In terms of section 185(b) of the LRA every employee has the right not to be subjected to unfair labour practices. According to the first statutory definition, an unfair labour practice was “anything the industrial court deemed to be an unfair labour practice”.
The current definition of “unfair labour practice” reads as follows:
(2) “Unfair labour practice” means any unfair act or omission that arises between an employer and an employee involving—
(a) unfair conduct by the employer relating to the promotion, demotion, probation (excluding disputes about dismissals for a reason relating to probation) or training of an employee or relating to the provision of benefits to an employee;
(b) the unfair suspension of an employee or any other unfair disciplinary action short of dismissal in respect of an employee;
(c) a failure or refusal by an employer to reinstate or re-employ a former employee in terms of any agreement; and
(d) an occupational detriment, other than dismissal, in contravention of the Protected Disclosures Act, 2000 (Act No. 26 of 2000), on account of the employee having made a protected disclosure defined in that Act.
Only persons who are in employment (employees) are protected against unfair labour practices. Employees are protected against the unfair labour practices committed by the employer. Unlike the situation under the old Act, an unfair labour practice can no longer be committed by an employee and no remedies are available to an employer. This point is dealt with by Grogan in the following article: Not for employers:
Can employers sue their employees for unfair labour practices? If not, is the Labour Relations Act unconstitutional because it fails to afford employers such a remedy? It took a trade union, acting in its role as an employer, to move the Labour Appeal Court to answer these questions. The dispute began with a spat between the National Entitled Workers Union (“NEWU”) and its deputy president, who resigned about a month after he was appointed “because of the manner in which this organisation is being run.” NEWU accused its former vice-president of “not following a fair procedure” and claimed that his resignation had violated the union’s constitutional right to fair labour practices.
The union referred the dispute to the CCMA, seeking compensation for the unfair labour practice to which it had been subjected. A case management officer told NEWU that the commission lacked jurisdiction to entertain the dispute. NEWU then launched an application for review of that “ruling”, and sought an order declaring the Labour Relations Act 66 of 1995 and the Employment Equity Act 55 of 1998 unconstitutional because they failed to provide employers with remedies against unfair labour practices. The Labour Court dismissed the application and refused leave to appeal (see NEWU v CCMA & others  2 BLLR 165 (LC) – Employment Law 22(5)).
Having been granted leave on petition, NEWU persisted only with the argument that the LRA and the EEA were unconstitutional because they failed to recognise unfair labour practices perpetrated by employees. In NEWU v CCMA & others  7 BLLR 623 (LAC), the Labour Appeal Court accepted that the Constitution of the Republic of South Africa, 1996 gives “everyone” a right to fair labour practices. While neither the LRA nor the EEA recognises unfair labour practices perpetrated by employees, the Basic Conditions of Employment Act 75 of 1997 makes provision for termination of employment by either party to an employment relationship.
Actions by employers against employees were possible under the 1956 LRA. But no cases were reported in which employers sued employees for an unfair or unlawful resignation. This suggested to the court that there was no need for a remedy that went any further than that provided for in the current BCEA or the common law (see, for example, Nationwide Airlines (Pty) Ltd v Roediger & another (2006) 27 ILJ 1469 (W), in which an airline pilot was ordered to work out his notice period). The court concluded that the omission of remedies for employers in the LRA and the EEA was deliberate, and designed to counterbalance the inherent inequality between employers and employees.
This objective had been approved by the Constitutional Court. The LAC observed that it was strange that the first employer to bring an action against an employee and to challenge current labour legislation on the ground that it did not protect employers happened to be a trade union. Since NEWU was not satisfied with its right to sue its former vice-president for breach of contract, it had to be assumed that the union was seeking an order of compensation or reinstatement. The latter order would entail compelling a reluctant employee to render service after his resignation, which, in the circumstances, a court would, naturally, be disinclined to do. The appeal was dismissed with costs.
Unfair discrimination: equal pay for equal work
Wage war/ Unequal pay for equal work (article by Professor John Grogan)
Of the many freedoms inherent in a free-market economy, the freedom to determine wages is as fundamental as the freedom to determine prices of goods and services according to market demands. Both freedoms have been subject to some state regulation in capitalist economies, and collective bargaining has standardised and raised wages paid to groups of workers. However, one principle has remained sacrosanct: namely, that employers may pay higher salaries to some workers than to others on the basis of skill, seniority, responsibility and experience.
Graduated remuneration is found in every corporate structure, private or public. Nobody seriously questions the morality of paying an accountant more than a sweeper, and sweepers have not (yet) been heard to claim that they are entitled to be paid the same salaries as accountants. However, when one accountant is paid more than another by the same company, the possibility of controversy arises. That possibility is enhanced, too, when there is no apparent justification for different values an employer attaches to various categories of work.
The value attached by an employer to each such category makes no difference to the person who is labelled a “plumber” where he or she is paid less than a person who is dignified with the title of “sanitary engineer”, if both persons do the same work and the plumber is paid less. Apart from differentials within job categories, the size of the gap between the remuneration paid to employees in jobs that are generally regarded as deserving of higher pay, such as accountants, and the wages of employees in humbler positions, such as sweepers, may also cause dissatisfaction.
Grievances over the amounts employees earn relative to their colleagues (sociologists call this phenomenon “relative deprivation”) have been with us for as long as formal employment. The law’s involvement in this issue is relatively recent. The common law has stayed out of it. Even in labour law, disputes over levels of pay were initially regarded as “disputes of interest” – ie, non-justiciable. The law took an interest in comparative levels of remuneration only when it began to take an interest in discrimination.
Graduated income levels are not normally regarded as discriminatory, any more than is graduated income tax. They can, however, be an indication of discrimination. The question the lawmakers must decide is: in what circumstances can wage differentials be regarded as infringing upon the universally recognised right not to be discriminated against for unacceptable reasons?
South Africa’s first attempt to control unfair discrimination in the workplace was given legislative expression in item 2(1)(a) of the so-called “residual unfair labour practice” in Schedule 7 to the Labour Relations Act 66 of 1995, which has now been repealed and replaced by the virtually identical section 6 of the Employment Equity Act 55 of 1998. Item 2(1)(a) of Schedule 7 prohibited any unfair act or omission between an employer and an employee based on any arbitrary ground including, but not limited to, race, ethnic or social origin, sex, sexual orientation, gender, family responsibility, political affiliation and belief. Relatively few employees went to court under this provision from the time it was put into force at the end of 1996 to the time of its repeal in 1999.
This is surprising. The very reason item 2(1)(a) was enacted was that the South African labour market is notoriously skewed as a result of past discrimination. A few employees and applicants for employment succeeded in proving that they were discriminated against by not being appointed (Whitehead v Woolworths (Pty) Ltd  8 BLLR 862 (LC) – subsequently reversed on appeal: see Woolworths(Pty) Ltd v Whitehead  JOL 6374 (LAC)) or given equal benefits (Leonard Dingler Employee Representative Council & others v Leonard Dingler (Pty) Ltd & others  11 BLLR 1438 (LC)). However, the Labour Court indicated from the first such application that it would not accept that the fact that an employee is black and paid less than a white colleague is proof in itself of unfair discrimination. Thus, in TGWU v Bayete Security Holdings  4 BLLR 401 (LC), the applicant employee, who had formerly been a security guard, was given what he termed a “marketing job”. He was paid R1 500 a month.
Soon afterwards, the company employed a white man at a salary of R4 500 a month. The applicant queried the difference between his salary and that of the new appointee. For his pains, the applicant was sent back to the beat. The Court held that the applicant had not passed what it described as “the first hurdle” in a discrimination action – to prove that he was discriminated against. All that the applicant had proved was “that he, a black, was earning R1 500 and that Louw, a white, was earning R4 500”. The applicant had conceded in evidence that he did not know what work his white colleague performed, what his educational qualifications were, for whom he had previously worked and for how long.
In Bayete Security Holdings, the Court was not required to go further than that. Now, in Louw v Golden Arrow Bus Services (Pty) Ltd  3 BLLR 311 (LC), a wage discrimination case was more fully and imaginatively presented. The case is worth close attention because it explains why the anti-discrimination provision in Schedule 7 did not prove to be the panacea some employees might have hoped it would be for assuaging feelings of relative deprivation about their salaries.
Mr Louw was employed, by a wholly owned subsidiary of Golden Arrow, as a buyer, in 1984, at a salary of R750 a month. By 1990, his salary had risen to R1 500. Louw did not complain that his increases over those 14 years were too slow. He could not, as the law is not concerned with the size of employees’ increments. However, he contended that the company had commenced discriminating against him from 1990 when it appointed a white man, a Mr Beneke, as a buyer on a salary of R2 300 per month and promoted him to warehouse supervisor in 1994.
Louw argued that the discrimination grew worse as his salary and that of Beneke were annually increased by the same percentage, resulting in a gap of R2 055 between their salaries by 1998. Louw’s case, as outlined in his pleadings, was that, at all material times, his work and that of Beneke were of equal value or, alternatively, that the difference in salary was disproportionate to the value of the two jobs. Fundamental to Louw’s case was that the reason for the difference between his salary and that of Beneke was that he was black and Beneke, white. However, Louw went further. He did not merely claim that the company had and was continuing to exploit him because he was black.
He contended, as a backstop, that the difference in his salary and that of Beneke constituted indirect discrimination on the grounds of race, colour or ethnic origin because the company “applied facts in its pay evaluation that had a disparate impact on black employees”. He listed these factors as performance, potential, responsibility, experience, education, attitude, skills, entry level and market forces. The company contended, quite simply, that there was a difference between the two employees’ salaries because their work was not of equal value, and that the difference was attributable to a number of considerations, none of which involved race discrimination.
Thus presented, the dispute went to the heart of the employer’s prerogative to decide how to categorise jobs and attach rewards to them. If Louw’s alternative claim were good in law, it would mean that no employer could take into account such factors as performance, potential, responsibility, experience, education, skills, the employee’s salary in previous employment, or supply and demand when appointing blacks or women to positions for which whites or white males had applied or in which whites or white males were employed.
Blacks and women would either have to be paid the same as their white male colleagues, or not be appointed. The latter possibility could, of course, compromise employers in respect of the obligation that they now bear under the Employment Equity Act to implement affirmative action. More seriously, every employee that has blacks and whites employed in similar positions would have to equalise their salaries or face a flood of claims from the previously disadvantaged.
However, before Louw reached the hurdle of proving that he was the victim of unfair discrimination as envisaged by item 2(1)(a), he had a preliminary one to surmount. This was the company’s objection that the Court lacked jurisdiction because the dispute had arisen before the implementation of the Act. That plea was considered in Louw & another v Golden Arrow Bus Services (Pty) Ltd (1998) 19 ILJ 1173 (LC). Basson J dismissed the point, holding that there was a difference between the form of unfair labour practice alleged by the applicant and a dismissal. Unlike a dismissal, the practice of which Louw complained was a “continuous act” which was allegedly
“the result of a (policy) decision introduced by the employer in terms of which the employer . . . pays employees who do the same work as other employees less on the basis of their race. This . . . clearly is a continuing activity which commences as soon as this practice is introduced and ceases only when the employer stops implementing the decision or policy. The employer is not committing a single and separate unfair labour practice each and every time an employee is either overpaid or underpaid but these payments are merely facts by way of which the existence of such continuous unfair discriminatory practice is indicated.”
On this basis alone, the Court had jurisdiction, even though Louw first objected to the discrepancy in salary between himself and his white colleague as long ago as 1991. However, there was another reason that founded jurisdiction. This was that the company had created the position of warehouse supervisor in 1994 and appointed Beneke to it in preference to Louw and other internal candidates.
Beneke’s promotion, said the Court, introduced a new dimension to the dispute. For, now, the applicant could claim not only that Beneke was earning more while he was employed in the same job as they were doing, but that he was earning more while doing work of equal value. The applicant had not articulated a grievance over Beneke’s promotion until after the commencement of the Act. This, said the Court, was enough in itself to found jurisdiction.
The latter finding is not without significance. It means, for example, that an employee who was overlooked for promotion many years before the commencement of the Act can create a dispute that is justiciable under the Act merely by raising a grievance. This may come suspiciously close to applying the Act retrospectively. However, it is difficult to escape the logic of the Court’s observation that an unfair promotion can have continuing effects, and that every benefit accorded the promoted employee by virtue of the initial promotion merely aggravates the consequences of the unfair promotion or failure to promote.
The jurisdictional hurdle crossed, Louw had to persuade the Court that his cause of action fell within the terms of paragraph 2(1)(a) of the residual unfair labour practice. This proved to be more complicated than might at first glance have appeared to be the case. Unlike the labour courts under the 1956 LRA, the Labour Court cannot decide what is and what is not an unfair labour practice on intuitive feelings about fairness alone. Identifying an unfair labour practice no longer entails merely the exercise of a “moral or value judgment” (NUMSA v Vetsak Co-operative Ltd & others  6 BLLR 697 (AD)), but is now, first and foremost, an exercise in statutory interpretation.
This is how the Court approached the matter in Golden Arrow. Landman J noted that item 2(1)(a) required an act or omission which involved unfair discrimination. Louw contended that the company was responsible for an omission. But did that omission “involve” discrimination? The question is critical because the connection between the act or omission and alleged discrimination lies at the heart of what item (2)(1)(a) proscribes. Even if it were accepted that Golden Arrow was responsible for an omission (presumably by failing to comply with Louw’s demand to increase his salary to the level of that of Beneke), the question remained whether its refusal to do so was because of Louw’s being “coloured”.
As Landman J noted, the word “involve” is not one of great precision. The closest the dictionary comes to defining what the legislature meant by “involve” in paragraph 2(1)(a), it seems, is when it assigns the following other possible meanings of the word: “to include covertly in or under something, wrap up” or “contain implicitly, include as essential; imply, call for, entail”. The word “covert” implies secrecy which, in turn, implies intention. The second definition posits only a connection: A is “involved” in B if it results in or contributes towards B, whether logically or causally.
The Court rejected the company’s contention that the involvement contemplated in item 2(1)(a) was limited to intentional discrimination, such as has been required by the United States courts in cases of so-called “disparate treatment”, as opposed to “disparate impact” (see, for example, Watson v Fort Bank & Trust 487 US 977 (1988)). According to the Harvard Law Review ((1996) 109 7), in the former class of cases, the complainant objects to “employment practices or incidents that intentionally subject people to impermissible discrimination”. In the latter class, the complaint is about “neutral employment policies, such as competency tests, that have the unintended effect of discriminating against individuals who belong to protected classes”.
It is difficult to comprehend the practical importance of this distinction. The difference between the two classes of discrimination turns, in effect, on the way in which the complaint is framed. If complainants allege that they have been intentionally discriminated against, they must, according to this distinction, prove that the employer intended to discriminate. If they have not made such an allegation, they need not prove intent. This, surely, cannot be the effect of a statute that prohibits certain forms of conduct. Such statutes must either require that dolus or culpa is a requirement for liability in respect of all conduct falling within its terms, or none. It would seem strange indeed if, in some cases of the same crime, intention or negligence is required and, in others, not.
This is why the Court was, with respect, correct when it declined to follow the intricacies of American jurisprudence in resolving this issue. Landman J observed, correctly it is submitted, that, in South African law, whether the commission of a prohibited act requires fault in the form of intention or negligence depends on the interpretation of the statute that creates the offence. Although the commission of an unfair labour practice is not an offence, the sameconsideration applies to an Act that imposes civil liability.
Why, then, must it be assumed that the legislature intended to make an employer liable for unfair discrimination even if it was unintended? If the use of the word “involving” does not provide a clear indication of the legislature’s intention in this regard, the express statement that an unfair discrimination can be “direct” or “indirect” certainly provides a clue.
The distinction between “direct” and “indirect” discrimination has its origin in American law. Like the distinction between “disparate treatment” and “disparate impact”, the distinction between direct and indirect discrimination differentiates between discrimination to which people are subjected only because they happen to have some personal identifying characteristic, and discrimination that results from barriers aimed at differentiating between people according to characteristics which are apparently “neutral” and, on the face of it, legitimate, but still have the effect of disadvantaging persons because they happen to belong to a particular group.
Direct discrimination must self-evidently be intentional. Indirect discrimination may be intentional, but need not be. The point is that the idea of indirect discrimination has evolved precisely to prevent discrimination resulting from social practices that have stacked the odds against groups of people in the past, and that continue to do so. Anti-discrimination law is part of a programme of social engineering that would be seriously compromised if people were permitted to engage in discriminatory practices simply because they are unaware that such practices are discriminatory.
The elimination of intention or fault does not remove the philosophical difficulties created by the prohibition on unfair discrimination: two problems remain. The first is to decide which differential treatment constitutes discrimination. The second is to decide on what basis discrimination can be held to be unfair. It is necessary to decide the first question because otherwise any form of inequality potentially falls within the net of the anti-discrimination clause. Not all forms of inequality amount to discrimination. That some people are born cleverer or stronger than others does not mean that the others are discriminated against because the cleverer or the stronger use their wiles or their strength to gain advantages. Discrimination arises only when some are favoured over others by persons with the power to confer advantages. It is, therefore, a social concept.
Society permits some forms of discrimination because they are considered legitimate, either because people are permitted to compete for advantages by using the strengths with which they are endowed by nature or because the denial of advantages is considered to be in the interests of those discriminated against – and of society. Wage discrimination is generally considered permissible for the former reason. In an ideal world, paying an accountant more than a sweeper would not be regarded as discrimination because there would be nothing other than natural ability or the capacity for work to prevent the sweeper from being an accountant. It is only when people are or have been prevented from exercising their natural talents in order to compete for advantages that differentiation becomes discrimination in the pejorative sense.
This is why, in Bayete Security, the Court required the applicant to prove that there was something other than the fact that he was black and his higher-paid colleague white before it was prepared to conclude that the differences in their wages amounted to discrimination. This is also why the legislature stated that discrimination is impermissible only when it is exercised against an employee on any arbitrary ground, including, but not limited to, those set out in item 2(1)(a).
A ground, in this context, means the reason why the person is discriminated against. Once intention is excluded, the “reason” for discrimination is the attribute which, objectively considered, explains why a person is relatively disadvantaged. So, there must be a causal connection between the possession of that attribute, on the one hand, and the relative disadvantage, on the other. The Court recognised this in Golden Arrow, and also the need for another limitation. As Landman J observed:
“It is necessary to distinguish clearly between discrimination on permissible grounds and impermissible grounds. An unfair labour practice is only committed (even by omission) if the impermissible grounds are the cause of the discrimination. Discrimination on a particular ‘ground’ means that the ground is the reason for the disparate treatment complained of.
The mere existence of disparate treatment of people of, for example, different race is not discrimination on the ground of race unless the difference of race is the reason for the disparate treatment. Put differently, for the applicant to prove that the difference in salaries constitutes direct discrimination, he must prove that his salary is less [than] Mr Beneke’s salary because of his race.” (Court’s emphasis.)
In yet other words, an applicant under item 2(1)(a) must prove that the reason he was discriminated against was impermissible, and that he was discriminated against for that reason and no other. The legislature stated that employers may not discriminate against their employees on “arbitrary grounds”, and provided a long list of grounds considered by it to be arbitrary. Item 2(1)(a) specifically stated, however, that the examples of listed impermissible grounds must not be considered closed. Other “arbitrary grounds” may, therefore, be conceivable.
The question is: how far should the net be spread? The Constitutional Court has already indicated that arbitrary grounds under the similar provision in the Constitution should be limited to grounds which, objectively considered, are “based on attributes and characteristics which have the potential to impair the fundamental human dignity of persons as human beings or to affect them adversely in a comparably serious manner” (see Harksen v Lane NO 1998 (1) SA 300 (CC)). In other words, whether discrimination is considered to be on an arbitrary ground – and, therefore, impermissible – depends either upon the reason the victim was discriminated against or on the effect of the discrimination.
This distinction is difficult to grasp. The attributes or characteristics mentioned in the anti-discrimination clause in the Constitution and in item 2(1)(a) cannot, in themselves, “have the potential to impair human dignity” – quite the reverse. It is not the characteristic that impairs human dignity but the manner in which its possessor is treated because he or she possesses that characteristic. It can, accordingly, only be the effect of the discrimination that impairs human dignity. The legislature appears to be saying that inequality that arises because of some arbitrary characteristic is unfair, which, as the Court said in Harksen’s case, is the next question.
In Golden Arrow, the Court did not have to deal with the scope of the expression “arbitrary grounds”. Had Louw alleged that he was paid less than Beneke simply because there was no rational basis for the differential, the Court would have had to wrestle with the problem of whether the employer’s mere failure adequately to justify the difference in the two men’s salaries rendered it arbitrary and accordingly unfair. However, Louw’s case was based on the contention that he was paid less than Beneke because of his (Louw’s) race, and he asked the Court to infer from the absence of a rational justification for the difference in salaries that it was race that accounted for the difference. This led the Court back to the question of causation.
Landman J noted that, in this regard, the English courts relied on the standard legal test for causation – namely, the sine qua non or “but for” test: would the complainant have received the same treatment but for his or her race, sex, religion, belief, etc? However, as Landman J noted, the test does not go far enough for the purposes of South African discrimination law, which raises the question of whether an impermissible ground must, as the learned judge put it, “be the sole cause of the discrimination or whether it is enough that it be a cause”.
Landman J identified three possible approaches to this question. The first is to determine whether “any contamination by impermissible unfair discrimination is sufficient to find that the act or omission complained of is caused or attributable to it”. The second is to find that there has been “contamination” only if the contamination is material.
The third is to find that there is unfair discrimination “to the extent that the discrimination in the case under investigation is caused or contaminated by it”. By “contaminate”, the learned judge clearly meant “cause” in the sense that contamination by arsenic causes food to be poisoned. According to the first test, the person who has caused the death of another by adding arsenic to the latter’s food will be liable for murder, irrespective of how miniscule the quantity of the poison.
According to the second test, liability will follow only if the arsenic added was sufficient in itself to cause death. According to the third, the poisoner will be guilty of murder, but his or her penalty will be determined by the amount of arsenic added. In other words, the degree of contamination affects not the decision as to whether murder (or unfair discrimination) has been committed, but the sentence (or remedy) that is called for. Landman J, while accepting the third approach, described the exercise as “akin to an attempt to unscramble an omelette”.
He is, with respect, correct. Take Mr Louw as an example. He is a member of an historically disadvantaged group. He entered the employment of Golden Arrow at a time when, to the extent that apartheid made it more difficult for people of colour to get jobs, the company could have exploited the situation, whether consciously or not, by paying him less. Similarly, Beneke might have earned a salary higher than he would otherwise have done in his previous j