, , , ,

Yesterday, the South African Constitutional Court handed down an important judgment at the interface of collective bargaining agreements, majority and minority unions, and the right to strike. From a comparative perspective, this judgment is interesting, because it reveals certain structural similarities as well as dissimilarities between South African and Indian labour law, the different nature of the problems that arise therefrom and, consequently, the shape of the judicial response to those problems.


The dispute arose out of the mining industry which, as the Court acknowledged at the beginning of the opinion, was the site of “the grievous struggle for better wages and conditions for the generations of mineworkers who have laid the foundations for this country’s wealth.” The facts were simple enough: in 2013, the Chamber of Mines of South Africa, on behalf of its members (various gold mining companies) entered into a collective bargaining agreement with three trade unions that represented a majority of the miners in the mining sector (these were the National Union of Mineworkers, Solidarity, and United Association of South Africa). The Association of Mineworkers and Construction Union (AMCU) – a relatively new union with a swiftly rising membership – was invited to join the negotiations; however, it refused to sign on to the collective bargaining agreement. The agreement was made applicable to all the company’s employees, including non-members of the unions.

While AMCU was not a majority union in the mining sector as a whole, it was a majority union in certain specific mines. Consequently, and regarding itself not bound by the collective bargaining agreement that it had never signed on to, AMCU gave a notice of strike at the five gold mines where it was the majority union. The strike was interdicted by the Labour Court, a decision that was upheld by the Labour Appeal Court.At issue was Section 23(1)(d) of the South African Labour Relations Act, which authorised employers and unions to make a collective agreement binding on non-parties (to that agreement) if, inter alia, the trade unions party to the agreement “have as their members the majority of employees employed by the employer in the workplace”. Workplace was defined under the LRA as “the place or places where the employees of an employer work“, with the proviso that “if an employer carries on or conducts two or more operations that are independent of one another by reason of their size, function or organisation, the place or places where employees work in connection with each independent operation, constitutes the workplace for that operation”.

The lower courts found that for the purposes of the gold mining industry, separate mines did not constitute separate workplaces. Consequently, under Section 23(1)(d), AMCU was bound by the majority unions’ decision to make the collective bargaining agreement binding on all workers. And therefore, the provisions of S. 65 of the LRA – which prohibited striking by anyone who “is bound by any arbitration award or collective agreement that regulates the issue in dispute” – were applicable, and AMCU was statutorily indicted from striking in the five gold mines.

AMCU then approached the Constitutional Court. The issues, as summed up by the Constitutional Court in paragraph 10 of its judgment, were as follows:

“The question is whether the agreement bound AMCU members at the five mines where it was in the majority. If it did, the statute prohibited its members from striking. If it didn’t, they were statutorily at liberty to strike. It all turns on what “workplace” means in the statute and, more specifically, in section 23(1)(d). Does it mean all the mines of the Chamber member companies overall – where AMCU was in the minority? Or the individual goldmines – where it had a majority? And if it was all the mines of the member companies overall, thus snatching away from AMCU members at the individual mines their right to strike, does the statutory provision withstand constitutional challenge?”

Meaning of “Workplace” 

The Constitutional Court began by noting that “two things are immediately notable about the way the statute defines “workplace”. The first is its focus on employees as a collectivity. The second is the relative immateriality of location. Both signal that “workplace” has a special statutory meaning.” (paragraph 24)

The Court analysed these two aspects in the broader context of the purpose of labour legislation, which was “the promotion of orderly bargaining by workers, collectively.” (para 25). For this reason, the definition of “workplace” was to be understood in the context of the statute’s focus on workers “as a collectivity“. From this, it followed that “location is not primary: functional organisation is.” (para 26) What of the proviso? Here, the Court noted that what constituted an “independent” workplace was also to be determined by applying the test of functional organisation. Since the “Labour Court and the Labour Appeal Court both found, in conclusory terms, that the individual AMCU-majority mines did not constitute independent operations“, and since AMCU had failed to demonstrate an alternative to the geographical interpretation of “workplace”, the Constitutional Court upheld the findings of the courts below. Consequently:

“It follows that the agreement was validly extended to AMCU members at the five AMCU-majority mines. The question now is whether the statutory provision that allowed this withstands constitutional scrutiny.” (paragraph 40)

Constitutionality: The Right to Strike

The South African Constitution – unlike the Indian – contains a fundamental right to strike. AMCU argued that the interpretation of S. 23(1)(d) privileged the principle of majoritarianism in the workplace, which – effectively – stifled dissident unions’ freedom of association and the right to strike. This is, of course, a core question, that goes to the heart of the idea of industrial democracy. And as the Court correctly noted, [this] challenge is freighted with history, and burdened by recent clashes between unions in many workplaces, including in the mining industry.” (paragraph 42) Referring to precedent and the scheme of the LRA, the Court noted that majoritarianism was its guiding principle: for example, “it is majoritarianism that underlies the statute’s countenancing of both agency shop agreements (deductions for majority union fees from all employees, both members and non-members),41 and closed shop agreements (collective agreement may oblige all employees to be members of the majority trade union).” (paragraph 43) The reason for this, according to the Court, was that “for orderly and productive collective bargaining, some form of majority rule in the workplace has to apply. What section 23(1)(d) does is to give enhanced power within a workplace, as defined, to a majority union: and it does so for powerful reasons that are functional to enhancing employees’ bargaining power through a single representative bargaining agent.” (paragraph 44)

The key question was whether this enhanced power vested in the majority union violated the constitutional scheme. Of course, as the Court correctly noted, AMCU was itself advocating a form of majoritarianism – one that operated on a mine-by-mine basis. Nonetheless, the Court considered the question of whether the constituency that was the site for majority rule – that is, the “workplace, determined from the point of view of collectivity, subject only to functionally-determined independence of operation” (paragraph 48) was constitutional. The Court held that it was. Among the many models of possible majoritarianism, the legislature had selected one, on the basis of its judgment that it “benefits orderly collective bargaining.” Now, in a previous case called Bader Bop, the Court had held that the LRA had to be interpreted so as to “allow minority unions to co-exist, to organise members, to represent members in relation to individual grievances and to seek to challenge majority unions.” (paragraph 52) Under the LRA, minority unions – including AMCU – retained their organisational rights, recruitment rights, deduction rights, shop steward rights, and bargaining rights. In this context, the loss of the right to strike had to be viewed in context of the fact that “the LRA, though premised on majoritarianism, does not make it an implement of oppression. It does not entirely suppress minority unions.” (paragraph 55) Consequently, the Court held:

“The limitation a section 23(1)(d) agreement imposes on the right to strike is strictly circumscribed – in both ambit and time. A collective agreement extended to non-parties does not apply to them indefinitely. It applies only for the duration of the agreement and regarding the specific issues it covers. Section 23(1) does not countenance indefinite or far-reaching extension. It directly ties the limitation of the right to strike to the outcome of the collective bargaining. It is narrowly tailored to the specific goal – orderly collective bargaining. Given the carefully circumscribed ambit of the limitation and the importance of its purpose, it is reasonable and justifiable.”

Rule of Law

AMCU raised a fascinating argument that Section 23(1)(d) was unconstitutional because in allowing an unsupervised extension of the collective bargaining agreement to non-members, it vested unsupervised public power upon private bodies. The Court, however, rejected any facile division of the public and the private, noting that history, as well as the constitutional scheme, was keenly aware of the interpenetration of the two domains. Consequently, the key issue was that:

“When legislation authorises private parties to exercise public power the question is thus how to ensure a rational relationship between their exercise of power and the attainment of legitimate legislative ends.”  (paragraph 70)

Here, there was clearly a rational relationship between the section, and the “legitimate legislative end” of promoting collective bargaining. Consequently, ex facie, the Section itself was not unconstitutional. However, the Court also noted that under accepted legal tests – following Justice O’Regan’s opinion in AAA Investments vs Micro Finance Regulatory Council – Section 23(1)(d) did vest public power in private bodies:

“The decision by private parties to invoke the power section 23(1)(d) affords them to extend their collective agreement to parties entirely alien to it has a coercive effect: it binds non-parties to the agreement, willy-nilly. And, as AMCU rightly points out here, the statute empowers contracting parties to do this with just about industry-wide effects. The extension of the agreement also has extensive implications for members of the public. For its duration, non-member employees are bound. Even more, they forfeit the right to strike if the collective agreement regulates the issue in dispute.”  (paragraph 78)

This meant that a Section 23 agreement would be subject to judicial scrutiny, under the principles of legality, lawfulness and non-arbitrariness. However, since the AMCU’s challenge was to the provision itself, and not to the specific agreement in the present case, it failed on this ground as well.

Interesting Features

The Constitutional Court’s judgment in AMCU vs Chamber of Mines raises issues that have comparative interest. For example, the Court’s decision to interpret the term “workplace” not in its ordinary linguistic sense (i.e., geographically and spatially) but in a non-geographical sense, based on the purpose of the LRA (the “collectivity” principle) is reminiscent of the Indian Supreme Court’s repeated holding on the meaning of the word “industry”, which it has interpreted in the context of the Industrial Disputes Act’s overall purpose of peaceably resolving labour disputes through a tripartite adjudicatory process.

Secondly, the Court’s focus on collective bargaining as the principle that is at the heart of the LRA demonstrates how South African labour law is based upon a model of trade union empowerment that goes beyond the Industrial labour regime. Under the Industrial Disputes Act, collective bargaining agreements between management and a union are not binding upon non-members, on the basis that non-members to a contract cannot be bound by its terms (subject to one exception in the ID Act, and subject to exceptions in various state labour laws, which provide for circumstances in which recognised “representative” unions can bind everyone in the establishment). As we have discussed before on this blog, the Indian labour regime is based on the idea that the State acts as a neutral arbiter of the competing interests of capital and labour, and the strength of trade unions is accordingly diluted. The South African model, on the other hand, seems to favour stronger trade unions, and thus overrides the principle of privity of contract in the interests of – as the Constitutional Court puts it – the “collectivity”.

And lastly, of course, the Court’s examination of the public nature of the power exercised under Section 23 speaks to a point that the Indian Supreme Court has also made repeatedly: labour legislation lies at an interface of the private and the public; the workplace is not to be considered as immune from the infusion of public law norms. Of course, which norms apply to the workplace, and how, remains a controversial question in both jurisdictions.