations should enjoy a lesser right of administrative review than that afforded under the interim Constitution. The repeal of the interim Constitution and its replacement by the Constitution in other words did not diminish the review entitlement under s 145(2). Section 6(2) of PAJA is the legislative embodiment of the grounds of review to which arbitration parties became entitled under the Constitution.

27. The extension of the grounds of review does not impinge on the time periods in s 145(1). PAJA requires that proceedings for judicial review be instituted without unreasonable delay and in any event not later than 180 days after exhaustion of internal remedies or after the person concerned became aware of the action challenged and the reasons for it (s 7(1)). That is a longer period than the six weeks s 145(1) affords. However, as both the CC¹⁴ and this court¹⁵ have emphasised, labour disputes require speedy resolution, and the legislature gave clear effect to this special imperative in s 145(1) by requiring a labour disputant to act quickly. The Constitution does not require that the legislation enacted to give effect to the right to administrative justice must embody any particular time periods. This is therefore a question on which the legislature may be expected to legislate differently in different fields, taking into account particular needs.¹⁶

The LAC’s approach to the test for review

28. It will be recalled that the LAC faulted the mine for not challenging the commissioner’s reliance on the Code of Good Practice and on Sidumo’s clean record and long service in its founding affidavit, but only in argument. This was incorrect. The mine considered from the outset that Sidumo’s clean record and long service were indeed relevant: both internal decisions carefully note their significance. The mine’s case was that despite these factors the continued employment relationship was intolerable under the Code. The mine’s complaint was therefore not that the commissioner took relevant factors into account, but that his decision was tainted by reliance on misconceived considerations. As Mr Gauntlett pointed out in argument, the mine could hardly be criticised for not complaining about reliance on factors that it considered legitimate itself. The mine’s true complaint – which in the nature of things its founding papers could not foresee, but which could be raised only in argument – was that the labour court and the LAC held that those factors insulated the commissioner’s decision from their intervention.

29. For what both Carephone and PAJA required the LAC to do was to consider whether the commissioner’s decision to reinstate Sidumo was ‘rationally connected’ to the information before him and to the reasons he gave for it. ‘Rational connection’ requires, as Froneman DJP explained in Carephone (para 37), in a passage this court approved and applied in the light of PAJA,¹⁷ that there must be a rational objective basis justifying the connection the commissioner made between the material before him and the conclusion he reached.

30. The LAC did not apply this test. Nor did it refer to Carephone, or indeed to PAJA. Instead, it asked whether considerations existed, which the commissioner had taken into account, that were ‘capable of sustaining’ his finding. In effect, the LAC asked whether there was material on record that could support the view that, despite his errors, the commissioner had nevertheless ‘got it right’. In so approaching the matter, the LAC treated the mine’s challenge to the decision as an appeal. In my respectful view, this was incorrect. The question on review is not whether the record reveals relevant considerations that are capable of justifying the outcome. That test applies when a court hears an appeal: then the inquiry is whether the record contains material showing that the decision – notwithstanding any errors of reasoning – was correct. This is because in an appeal, the only determination is whether the decision is right or wrong.¹⁸

31. In a review, the question is not whether the decision is capable of being justified (or, as the LAC thought, whether it is not so incorrect as to make intervention doubtful), but whether the decision-maker properly exercised the powers entrusted to him or her. The focus is on the process, and on the way in which the decision-maker came to the challenged conclusion. This is not to lose sight of the fact that the line between review and appeal is notoriously difficult to draw. This is partly because process-related scrutiny can never blind itself to the substantive merits of the outcome. Indeed, under PAJA the merits to some extent always intrude, since the court must examine the connection between the decision and the reasons the decision-maker gives for it, and determine whether the connection is rational. That task can never be performed without taking some account of the substantive merits of the decision.

32. But this does not mean that PAJA obliterates the distinction between review and appeal. As this Court has observed:

‘In requiring reasonable administrative action, the Constitution does not … intend that such action must, in review proceedings, be tested against the reasonableness of the merits of the action in the same way as an appeal. In other words, it is not required that the action must be substantively reasonable, in that sense, in order to withstand review. Apart from that being too high a threshold, it would mean that all administrative action would be liable to correction on review if objectively assessed as substantively unreasonable …’¹⁹

In Carephone, Froneman DJP explained that in determining whether administrative action is justifiable in terms of the reasons given for it (or, in PAJA’s formulation, whether the connection made is ‘rational’) –

‘value judgments will have to be made which will, almost inevitably, involve the consideration of the ‘merits’ of the matter in some way or another. As long as the Judge determining this issue is aware that he or she enters the merits not in order to substitute his or her own opinion on the correctness thereof, but to determine whether the outcome is rationally justifiable, the process will be in order.’²⁰

Application of process-review under PAJA

33. In my respectful view, the LAC lost from sight that it was required to subject the commissioner’s decision to process-related scrutiny, and not to inquire whether his decision was capable of being sustained by recourse to other factors emerging from the record. Turning to that task, it is evident that apart from the employee’s clean record and long service, the commissioner took four considerations into account. Three of these (absence of loss; ‘mistake’; no dishonesty) the LAC rightly rejected as bad. The fourth – that the misconduct did not go to the heart of the relationship, ‘which is trust’ – was in my view also incorrect. This is for two reasons. First, the failure to search represented not a peripheral malperformance, but a profound failure at the very core of the employee’s job functions. Second, the employer trusted him to carry out searches unsupervised while he was on watch: his failure to do so necessarily violated that trust.

34. Given that the commissioner took four bad reasons into account in reinstating the employee, but that other legitimate reasons existed that were capable of sustaining the outcome, can it be said that the employee’s reinstatement was ‘rationally connected’ to the information before the commissioner, or the reasons given for it, as PAJA requires? In my view, it can not. It can certainly not be said that the outcome was ‘rationally connected’ to the commissioner’s reasons as a whole, for those reasons were preponderantly bad, and bad reasons cannot provide a rational connection to a sustainable outcome. Nor does PAJA oblige us to pick and choose between the commissioner’s reasons to try to find sustenance for the decision despite the bad reasons. Once the bad reasons played an appreciable or significant role in the outcome, it is in my view impossible to say that the reasons given provide a rational connection to it. This dimension of rationality in decision-making predates its constitutional formulation. In Patel v Witbank Town Council,²¹ Tindall J set aside a decision which had been ‘substantially influenced’ by a bad reason. He asked:

‘[W]hat is the effect upon the refusal of holding that, while it has not been shown that grounds 1, 2, 4 and 5 are assailable, it has been shown that ground 3 is a bad ground for a refusal? Now it seems to me, if I am correct in holding that ground 3 put forward by the council is bad, that the result is that the whole decision goes by the board; for this is not a ground of no importance, it is a ground which substantially influenced the council in its decision. …

This ground having substantially influenced the decision of the committee, it follows that the committee allowed its decision to be influenced by a consideration which ought not to have weighed with it.’

The same applies where it is impossible to distinguish between the reasons that substantially influenced the decision, and those that did not.²²

35. For these reasons, the commissioner’s determination should have been set aside.

The LAC’s oversight over CCMA commissioners’ determinations

36. There is a further reason why the commissioner’s decision was vulnerable to review. In my view, the commissioner failed to appreciate the ambit of his duties under the statute, and therefore incorrectly approached the task entrusted to him of determining whether the employee’s dismissal was fair. Because of the general importance of this point, its relation to the LAC’s oversight of the CCMA as evidenced in the judgment requires emphasis.

37. As already observed (para 29 above), the test the LAC applied in scrutinising the commissioner’s decision was more appropriate to the hearing of an appeal than to the task of review. The effect of this approach has been to give CCMA commissioners more leeway than the statute allows, since it insulates their decisions from intervention unless the record is devoid of reasons that are capable of justifying the outcome. Other tests propounded in the LAC have gone even further in this direction. One suggests that intervention is appropriate only where there is a ‘yawning chasm’ between the sanction the court would have imposed and that of the commissioner, or where the sanction ‘is so egregious that it shocks and alarms the court’.²³ This narrows the ground for intervention even further, since it likens the proceedings to an appeal against sentence: and the scope for intervention in such proceedings is notoriously narrow.

38. It is hard not to feel considerable appreciation for the LAC’s dilemma in attempting to find a constrictive test for review. The LAC is a specialist tribunal with experience in and knowledge of the labour field, and is particularly aware of the danger that the labour courts could be flooded with review applications from importunate parties who resent CCMA determinations.²⁴ But it seems to me that the LAC sought to tackle the problem at the wrong end. The answer to the spectre of a flood of disgruntled litigants lies not in an unduly constricted or misfashioned standard of review, but in directing CCMA commissioners more closely to the proper scope of their powers and duties under the LRA.

39. For the LRA embodied a historic compromise between labour and employers, both being represented by experts on the drafting committee that produced it. And the statute’s formulation of the employer’s powers, and those of the CCMA in overseeing their exercise, reflected the careful balance that compromise entailed. It is therefore vital that the LRA’s wording should be given proper effect.

40. With great respect to the LAC, I do not think the review tests it sought to fashion are either statutorily warranted or constitutionally sound. A sentence in a criminal case is insulated against intervention because its imposition involves the exercise of a discretion entrusted to the judicial officer, which is not readily overturned. By contrast, a CCMA commissioner is not vested with a discretion to impose a sanction in the case of workplace incapacity or misconduct. That discretion belongs in the first instance to the employer. The commissioner enjoys no discretion in relation to sanction, but bears the duty of determining whether the employer’s sanction is fair. This was clearly explained by Ngcobo JA in the LAC in Nampak Corrugated Wadeville v Khoza:²⁵

‘The determination of an appropriate sanction is a matter which is largely within the discretion of the employer. However, this discretion must be exercised fairly. A court should, therefore, not lightly interfere with the sanction imposed by the employer unless the employer acted unfairly in imposing the sanction. The question is not whether the court would have imposed the sanction imposed by the employer, but whether in the circumstances of the case the sanction was reasonable.’

41. This statement of the law elicited a spirited debate within the LAC, after Ngcobo JA’s departure from it, which it is unnecessary to review here,²⁶ about the applicability of the ‘reasonable employer’ test. It suffices to say that the key elements of Ngcobo JA’s approach are: (a) the discretion to dismiss lies primarily with the employer; (b) the discretion must be exercised fairly; and (c) interference should not lightly be contemplated.

42. That is indeed what the statute requires, and in County Fair Foods (Pty) Ltd v CCMA,²⁷ Ngcobo AJP returned to the fairness criterion and to the just ambit of employer discretion. He now emphasised (d) that commissioners should use their powers to intervene with ‘caution’, and (e) that they must afford the sanction imposed by the employer ‘a measure of deference’. The rationale is that the duty of imposing the workplace sanction rests primarily with the employer:

‘Given the finality of the awards and the limited power of the Labour Court to interfere with the awards, commissioners must approach their functions with caution. They must bear in mind that their awards are final – there is no appeal against their awards. In particular, commissioners must exercise greater caution when they consider the fairness of the sanction imposed by the employer. They should not interfere with the sanction merely because they do not like it. There must be a measure of deference to the sanction imposed by the employer subject to the requirement that the sanction imposed by the employer must be fair. The rationale for this is that it is primarily the function of the employer to decide upon the proper sanction.

…

The mere fact that the commissioner may have imposed a somewhat different sanction or a somewhat more severe sanction than the employer would have, is no justification for interference by the commissioner.

…

… In my view, interference with the sanction imposed by the employer is only justified where the sanction is unfair or where the employer acted unfairly in imposing the sanction. This would be the case, for example, where the sanction is so excessive as to shock one’s sense of fairness. In such a case, the commissioner has a duty to interfere.’

43. This analysis is firmly rooted in the prescripts of the statute, and affords an approach to the duties of commissioners that is not only fair and practicable, but would also shield the labour courts from the very flood of litigation the alternative tests have mistakenly been designed to avoid. It is in my view regrettable that the LAC has not consistently affirmed and applied the analysis. Although some panels have affirmed Ngcobo AJP’s approach,²⁸ this case indicates how far the practice of the LAC has on occasion strayed from it. Instead of insisting that under the LRA the discretion to impose the sanction lies primarily with the employer, to be overturned only with caution, the approach evidenced in the present case appears to have upended the due order and conferred the discretion instead on the commissioner. Instead of exhorting commissioners to exercise greater caution when intervening, and to show a measure of deference to the employer’s sanction, so long as it is fair, it has insulated commissioners’ decisions from intervention by importing unduly constrictive criteria into the review process.

44. In view of this, it is worth emphasising some of the reasons that underlay the analysis of Ngcobo AJP. One is textual; another conceptual; and a third institutional.

45. Textual: Section 188(2) of the LRA requires any person considering whether or not the reason for dismissal is ‘a fair reason’ to take into account the Code of Good Practice on dismissal (Schedule 8 to the LRA, set out in the footnote to para 16 above). The Code is significant in locating the first-line responsibility for workplace discipline and sanction with the employer. I say this for three reasons:

(i) Item 7(b)(iv) of the Code requires the commissioner to consider whether dismissal was ‘an’ appropriate sanction for the contravention. It does not require the commissioner to consider whether it was ‘the’ appropriate sanction. The use of the indefinite, as opposed to the definite, article is important.²⁹ It shows the legislature’s awareness that more than one sanction could be considered ‘fair’ for the contravention.

(ii) The benchmark the Code repeatedly sets is whether the sanction is ‘appropriate’. This requires the sanction to be suitable or proper. As Myburgh and van Niekerk observe, ‘The benchmark of appropriateness necessarily implies a range of responses.’³⁰

(iii) The Code states that generally it is not appropriate to dismiss for a first offence unless the misconduct is serious and of such gravity that it makes a continued employment relationship ‘intolerable’. ‘Intolerable’ means ‘unable to be endured’ (Concise Oxford Dictionary). This necessarily imports a measure of subjective perception and assessment, since the capacity to endure a continued employment relationship must exist on the part of the employer. It follows that the primary assessment of intolerability unavoidably belongs to the employer. This is not to confer a subjective say-so. Allowing some leeway to the employer’s primacy of response does not permit caprice or arbitrariness. A mere assertion on implausible grounds that a continued relationship is intolerable will not be sufficient. The criterion remains whether the dismissal was fair.

46. Conceptual: The text of the Code has its roots in the inherent malleability of the criterion it enshrines, namely fairness, which is not an absolute concept. The criterion of fairness denotes a range of possible responses, all of which could properly be described as fair. The use of ‘fairness’ in everyday language reflects this. We may describe a decision as ‘very fair’ (when we mean that it was generous to the offender); or ‘more than fair’ (when we mean that it was lenient); or we may say that it was ‘tough, but fair’, or even ‘severe, but fair’ (meaning that while one’s own decisional response might have been different, it is not possible to brand the actual response unfair). It is in this latter category, particularly, that CCMA commissioners must exercise great caution in evaluating decisions to dismiss. The mere fact that a CCMA commissioner may have imposed a different sanction does not justify concluding that the sanction was unfair. Commissioners must bear in mind that fairness is a relative concept, and that employers should be permitted leeway in determining a fair sanction. As Myburgh and van Niekerk suggest:

‘The first step in the reasoning process of the commissioner should be to recognise that, within limits, the employer is entitled to set its own standards of conduct in the workplace having regard to the exigencies of the business. That much is trite. The employer is entitled to set the standard and to determine the sanction with which non-compliance with the standard will be visited.’³¹

Todd and Damant explain:

‘The court must necessarily recognise that there may be a range of possible decisions that the employer may take, some of which may be fair and some of which may be unfair. The court’s duty is to determine whether the decision that the employer took falls within the range of decisions that may properly be described as being fair.’³²

This passage equally describes the duty of a commissioner. It follows that in determining the fairness of a dismissal, a commissioner need not be persuaded that dismissal is the only fair sanction. The statute requires only that the employer establish that it is a fair sanction. The fact that the commissioner may think that a different sanction would also be fair, or fairer, or even more than fair, does not justify setting aside the employer’s sanction.

47. Institutional: As the approach of Ngcobo AJP implies, the solution to the flood of cases the LAC understandably fears does not lie in unduly constricting the grounds of review to permit. It lies in pointing commissioners firmly to the limits the statute places upon their power to intervene:

‘If commissioners could substitute their judgment and discretion for the judgment and discretion fairly exercised by the employers, then the function of management would have been abdicated – employees would take every case to the CCMA. This result would not be fair to employers.’³³

Summary

48. To summarise:

(a) PAJA applies in the review of the decisions of CCMA commissioners.

(b) The review criterion relevant to this case is whether the decision is rationally connected with the information before the commissioner and with the reasons given for it.

(c) In applying this criterion the question is whether there is a rational objective basis justifying the connection the commissioner made between the available material and the conclusion.

(d) Commissioners must exercise caution when determining whether a workplace sanction imposed by an employer is fair. There must be a measure of deference to the employer’s sanction, because under the LRA it is primarily the function of the employer to decide on the proper sanction.

(e) In determining whether a dismissal is fair, a commissioner need not be persuaded that dismissal is the only fair sanction. The statute requires only that the employer establish that it is a fair sanction. The fact that the commissioner may think that a different sanction would also be fair does not justify setting aside the employer’s sanction.

Conclusion: substitution of commissioner’s determination

49. It follows that the commissioner’s determination cannot stand and must be set aside. This conclusion makes it unnecessary to consider the mine’s further argument, that Sidumo’s absence of remorse, and his conduct in untruthfully denying the ambit of his duties before the commissioner, and in persisting in that defence in his affidavits in the review, itself rendered his continued employment intolerable. There is long-standing LAC authority for this proposition,³⁴ but the present case does not require that it be considered.

50. The parties were agreed, should the commissioner’s determination be set aside, because of the long time that has passed since the dismissal (now more than six years), and the fact that this court has before it all the necessary information, and that the issues have been fully traversed, that it would not be in the interests of justice for the matter to be referred back to the CCMA for arbitration afresh. Both parties instead asked that in view of the circumstances this court substitute its own decision.

51. Addressing, then, the question the commissioner had before him, it is necessary to consider the gravity of the misconduct (which exposed the mine to the risk of serious loss) in the light of the employee’s unblemished fourteen-year service record. Here the factors set out in para 33 above are relevant: the employee’s misconduct went to the heart of the employment relationship and violated the trust the employer placed in him. What is more, the failure to search unsupervised – which constituted his core duty – was sustained over three shifts. Though the sanction of dismissal is undoubtedly severe, especially in its effects on the employee, it is in my view impossible to say that it is not a fair sanction. It certainly seems to me to fall within the range of sanctions that the employer was fairly permitted to impose. The employee should therefore have been refused the relief he sought.

Costs

52. Given the broader dimensions of this litigation, and that leave to appeal was sought, and granted, because this was a ‘test case’, I am of the view that it would be unfair to burden the employee with the costs in this court.

53. The order is as follows:

1. The appeal succeeds.

2. The order of the LAC is set aside. In its place is substituted:

‘(i) The appeal succeeds with costs.

(ii) The decision of the Labour Court is set aside. In its place is substituted:

“(a) The review succeeds with costs.

(b) The decision of the second respondent, the CCMA commissioner, is set aside.

(c) In its place there is substituted a determination that the dismissal of the employee was fair.”’

E CAMERON

JUDGE OF APPEAL

CONCUR:

HARMS JA

CLOETE JA
LEWIS JA
MAYA JA