Amcu v Chamber of Mines of SA (CCT87/16) [2017] ZACC 3 (21 February 2017) per Cameron J (Nkabinde ACJ, Froneman J, Jafta J, Madlanga J, Mbha AJ, Mhlantla J and Zondo J concurring)

In a unanimous landmark judgment the Constitutional Court held that the exercise of power by the Chamber of Mines entails public law consequences but does not mean that it is “administrative action” as defined in PAJA.  So the decision to conclude an agreement that the LRA extends to non-parties, was not “of an administrative nature”.  The parties were agreeing between themselves but not administering policy or statutory powers.  But as the agreement had wide-ranging public consequences their conduct was public, but not administrative, in nature.  Amcu’s appeal was disallowed.

[87]   One might ask how, if the statutory provision itself is not irrational, and indeed passes limitations analysis, there can be scope for irrationality review in its application. But a provision can rationally grant a power that may be irrationally exercised.  That is a matter for practical enforcement.  A particular agreement may be vulnerable to attack for irrational and undue effects on minority unions and non-members.  An instance might be where parties to a section 23(1)(d) agreement conclude it in flagrant breach of an express agreement with minority unions protecting them from the exercise of the power. [emphasis added]

See also: discussion at pages 283, 313 and 319 in Du Toit et al Labour Relations Law: A Comprehensive Guide 6th ed (LexisNexis 2015)

See also:

Excerpts without footnotes

Introduction

[1]   At issue is whether workers at five gold mines may exercise the right to strike while an agreement prohibiting strikes, to which they were not party, is in force.[1] The union representing the majority of workers at each of the mines is the first applicant, the Association of Mineworkers and Construction Union (AMCU).  The second and further applicants are its members at those mines.  But AMCU is not the majority union at any of the mining companies who own the mines.  The question is whether an agreement concluded between mining companies and their collective representative, on the one hand, and unions representing a majority of workers of those companies, on the other, binds employees at individual mines where their own union, which is not party to the agreement, is the majority union.

. . . . .

Rule of law

[60]   AMCU takes aim also at the mechanism by which the extension to non-parties was effected, and its impact once extended. The nub of its complaint is that section 23’s unsupervised private extensions license lawless exercises of power.  The provision thereby violates the rule of law, and is constitutionally offensive.

[61]   Its argument is that while the extension of the agreement to non-parties, including AMCU, did not in itself constitute an exercise of public power, it was “tantamount to the exercise of public power”. It points out that the three mining companies who through the Chamber concluded the agreement employ nearly three-quarters of all employees in the entire gold mining sector.  So the private extension of the agreement “by and large impacted an entire sector in the [country’s] economy”.  All this, it says, in private hands, with no judicial or legislative checks and balances.

[62]   Here, AMCU submits, “the rule of law is both relevant and applicable”. The section 23 scheme permits private actors to exercise public power arbitrarily.  This is antithetical to the open, accountable, democratic principles of our constitutional state.  In this way, section 23 offends the principle of legality – and indeed “poses a serious threat to our democratic [s]tate”.

[63]   Hence as an alternative to its interpretative challenge, AMCU asks the Court to declare section 23(1)(d) unconstitutional and invalid. AMCU’s rule of law challenge is not a rights-violation plus limitation challenge.  Instead, it impugns section 23(1)(d) as irrational.  It invokes the foundational value of the rule of law to contend that the provision violates the principle of legality.[1]  This is because it grants “private actors the right to effectively exercise public power arbitrarily, that is without observance of the rule of law”.

[64]   The respondents urged that section 23(1)(d) did not involve the exercise of public power. It was, they said, merely an ordinary statutory provision that allowed legal consequences to flow from private parties’ conduct.  This was unobjectionable.  Both the Labour Court and the Labour Appeal Court endorsed this approach.  The Labour Court held that there is “nothing inimical to the rule of law for legislation to provide for legal consequences to flow from the conduct of private parties”.[2]  The Labour Appeal Court similarly characterised the provision approvingly as one that allowed self-regulation as a means to supporting collective bargaining.[3]

[65]   In adopting this approach, the respondents’ argument and the Labour Appeal Court’s conclusion in effect treats section 23(1)(d) as though it is similar to the Wills Act.[4] This allows testators to determine the post-mortem distribution of their property.  That distribution obviously affects others – some may receive, others may be excluded – but the invocation of the statutory power remains in essence private: one between the testator and her earthly goods.  The power is statutorily sourced, but its exercise remains largely that of a private property owner choosing the devolution of her property after death.  As will appear, though the endpoint of the Labour Court and the Labour Appeal Court was unimpeachable, the path I take differs from theirs.

[66]   We start with the springboard for AMCU’s rule of law attack. The rule of law is enshrined as a foundational value in the Constitution.[5]  From this it flows as “axiomatic” that the exercise of public power must comply with the doctrine of legality, which stems from the rule of law.[6]  This foundational principle binds Parliament.  Its legislation must show “a rational relationship between the scheme which [Parliament] adopts and the achievement of a legitimate governmental purpose”, since “Parliament cannot act capriciously or arbitrarily”.[7]

[67]   Invoking Law Society of South Africa,[8] AMCU complained that section 23(1)(d) failed the test of legislative rationality. There this Court emphasised that the requirement that a legislative scheme must be rational “is not directed at testing whether legislation is fair or reasonable or appropriate”, but “is restricted to the threshold question whether the measure the lawgiver has chosen is properly related to the public good it seeks to realise”.[9]

This, the Court explained, is a lower threshold than a limitations analysis under section 36 of the Constitution.  That asks whether an infringement of an entrenched right is “reasonable and justifiable”.[10]  And indeed, the Court noted, to constitute a justifiable limitation, a provision must necessarily be rational: “It is self-evident that a measure which is irrational could hardly pass muster as reasonable and justifiable for purposes of restricting a fundamental right”.[11]

[68]   That reasoning applies here. As already explained, section 23(1)(d) infringes on the right to strike, but this is justifiably limited.[12]  Embedded in this conclusion, as pointed out in Law Society of South Africa, is that the provision is also rational.  Section 23(1)(d) furthers the legitimate governmental purpose of promoting effective collective bargaining by way of a scheme premised on majoritarianism.  As the provision is a constitutionally permissible limitation on certain entrenched rights, it is by corollary rational.[13]

[69]   And AMCU’s argument stems from an incorrect premise. Permitting a private actor to exercise public power does not inherently violate the rule of law.  Our constitutional scheme is more complex.  It casts up no impenetrable wall between the public and the private.  This is not least because the unjust grief of past exclusionary subordination and oppression based on race, which the Constitution is dedicated to eradicating, was perpetrated through both public and private means.  And the risk our Constitution recognises and confronts is that patterns of exclusion and discrimination could be perpetuated through ostensibly “private” exercises of power.

[70]   Thus the constitutional dispensation recognises that state organs and public authorities may perform acts that are not public in nature,[14] but conversely, and more pertinently to the present, that private actors may perform acts that entail the exercise of public power.[15]

This is because “public powers and public functions are wider than governmental powers and governmental functions”.[16]  The Bill of Rights itself binds natural and juristic persons if, and to the extent, that it is applicable, taking into account the nature of the right and the nature of the duty it imposes.[17]  The constitutional guarantee of just administrative action is conferred without distinction as to whether the actor is governmental or non-governmental.[18]  And so the Promotion of Administrative Justice Act (PAJA),[19] enacted to fulfil this guarantee, expressly covers administrative action taken by a natural or juristic person, other than an organ of state, when that person exercises a public power or performs a public function in terms of an empowering provision.[20]

[71]   Hence, it is trite that state organs do not alone exercise public power. Non-state organs may and do exercise public power.[21]  Beyond the initial question of typology (private vs public) lies the practically more crucial inquiry as to how the particular exercise of power is regulated and what safeguards exist for its exercise.  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.

[72]   Neither in its papers nor at the hearing did AMCU fully explain how section 23(1)(d) is irrational. As shown earlier, section 23(1)(d) promotes and serves the goals of collective bargaining, which is clearly a legitimate legislative end.[22]

[73]   AMCU framed its rule of law argument as a challenge to the constitutionality of section 23(1)(d), rather than to the particular extension the Chamber and the contracting unions effected here. But an extension could itself conceivably be subject to a legality challenge.  The typology thus answers the question whether invocation of the provision entails the exercise of a public power.  From there follows the more important substantive inquiry as to what safeguards apply to the exercise of the power.[23]

[74]   If the invocation of the powers section 23(1)(d) confers is public, then its exercise must comply with the principle of legality – and from there a range of review mechanisms is available to a party claiming to be unfairly affected. The actual exercise of the power the provision confers on private parties can never occur lawlessly.  It is subject to review under the principle of legality and, if it is administrative action, under PAJA.  So AMCU’s submission that section 23(1)(d) – in contrast to section 32 – does not allow for judicial checks on extensions of collective agreements is wrong.

[75]   So does using section 23 to extend a collective agreement to non-parties, including minority unions, entail the exercise of public power? In answering this, the predominant focus is on the nature of the power that is being exercised.[24]  The question is not so much, who exercises the power, nor even, where does the power come from: but what does the power look and feel like?  What does it do?  Pointers here include—

  • the source of the power;
  • the nature of the power;
  • its subject matter; and
  • whether it involves the exercise of a public duty.[25]

[76]   What do “public function” and “public power” mean? Langa CJ illuminatingly noted in a minority judgment in Chirwa:

“Determining whether a power or function is ‘public’ is a notoriously difficult exercise. There is no simple definition or clear test to be applied. Instead, it is a question that has to be answered with regard to all the relevant factors, including:

(a) the relationship of coercion or power that the actor has in its capacity as a public institution;

(b) the impact of the decision on the public;

(c) the source of the power; and

(d) whether there is a need for the decision to be exercised in the public interest.

None of these factors will necessarily be determinative; instead, a court must exercise its discretion considering their relative weight in the context.”[26]

[77]   And in Grey’s Marine, the SCA correctly stated that “the exercise of public power generally occurs on a continuum with no bright line marking the transition from one form to another”.[27]

[78]   In AAA Investments, the Micro Finance Regulatory Council, whose existence and functioning was recognised and approved by the Minister of Trade and Industry, played a regulatory role in supervising financial transactions. It took on many of the features of an organ of state.  In a minority analysis, concurring in the majority’s order, O’Regan J determined whether a private actor exercised public power by asking whether the decision is “coercive” in effect, and whether the decision is related to a “clear legislative framework”.[28]

Though the majority took a different path, nothing in its judgment disavows the more general significance of O’Regan J’s analysis.

[79]   Those features are present here. 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.[29]

[80]   These are far-reaching effects. They show that invoking section 23(1)(d) is not simply a private matter between private parties.  It affects the public.  The rationale for the power and its exercise is the public interest in improving workers’ conditions through collectively agreed bargains.  What is more, the decision to conclude and extend the agreement is rooted in Parliament’s exercise of its legislative power:  it is licensed by legislation.  The power is statutorily sourced.

[81]   This all points distinctly to a power more than private in nature. It is public. The conclusion of a collective agreement triggering a statutorily licensed extension under section 23(1)(d) is in its effects and substance an exercise of legislatively conferred public power.  It’s not the same as the statutory power the Wills Act confers on testators.

[82]   Features pointing to “public” are:

(a) the decision is rooted in legislation and its effects are circumscribed by the statute;

(b) the effect of the decision is mandatory on non-parties and coercive on their constitutional entitlements;

(c) the decision results in binding consequences without those parties’ acquiescence; and

(d) the rationale for extension is a plainly public goal, namely the improvement of workers’ conditions through collectively agreed bargains.

[83]   The Chamber and the mining houses it represented, together with the workplace-majority unions party to the agreement, were not governmental actors. Nevertheless their conduct had a sufficiently public character, and entailed sufficient public consequences, to make what they did the exercise of public power.

[84]   That their exercise of power entailed public law consequences does not mean that it was “administrative action” as defined in PAJA. This is because the decision to conclude an agreement that the statute, upon fulfilment of the conditions it specified, extends to non-parties, was not “of an administrative nature”.[30]

The parties were not administering policy or statutory powers; they were agreeing amongst themselves.  Their agreement had wide-ranging public consequences.  But in concluding it they did not act administratively.  Their conduct was public, but not administrative, in nature.[31]

[85]   This typology has the important consequence that the conclusion of an agreement under section 23(1)(d) is subject to judicial scrutiny. An agreement concluded under the provision is reviewable under the principle of legality.  The principle requires that all exercises of public power – including non-administrative action – conform to minimum standards of lawfulness and non-arbitrariness.[32]  Invoking the statute’s enormous clout by using a statutory power may not occur irrationally or arbitrarily.

[86]   AMCU’s grievance was of course not that the collective agreement here was capricious, irrational or arbitrary. It was that the statute offended the Constitution because a section 23(1)(d) agreement might be.  But since agreements concluded under the provision are liable to scrutiny, that grievance is abated.

[87]   One might ask how, if the statutory provision itself is not irrational, and indeed passes limitations analysis, there can be scope for irrationality review in its application. But a provision can rationally grant a power that may be irrationally exercised.  That is a matter for practical enforcement.  A particular agreement may be vulnerable to attack for irrational and undue effects on minority unions and non-members.  An instance might be where parties to a section 23(1)(d) agreement conclude it in flagrant breach of an express agreement with minority unions protecting them from the exercise of the power. [emphasis added]

[88]   But that question need not be answered now. The facts and the ambit of the parties’ arguments before us do not require a speculative quest for instances of in-practice (as opposed to facial) irrationality.  It is enough to note that parties extending agreements in terms of section 23(1)(d) may not irrationally exercise the power the statute confers.

[89]   It follows that AMCU’s challenges to the constitutional scheme that permits extensions of collective agreements to non-parties under section 23(1)(d) cannot succeed.