The Employment Equity Act, 1998 (And Other Myths About the Pursuit of 'Equality', 'Equity' and 'Dignity' in Post-Apartheid South Africa) Part 1

The author critically examines the organising principle of the affirmative action provisions of the Employment Equity Act (or EEA), as well as the implications of the recent judgment by the Constitutional Court in its first case involving the application of affirmative action in the employment context (and in terms of the EEA) -- SAPS v Solidarity obo Barnard. While reiterating the need for restitutionary measures such as affirmative action in South Africa, the author concludes -- probably quite controversially -- that the EEA's treatment of affirmative action has nothing to do with the equality right in the Bill of Rights, and that the Act pursues a different (and omnipresent) social engineering agenda by the state. The author calls for this realisation to prompt future affirmative action cases arising from the application of this Act to be removed from the scheme of (and potential defences available under) the equality jurisprudence, and for the courts to critically interrogate the constitutionality of the EEA's affirmative action scheme within its own context. The author believes that Chapter III of the Act is unconstitutional in this sense, and he calls for the scrapping of its provisions. He also calls for a (more) constitutionally-compliant exposition from the Constitutional Court of the parameters of legitimate affirmative action under the Bill of Rights, and adds his voice to the numerous calls for reconsideration of the "rationality test" expounded in Minister of Finance v van Heerden. More generally, the author considers the apparently all-pervasive application of the government ideology of the pursuit of demographic representivity in "transformation" of employment and other contexts (expressing grave doubts about its constitutionality along the way). This article forms Part 1 of this piece and the author considers the constitutional requirements for a legitimate affirmative action programme or measure. He then examines the affirmative action scheme of the Employment Equity Act, and explains his views on why such scheme is, in fact, unconstitutional. In Part 2 of this piece (which follows in this edition), the author continues to critically evaluate the Constitutional Court judgment in the Barnard case, and he highlights the biggest areas of disappointment of this judgment within the context of South Africa's equality jurisprudence. After a very brief consideration of the recent amendments to the Employment Equity Act, the author concludes by providing reasons why the Act's approach to affirmative action needs to be rejected, and soon.


Introduction
As I write this, many pundits are probably still reeling -in disappointment, if nothing else -from reading the various, variegated and sometimes verbose judgment(s) in what promised to be one of the most important employment law cases to confront the Constitutional Court (or CC) in recent years. Labour lawyers have not always had the best of times before this court; the clear-as-mud judgment in Chirwa v Transnet 4 still haunts classrooms and courtrooms alike, and it was probably hoped that the CC's eventual decision in SAPS v Solidarity obo Barnard 5 would have a less chequered reception. Barnard's case is not only the first on affirmative action in the employment context to grace the hallowed halls of the Braamfontein court (following its landmark (other) affirmative action judgment in Minister of Finance v Van Heerden 6 ), it is also a case that has run the gamut of our judicial system, having ended up before the CC after a protracted innings that included airings before the Labour Court, the Labour Appeal Court and the Supreme Court of Appeal. Eventually, the claimant ended up with 2 wins and 2 losses -reminiscent of what would have been a very exciting 5match Test cricket series between the Proteas and Australia. Ms Barnard, however, ended up one Test match short of clinching a series win in her favour, and with the CC being her final port of call, effectively a rained-out final match which saw her losing the series. Our highest court closed the door on her case, and, more importantly, on many of the legal issues with broader relevance that were raised in her case, and one can only hope for a future tour from some other team to definitively determine the state of the pitch to everyone's satisfaction. At the very least, the history of the AM LOUW PER / PELJ 2015(18)3 596 litigation in this matter appears to have uncovered some rather significant fault lines (foot marks?) in the respective judges' understanding of the constitutionally mandated framework of affirmative action, and of the boundaries of legitimate measures employed in its name. We all know that hard cases, generally, make for bad law. It is doubtful that Barnard's case can easily be classified as such (which, I'm sure, is why this judgment was received by many with a sense of shock), so hopefully protracted cases don't make for the same standard of legal (un)certainty. Only time will tell, although early indications are that the Barnard judgment will probably not end up in any future compilation of the Constitutional Court's greatest hits.
In the meantime, and because the CC was not called upon to do so, I hope (once again 7 ) to consider the legitimacy of the legislative instrument that is behind the affirmative action measures employed in Ms Barnard's case and in so many others, and which I believe is also behind most of the problems experienced with the application of affirmative action in our workplaces. I intend to ask whether the Employment Equity Act (or EEA) 8 is still (or really?) the best way we have been able to find in our democratic dispensation to deal with the thorny issues of restorative justice, the promotion of substantive equality, and the much-vaunted process of nation-building. Also, seeing that the EEA was recently amended (for the first time since its inception 16 years ago), this is an opportune time to also consider the changes brought about so recently by the legislature. 9 The amendments have been significant, and they warrant closer attention, even though -and here I will start to pin my colours to the mast -I believe they may constitute little more than the equivalent of sticking a Band-Aid on an amputated limb. 7 I have written previously, at length, on the problems I perceive with this Act and also other forms of affirmative action in the South African context which appear to follow its lead (specifically, the always controversial sports transformation agenda pursued by government to date): see  Stell LR (Part 1);  Stell LR (Part 2);  Stell LR (Part 3); Louw 2005 LDD; Louw 2006 ISLJ; Louw 2007 De Jure; Louw 2008 De Jure. On the Employment Equity Act 55 of 1998, specifically, see Louw 2006 SA Merc LJ (which will be referred to in more detail in the text below). 8 Employment Equity Act 55 of 1998 (or EEA), as amended by the Employment Equity Amendment Act 47 of 2013. 9 The amendments in terms of Act 47 of 2013 will be (very briefly) considered in s 5 of the text below.
I believe it is important to record here that I do fervently believe that we need some species of affirmative action in South African workplaces, and that we will probably need it still for some time to come. This is not yet another call for a 'sunset clause' to the Employment Equity Act (or, more specifically and relevantly, its chapter on affirmative action in the workplace). Forget sunset, that would be much too natural a way for this Act to go "gentle into that good night". 10 I would like to see a more drastic (and much quicker) end to it. The EEA is not and has never been the way to regulate affirmative action. We definitely can, and should, do better than this controversial piece of legislation, which -as I will argue -is unconstitutional in respect of its affirmative action scheme. In fact, had Ms Barnard's legal team decided to challenge the constitutionality of the relevant parts of this Act, the CC's judgment may have contributed more to our equality jurisprudence than it promises to have done on the basis of how the case was argued before this court. More will be said on this later. For now though, I feel it is important to note that this piece will follow a different approach and focus to that found in most of the other literature on affirmative action as published in South Africa in the past decade or so. Much has been written (and much of it is quite convincing) about the justifications for affirmative action and why we (still) need it, 11 and how it can/should contribute to our continuing project of the embodiment of transformative constitutionalism in everything we do. This piece, however, will focus on the practical experience of affirmative action (under the EEA).
It is all good and well to make lofty proclamations in this regard, but I hope to rather bring things down to the level of the experience of these lofty ideals and policies in action. The experience of the application of affirmative action (especially in the public sector) displays a clear dissonance between principle (as pronounced by the Constitutional Court, in terms of the theoretical framework for such measures in respect of the limits set by the Constitution) and practice. 12 It is in the latter that we 10 With apologies to Dylan Thomas. 11 See, for example, Dupper 2004 SALJ; Dupper and Garbers 2012 Reinventing Labour Law. 12 A press release by the Helen Suzman Foundation, published shortly after the Barnard (CC) judgment was handed down by the Constitutional Court, remarked on the fact that the implementation in practice of affirmative action might be more in need of constitutional adjudication than the principle involved: "Even though they may have reached the same decision by different means, the unanimity of our court's decision speaks volumes: transformation is here to stay. The real question that remains to be asked is how the CC properly evaluates cases where find the major problems, and the (sometimes understandable) opposition to affirmative action which it generates. And it is also here that we find the evidence of a rather shady governmental agenda at work in the Act's affirmative action provisions, which I hope to deal with in some detail in this piece.
In the section that follows I will delve into the question of whether the Act (or, at least, the scheme that it established for the application of affirmative action in our workplaces) is unconstitutional. I will do so firstly by considering the constitutional framework for legitimate affirmative action, and then by considering how (if at all) the EEA fits in complying with this, with a particular (critical) focus on the organising principle of the Act's affirmative action provisions and on how it pervades the Act's scheme for such measures. I will then, in section 3, address the constitutionality of the EEA in the light of the conceptual and other problems with the relevant parts of the Act. In section 4 I will briefly discuss why the Constitutional Court's judgment in Barnard is so disappointing in the landscape of our equality jurisprudence. I will then, in section 5, briefly include some remarks regarding the recent amendments to the EEA (specifically, for the purposes of the discussion in this piece, those concerning section 42 of the Act). In the concluding section, section 6, I will provide some more reasons why the EEA's affirmative action provisions should be shown the door, and why I believe the adjudication of disputes about the implementation of affirmative action under this Act should not take place within the paradigm of the equality right in our Bill of Rights.

2
Are the affirmative action provisions of the Employment Equity Act unconstitutional?
In our constitutional dispensation, and in the light of its non-negotiable equality guarantee, any differentiation in treatment between groups and individuals on the basis of race and sex (or disability) -the three grounds for the entitlement of suitably qualified members of designated groups under the Act for the application of affirmative action -can be legitimate and legal only if such measures comply with the that principle is not in contention, but its implementation is." (Premhid 2014 http://hsf.org.za/resource-centre/hsf-briefs/barnard-case-a-missed-opportunity). 599 constitutional equality guarantee. 13 The Bill of Rights itself proclaims differentiation on such grounds to be suspect, per se. 14 The mere fact that we are dealing with affirmative action (or, as Moseneke J reminded us in Van Heerden, a more accurate term in our context would be "regstellende aksie" or corrective action 15 ), in itself, is not problematic, bearing in mind that our equality right (in section 9(2)) specifically makes provision for such measures in the pursuit of substantive, as opposed to formal, equality. 16 In fact, the section does more than simply make provision for affirmative action. A substantive notion of equality demands it. So, affirmative action in itself is not bad. But we must remember that any such differential treatment of persons which does not conform to the equality right would be bad. It would amount to discrimination in the meaning of section 9(3), on listed grounds, and then, quite probably, unfair discrimination. I will argue that the Employment Equity Act's affirmative action provisions, very fundamentally, do not comply with the constitutional requirements for legitimate restitutionary measures, and that this could expose employers who implement such policies and programmes in the name (and guise) of the Act to potential claims of unfair discrimination. 13 Contained in s 9 of the Bill of Rights in the Constitution, which reads as follows: "Section 9: Equality (1) Everyone is equal before the law and has the right to equal protection and benefit of the law.
(2) Equality includes the full and equal enjoyment of all rights and freedoms. To promote the achievement of equality, legislative and other measures designed to protect or advance persons, or categories of persons, disadvantaged by unfair discrimination may be taken.
(3) The state may not unfairly discriminate directly or indirectly against anyone on one or more grounds, including race, gender, sex, pregnancy, marital status, ethnic or social origin, colour, sexual orientation, age, disability, religion, conscience, belief, culture, language and birth. (4) No person may unfairly discriminate directly or indirectly against anyone on one or more grounds in terms of subsection (3). National legislation must be enacted to prevent or prohibit unfair discrimination. (5) Discrimination on one or more of the grounds listed in subsection (3) is unfair unless it is established that the discrimination is fair."

The constitutional licence to pursue affirmative action (and its limits)
In order to set the frame of reference for the enquiry to follow, we must ask: What then does the Constitution demand of legitimate affirmative action? It is here where we are confronted with two possibilities. The first is based in a landmark unfair discrimination case, Harksen v Lane 17 (which did not involve affirmative action, but is highly relevant to the testing of whether a purported affirmative action measure violates the equality clause, as the SCA explained in its judgment in Barnard 18 ). The second was expounded in an affirmative action case -the above-mentioned Minister of Finance v Van Heerden. There is a significant difference between the two approaches, which relates to the standard of review of affirmative action measures: the Harksen approach has come to be called the "fairness test"; the Van Heerden approach has come to be called the "rationality test". 19 More will be said below about these differences, but for now the point is that any affirmative action measure (or legislative or other instrument which mandates the application of affirmative action) must satisfy the requirements of either or both of these approaches to the differential treatment of persons on grounds listed as suspect in the equality right, in the sense of constituting (or leading to a presumption of the existence of) unfair discrimination. 17 Harksen v Lane 1998 1 SA 300 (CC) (hereinafter Harksen). 18 In Solidarity obo Barnard v SAPS 2014 2 SA 1 (SCA) para 50 (hereafter Barnard (SCA)), where the court (by way of Navsa ADP) held as follows: "The starting point for enquiries of the kind under consideration is to determine whether the conduct complained of constitutes discrimination and, if so, to proceed to determine whether it is unfair. When a measure is challenged as violating the Constitution's equality clause, its defender could meet the claim by showing that it was adopted to promote the achievement of equality as contemplated by s 9(2), and was designed to protect and advance persons disadvantaged by prior unfair discrimination. Similarly, as stated above, s 11 of the EEA provides that whenever unfair discrimination is alleged, the employer against whom the allegations is made must establish that it is fair." The court referred, as authority for this proposition, to Harksen paras 43-46. The CC, in Van Heerden, by implication rejected the application of the Harksen test in affirmative action cases, on the grounds that affirmative action under s 9(2) of the equality clause is not presumptively unfair or subject to strict scrutiny review.
While the Harksen test may not apply directly to s 9(2), it is suggested that its role in this context should be carefully reconsidered. It is submitted that it should apply in any case where a complainant alleges that what purports to be an affirmative action measure does not satisfy the constitutional requirements for such a measure, and thus unfairly discriminates against him or her in terms of s 9(3 Before considering the scheme and approach of the EEA to affirmative action, let's consider what those requirements are in terms of the above case law. Undeniably, the least strict of the above two approaches is that found in Van Heerden's case (the rationality test). According to the majority of the court (per Moseneke J) in this case, a legitimate affirmative action programme or measure must pass the internal test set in section 9(2) for the legitimacy of affirmative action as a means to pursue the redress of past disadvantage in order to achieve substantive equality. This test requires three distinct things of such an affirmative action policy or measure: it must target those (groups or persons) previously disadvantaged through unfair discrimination; it must be designed to protect or advance such persons or categories of persons; and it must, ultimately, promote the achievement of (substantive) equality. 20 If the above requirements are met, the effect of such a policy or programme on those who do not benefit from it or are, in fact, intentionally excluded from any such benefits (for example, an able-bodied white male -the only non-designated group under the EEA that is excluded from the benefits of affirmative action) 21 is, apparently, irrelevant.
Such a non-beneficiary will not be able to claim that the relevant policy or programme unfairly discriminates against him, as the relevant policy or programme will be deemed to constitute legitimate affirmative action under section 9(2) and, thus -presumably would determine whether affirmative action is such a special case that it might be completely immune from constitutional scrutiny (beyond the borders of section 9(2)) in respect of its fairness. Reading through some of the literature it sometimes appears as if some commentators (and judges) believe this to be the case, but I would submit that that is not what was said in Van Heerden. Moseneke J pointed out that the use of such measures is an integral part of the pursuit of substantive equality (the full and equal enjoyment of all rights guaranteed in section 9(2)), and that this means that there could be no presumption of unfairness (in terms of section 9(5) read with section 9(3)) when such a policy or programme discriminates on a listed ground: Legislative and other measures that properly fall within the requirements of section 9(2) are not presumptively unfair. Remedial measures are not a derogation from, but a substantive and composite part of, the equality protection envisaged by the provisions of section 9 and of the Constitution as a whole. Their primary object is to promote the achievement of equality. To that end, differentiation aimed at protecting or advancing persons disadvantaged by unfair discrimination is warranted provided the measures are shown to conform to the internal test set by section 9(2). 24 It is important to note that this does not mean that affirmative action cannot be unfair, only that it will not be presumed to be so in terms of the equality provisions: [I]f restitutionary measures, even based on any of the grounds of discrimination listed in section 9(3), pass muster under section 9(2), they cannot be presumed to be unfairly discriminatory … If a measure properly falls within the ambit of section 9(2) it does not constitute unfair discrimination. However, if the measure does not fall within section 9(2), and it constitutes discrimination on a prohibited ground, it will be necessary to resort to the Harksen test in order to ascertain whether the measures offend the anti-discrimination prohibition in section 9(3). 25 More will be said about fairness and Van Heerden (and Barnard (CC)) later. Instrumentalist deference typically absolutises specific socio-political goals and disconnects them from their historically contingent and contextually relative settings. Such an approach flies in the face of the fact that in reality no social good is pursued in a space devoid of competing interests and to treat them as such would be tantamount to judging the constitutionality of measures designed to promote such goods in terms of their own stated objectives only. This would of course result in no meaningful constitutional scrutiny at all, since such measures would in effect be constitutionally self-justifying. 28 This last is aptly illustrated by the Labour Appeal Court's judgment in Barnard, which one observer has characterised (for this very reason) as "nothing but a dreadful miscarriage of justice". 29 Instead, the fairness approach in terms of Harksen would require a remedial measure to be both proportional and fair. 30 In terms of this approach, the previously advantaged complainant is treated as more than just collateral damage in the pursuit of our constitutional idyll of a substantively equal society. Such a complainant is not treated as a means to an end, but as an end in 26 See Harksen paras 46, 49, 50 et seq. 27 Pretorius 2013 SALJ 40. 28 Pretorius 2010 SAJHR 554. 29 Malan 2014 De Jure 125. 30 Pretorius 2010 SAJHR 539 argues that what was said in Harksen para 51 suggests that the remedial or restitutionary nature or purpose of the differential measure does not place it beyond the reach of the right against unfair discrimination, since the remedial objective is firmly imbedded within the unfair discrimination inquiry itself: "The remedial purpose is integrated, as a contextual factor, with the substantive equality-based unfair discrimination analysis as a whole. [I]t is important to ensure that the process of achieving equity is conducted in such a way that the baby of non-racialism is not thrown out with the bath-water of remedial action. Thus while I concur fully with Moseneke J that it would be illogical to permit a presumption of unfairness derived from section 9(3) (read with section 9(5)), to undermine and vitiate affirmative action programmes clearly authorised by section 9(2), by the same token I believe it would be illogical to say that unfair discrimination by the state is permissible provided that it takes place under section 9(2 whole. Worse, though, it holds very serious implications for those persons who may find themselves disadvantaged by affirmative action measures (including both the "previously advantaged" and that more sympathetic category of such nonbeneficiaries, those "less previously disadvantaged than others" -the minority designated groups under the Act). Even though our courts have generally rejected such differentiation between previously disadvantaged groups, in this farmyard some animals are definitely more equal than others, 36 if just for the reason that they happen to be more numerous than others. More will be said on this later, when we will encounter the fascinating world of demographics.
However, for the present purposes, and while we unfortunately still (even following Barnard (CC) 37 ) await the final word from the Constitutional Court on which approach is to be followed in future, let's assume that, at the very least, affirmative action must pass the rationality test under Van Heerden before it can be earmarked as constitutional (and, thus, legitimate). And I believe that the following definition (or description) of affirmative action, as per Shaik AJ in Naidoo v Minister of Safety and Security, does well to incorporate the elements for such a programme as contained in section 9(2) and identified by Moseneke J in van Heerden: The essence of affirmative action is to differentiate and to prefer a member of a designated group in order to promote and attain substantive equality. Its purpose is to redress the effects of past discrimination and to end discrimination, and by these means to promote equality In s 4 of the text below I will briefly revisit the Van Heerden "rationality test", academic criticism of it, and its treatment in Barnard (CC).

The problems with the Employment Equity Act
So … enter the Employment Equity Act. I will briefly test its most important provisions which deal with affirmative action (and, in fact, the whole scheme of the Act's affirmative action provisions as contained in Chapter III of the Act) against the Constitution, and more specifically, the rationality test under Van Heerden. I will submit that the Act is in fact, unconstitutional in its approach to the issue. In order to explain this view, I will consider the wording of the Act and briefly hold up the relevant provisions (and the form of affirmative action that it establishes and sanctions) to the mirror of the requirements set in Van Heerden (which I view to constitute the lowest bar 40 which a constitutionally-sanctioned affirmative action programme must be able to cross 41 ).
The most important aspect of the Act that will be highlighted here is one that has received surprisingly little -if any of real significance -attention from (labour law) commentators to date. 42 This is the fact that the Act introduced a new, and completely alien, concept into the scheme of the legitimate objective(s) of affirmative action, which is nowhere to be found in the constitutional equality guarantee or elsewhere in the Bill of Rights. And this new objective has come to colour the design and implementation of affirmative action plans (especially in public sector employmentviz the Barnard and other such litigation 43 ) as well as the approach to the adjudication 40 The Constitutional Court agrees: see Barnard (CC) paras 39 (per Moseneke ACJ) and 94 (per Cameron J et al), where the rationality standard is described as the "bare minimum requirement" for the constitutionality of an affirmative action measure.

41
As Pretorius puts it: "Although rationality review requires a minimum measure of justification in respect of the legitimacy of the purposes pursued and of ends-means coherence, it relieves the state of the duty to justify actions in two significant respects. Requiring a rational relationship between means and ends is a far less exacting standard than demanding that means should be proportional to ends … Rationality review does not express the same responsiveness to situations where the infringements of rights are unnecessarily intrusive. To the extent that a rights-limiting act can be rational, even if disproportional or unfair, a mere rationality standard of justification The legislature is considered the most fundamental arm of democratic governance. In its purest form it serves to secure the foundations of democracy by translating the will of people into the law of land. At its core, the legislature is a mirror of society's soul. 45 If the Employment Equity Act mirrors South African society's soul, we may be urgently in need of an exorcist. In any event, the EEA is an example of legislation passed specifically in terms of the constitutional instruction to the legislature to actively promote equality. 46 As such, the Act must comply with the constitutional requirements for remedial action in terms of the equality guarantee. I will accordingly briefly examine the provisions of the Act to determine whether this is, in fact, the case.

The mysterious numbers game
Let's start with the Preamble to the Act, which lists six separate objectives in explanation of the need for its promulgation. to promote the constitutional right of equality, and the exercise of true democracy; to eliminate unfair discrimination in employment; to ensure the implementation of employment equity to redress the effects of unfair discrimination; to promote economic development and efficiency in the workforce; and to give effect to the obligations of the Republic as a member of the International Labour Organisation.
These five objectives are relatively uncontroversial. It is, however, in the fourth objective (in the order as listed in the Preamble) that we first encounter a concept that is apparently completely alien to the constitutional equality guarantee and the Bill of Rights as a whole: "to achieve a diverse workforce broadly representative of our people".
The concepts of diversity and representivity, 47 which we encounter for the first time here, are thus elevated to the status of fundamental justification grounds for the Act, and these concepts (especially representivity) then pop up consistently throughout the Act, and especially in its Chapter 3, which deals with affirmative action. One of the most prominent examples is found in what must certainly be one of the most important provisions in the scheme of this legislation, section 2, which explains the purpose of the Act. This is not some "empty lip service to grand ideals"-type of provision which is untethered to the mechanics of how the Act actually works or how employers implement affirmative action in terms of it. Remember that section 6(2) -another pivotal provision in the Act -provides that an employer may defend a claim of unfair discrimination under the prohibition of unfair discrimination contained in section 6(1) by showing that differential treatment occurred in terms of affirmative action 47 If any reader had seen this article in draft form, s/he would have noted that it was riddled with alerts from my word processor's spell-check function. The reason is, of course, that the word "representivity" is not part of the English language. A Google search reveals that the wordnik.com website does mention it, and contains 10 examples of its use. All 10 of these examples derive from use by the African National Congress (ANC) in official documents or media statements. It appears that this word is, in fact, a creation of the ANC, and it is, of course, a pivotal component of its ideology of "demographic representivity" (as discussed in more detail in the text below). I will continue to use the word with this spelling in this piece. When in Rome … "consistent with the purpose of this Act". It is then, within this scheme of justification for unfair discrimination, very interesting to consider the wording of section 2, which provides as follows: Section 2: Purpose of the Act The purpose of the Act is to achieve equity in the workplace by - Not only does section 2 incorporate the achievement of representivity as one of the purposes of the Act, but it does so in a truly surprising way. We have seen that the constitutional imprimatur for affirmative action has been unequivocally stated by the Constitutional Court to be one that requires the pursuit of a "remedial or restitutionary equality". According to Van Heerden, what section 9(2) requires is that affirmative action must have the objective of (and must be designed to achieve) the redress of past disadvantage. But here, in section 2 of the Act, the drafters of the EEA tell us that the redress of past disadvantage is apparently just a means to another end! The purpose of the Act is to implement affirmative action to redress disadvantage, in order to ensure the equitable representation of members of designated groups in the workplace. And this is truly strange, for a number of reasons. provisions regarding South African workplaces. And it should be noted that these provisions regarding certain institutions in the public service refer to the objective of making the relevant institutions "broadly representative" (ie to "reflect broadly the racial and gender composition of South Africa"). This is something very different from the application of targets based on the demographic representation of different groups within the population 51 -as has become the norm in affirmative action target-setting in many workplaces, most notably in most (if not all) of the public service. There is no sign in the above-quoted provisions of the Constitution of the race-based (or is it just racist?) lunacy of an SA Police Service "equity" plan which concerns itself with the representation of Indian women calculated to the third decimal (as per the example that presented in the case of Naidoo v Minister of Safety and Security 52 ).
Secondly, when one considers that section 3(a) of the EEA provides that the Act must be interpreted "in compliance with the Constitution", it is unclear how this concept of equitable representation has snuck into the Act in the first place, or how one should consider its role and importance in vacuo in the light of its having been parachuted into the Act with apparently no constitutional connection whatsoever. 53 Section 3(a) would, to my mind, demand that the "equitable representation" standard as employed in the Act must be tested in respect of its compatibility with the provisions of section 9(2) and the Constitutional Court's emphasis on the objective of the redress of past objectivity, fairness, and the need to redress the imbalances of the past to achieve broad representation." 51 As recognised by Shaik AJ in Naidoo paras 131-132: "For race, it is said [in the SAPS employment equity plan at issue] the ideal workforce profile and numeric targets be 79:9:8:2 in respect of Africans, white, coloured and Indians, respectively ... It is important to note this construct is at variance with the stated purpose in the [Preamble of the] Equity Act and equity plan, namely to create a workforce that is 'broadly representative of the South African community." 52 Referred to in the text to fn 3 above. disadvantage for any measure aimed at advancing the cause of designated groups.
Again, and in any event, it is quite a conceptual leap from a broadly representative public administration 54 to a requirement of a demographically representative workforce also in the private sector (as has become the norm in the implementation of affirmative action programmes under the EEA -more to be said on this later).
Thirdly, and most importantly (and most fundamentally disturbing), it is in any event unclear how redressing past disadvantage would/can logically lead to ensuring the equitable representation of groups, or vice versa, in any context. There is no apparent link between the representation of any particular group (be it based on race, or gender) and the redressing of past disadvantage suffered by such group. I will revisit the conceptual problems with the interaction between representivity, equity and equality in s 2.2.3 in the text below. For now it bears mentioning that one (and, I would suggest, the only) way in which demographic representation of a group may be linked to the pursuit of (substantive) equality and the redress of past disadvantage is that it may provide evidence which may point towards exclusionary practices or other barriers to the achievement of equality (eg where the significant under-representation of a group in a specific context, when measured against such a group's representation in the population, more generally, might point to the fact that members of such a group may have been or currently are being unfairly excluded from opportunities). It should be noted, however, that such information regarding the representation of a group should be used with extreme care, and with due recognition of the fact that differences between group representation in a specific context (eg a workplace) and demographic statistics reflecting representation in the broader population may at best highlight a likelihood of "something being wrong with this picture" -there are a myriad reasons why any group's representation (in any context) might be different from such a group's demographic representation in the population. More will be said on this in the text below. 56 The closest a court has come to investigating this question, although quite superficially, is in the concurring judgment of van der Westhuizen J in Barnard (CC). The following was said at para 149: "Before focusing specifically on the facts of this case, it must be pointed out that equality can certainly mean more than representivity. Affirmative measures seek to address the fact that some candidates were not afforded the same opportunities as their peers, because of past unfair discrimination on various grounds. By focusing on representivity only, a measure's implementation may thwart other equality concerns. For example, if a population group makes up 2 or 3 percent of the national demographic, then, in an environment with few employees, the numerical target for the group would be very small or even non-existent. If a candidate from this group is not appointed because the small target has already been met, this may unjustly ignore the hardships and disadvantage faced by the candidate or category of persons, not to mention the candidate's possible qualifications, experience and ability." And at para 150: "Although equality can manifest regarding the rationality (and hence, legitimacy) of any affirmative action measure which states as its objective the achievement of (demographic) representation -and this need for rationality was confirmed by Moseneke J in Barnard with reference to the principle of legality. 57 I will say more on this later.
The EEA, in Chapter 3, continues to display a preoccupation (although obsession might be a more apt description, and I'll use it from here on) with the issue of race and gender representivity, which is conspicuously untethered from its constitutional mandate.
-Another very important section of the Act (to be found in its chapter 3, which deals specifically with affirmative action) is section 15. Section 15 (1)  -Then section 15(2)(d)(i) provides that affirmative action measures must also include measures to "ensure the equitable representation of suitably qualified people from designated groups in all occupational categories and levels in the workforce".
in various forms, in the context of this case it takes the form of representivity. By appointing Ms Barnard, her designated group would have been significantly over-represented and her appointment would have aggravated racial inequality." These views will be interrogated more closely in the text below. 57 Barnard (CC) para 39, where the learned judge made the following observations: "As a bare minimum, the principle of legality would require that the implementation of a legitimate restitution measure must be rationally related to the terms and objects of the measure. It must be applied to advance its legitimate purpose and nothing else. Ordinarily, irrational conduct in implementing a lawful project attracts unlawfulness. Therefore, implementation of corrective measures must be rational. Although these are the minimum requirements, it is not necessary to define the standard finally." The culmination of the EEA's obsession with representivity is found in its section that deals specifically with the assessment of compliance with its provisions by designated employers -section 42. It is this section that first merges the concept of demographics into the concept of representivity. Section 42 of the Act, prior to its recent amendment, 60 listed a number of indicators to be used in assessing an employer's compliance with the Act, and specifically the determination of whether designated groups are equitably represented in a specific workplace and the legitimacy of the targets set. These include(d) the following: the demographic profile of the national and regional economically active population; the pool of suitably qualified persons from designated groups from which the employer may reasonably be expected to promote or appoint employees; The amendments to the Act (specifically to s 42) occasioned by the Employment Equity Amendment Act 47 of 2013 will be considered briefly in s 5 in the text below. I would submit that such amendments, in the context of the subject of this piece, are significant.
I will (only briefly) revisit the recent amendments to the Act (more specifically, just to this section) later, 61 and I will also include some discussion on the relevance of the contents of section 42 (in consideration of the conceptual problems with the Act's scheme of affirmative action based on the pursuit of (demographic) representivity). 62 For now, though, the point is that this section provides the only attempt contained within the Act at providing a definition of the concept of '"equitable representation", which is so central to the very purpose of the Act as well as the scheme of its Chapter III. Any designated employer doubtful as to what he or she should do to comply with the Act when designing and implementing an affirmative action measure, therefore, is first and foremost told to refer to demographic statistics regarding the representation of members of designated groups (and of the non-designated group) in the population. The section, as read with sections 19 and 20, especially, thus leads such employers down a clearly demarcated path to where the legislature wants them to arrive. Having, for want of a better term, thus been led down the garden path, it is no surprise that we find the EEA's obsession with demographics -its numbers game -having also become the inevitable obsession of human resource managers across the board. I will say more about section 42 and its implications below. For now though, while the role of regional as opposed to national demographics as a yardstick to assess compliance with the Act by designated employers has been the subject of litigation, 63 I am not aware of a single challenge to date regarding the legitimacy, more fundamentally, of the very concept of demographics in this context. Ms Barnard failed to persist in any such argument before the Constitutional Court, and we will have to wait (although, hopefully not another 10 years or so) before our highest court might again be charged with considering these issues.
We have all been faced with the EEA for so long now that all of what was said above about the contents of the Act must seem to be just a rehashing of common knowledge, and thus might lead to the assumption that it is also quite uncontroversial ( why have we not questioned this fundamental aspect of the Act more closely and critically to date? It is clear that the Act is simply riddled with this notion of (demographic) representation of groups. But where does it come from? And, more importantly, why is it there? I hope to return to these questions later.

The deafening silence over the mysterious numbers game
The EEA's astounding obsession with race and gender representivity (and with demographic statistics) -its "numbers game" approach -has to my knowledge never [The Employment Equity Act's] concern is not with disadvantage, but with racial representativeness, which it uses as its organising concept. Since demographic testing of this sort can find no justification in the Constitution, the Act can be rescued only if representativeness is considered to be a legitimate proxy for past disadvantage. To prove this, the court will need to be satisfied that no reasonable alternative exists by which past disadvantage might be tested directly. That is not an easy conclusion to reach … In the sphere of employment … degrees of disadvantage can be assessed in the course of appointing or promoting a person. Such an individualised assessment is, broadly speaking, what the US Constitution expects of a legitimate affirmative action programme within the employment sphere, and the same is, arguable, true of ours. 65 Rycroft (in a piece published very shortly after the enactment of the EEA) seems to have foreseen the probability that the Act's drastic remove from the constitutional scheme for remedial action would lead to future uncertainty regarding its application: It is to be noted that whilst s 9(2) of the Constitution sanctions legislative and other measures designed to protect or advance persons, or categories of persons, 64 Although this highly germane aspect of the Act's organising principle when it comes to affirmative action has been addressed in respect of the more general trend of the large-scale legislative embedding of representivity in post-apartheid South Africa, which has been the subject of an indepth and highly critical piece (which will be referred to extensively in the text that follows) -see  TSAR. 65 Brassey 1998 ILJ 1363. disadvantaged by unfair discrimination , the target in the Employment Equity Act is not explicitly the concept of disadvantage but 'designated groups', defined not by disadvantage but by race, gender and disability. There is thus a moot constitutional point as to whether the Employment Equity Act is tailored narrowly enough to meet the declared constitutional purpose that affirmative action measures must be "designed" to protect or advance persons, or categories of persons, disadvantaged by unfair discrimination. Whilst an affirmative action policy will ordinarily constitute discrimination, it is saved from being unfair discrimination because of the constitutional and legislative mandate. But precisely because these provisions are seen by many as an exception to the right equality and the prohibition against unfair discrimination, they will be revisited many times in the coming years. 66 Grogan, in his popular labour law textbook, commences the chapter on unfair discrimination with the following sentence: The Employment Equity Act 55 of 1998 aims to correct the demographic imbalance in the nation's workforce by compelling employers to remove barriers to advancement of "blacks", "coloureds", "Indians", women and the disabled, and actively to advance them in all categories of employment by "affirmative action". 67 [My emphasis] In the chapter on affirmative action, Grogan declares in the first paragraph that "[the goal of the EEA] is to ensure that the nation's working population reflects the demographics of the population as a whole". 68 Basson et al, in Essential Employment Discrimination Law, also grasp the apparent object of the Act and come to the point in describing it, quite succinctly, when they declare that "affirmative action, according to the EEA, is primarily about numbers". 69 Mushariwa tells us -without more -that "[r]epresentivity is crucial". 70 Of course, these works are aimed mostly at a rather general overview of the most important principles of our labour law, and as such are not ideal vehicles for in-depth (and possibly critical) analysis of some (controversial) concepts. I do still wonder, however, whether law students and the other targeted readers of texts such as these do not deserve some deeper interrogation of the 66 Rycroft 1999 ILJ 1413-1414. 67 Grogan Workplace Law 94. 68 Grogan Workplace Law 121. 69 Dupper et al Essential Employment Discrimination Law 259. Mushariwa also does not interrogate this issue of representivity, when she states that "in order for an affirmative action policy to be of a standard capable of withstanding constitutional scrutiny, it needs to set clear targets of how it will achieve a representative workforce within a particular employer setting … The aim of affirmative action is to create a representative workforce with respect to race, gender and disability, yet there is a clear focus in [the case of UNISA v Reynhardt] on race alone. Affirmative action is not just about race. The quest for representivity requires an investigation into the gender and disability composition of incumbents also." (Mushariwa 2012 PELJ 413, 416-417). 70 Mushariwa 2011 Obiter 445.
legitimacy of this fundamental characteristic of the Act, or even just a stab at speculation regarding its constitutionality. Marie McGregor, at least, refers to the issue in her discussion of the affirmative action scheme of the EEA: In essence, then, affirmative action measures or employment equity plans strive to achieve "equitable representation" in the workplace. This is differently worded from the constitution, which sets out to promote the achievement of equality and authorising measures which may be taken to assist with achieving equality. The constitution does not use the notions "equitably represented," "equitable representation," or "designated groups," the last of which is defined not by disadvantage, but by race, sex and disability in the Employment Equity Act. Moreover, these notions are found only in the Employment Equity Act. Whether the notion of "equality" may be equated to "equitable representation" is debatable. 71 This same author has also expressed a more definite view on the issue, although only in passing. 72 It is, in fact, necessary to leave our shores if one hopes to find any critical scholarship specifically regarding the legality of the Employment Equity Act, although even this is for the most part rather equivocal in its findings and does not express any Solidarity v Department of Correctional Services 2014 35 ILJ 504 (LC) para 30, where Rabkin-Naicker J observed: "Given that there is no attack on the constitutionality of the EEA and s 6 in particular, I must find on the basis of the jurisprudence of our highest court that affirmative action will be said (in section 4) below regarding "representivity" before the Constitutional Court in Barnard.

A closer reflection on the EEA's numbers game, in its broader (political) 75
context Outside the legal fraternity, however, criticism of the Employment Equity Act or, more pertinently, the particular brand of affirmative action that it sells -and its most controversial characteristic -is a little easier to find. Of course, much of such criticism regularly emanates from amongst the ranks of the usual suspects. While it is inapt, illogical and extremely unfair to automatically label critics of transformational policies as conservative or as being opposed to transformation (or even worse, racist), the perception is probably well-founded in some instances. But some of the more vocal opponents happen to be those with a proven track record, such as veteran journalists with no apparent political platform to promote. One such critic recently published a thought-provoking short piece on the internet regarding the government-driven drive towards demographic representivity in all spheres of our society: Over the past year or so the ANC government has passed a number of laws designed to allow for the more effective enforcement of "demographic representivity" in the economy and professions. This is the principle that all spheres of life in South Africa should be made to conform, at all levels, to the racial (and increasingly now gender) composition of the total economically active population of the country: 74.9% black, 11.3% white, 10.8% Coloured and 3% Indian. This principle, the enforcement of which relies upon the perpetuation of apartheid-era racial categories, is seen to negate the right of any individual from a minority community to equal treatment before the law. It is also seen as justifying the extension of party and state control into all "centres of power" in society -most recently over the legal profession. President Jacob Zuma stated earlier this year that his government's priority now is to measures in conformity with the purposes of the EEA are those that meet the requirement of substantive equality." 75 We should not beat about the bush here; the demographic representivity drive is a political agenda designed and enforced by the ruling party, the ANC, which underpins and pervades government policy. Compare the following, from Public Service Commission Audit of Affirmative Action xii: "[Representivity targets within the public service] have been set within a broader political and policy environment, and therefore have a legitimacy which needs to be adhered to." This "legitimacy" may very well be claimed to be sourced from s 197(1) of the Constitution, which proclaims that the public service "must loyally execute the lawful policies of the government of the day", but I would argue that the lawfulness of such policies is highly suspect in terms of the rest of the Constitution (and especially, of course, the Bill of Rights). ensure that the economy "reflects the demographics of the country" in both management and ownership. The Employment Equity Amendment Act, passed late last year, gives government massively increased leverage to enforce its will on this matter. 76 What stands out in this particular comment on the current lie of the land is its recognition of the insidious nature of such (alleged) social engineering agenda and, when one reads between the lines, its especially dangerous context: criticism of the liberators of apartheid South Africa is possibly still, 20 years into our democracy, rather unpalatable to most. This is often, of course, due to a fallacious assumption, namely that criticism of the current state or conduct of those who opposed apartheid in the past equates to implied approval of the system they opposed. But a greater threat to our Constitution may be the apathy of the ignorant both within and outside our borders: Among many South African intellectuals this principle [of demographic representivity] is regarded as self-evidently natural and just. It is seen as compatible with "nonracialism", synonymous with "substantive equality" and a matter of "distributional" and "economic justice". Opponents of the EE Amendment Act were derided last year as "conservatives", "rank conservatives", "arch conservatives" and "dinosaurs". In the Western media too the enforcement of this principle in South Africa seems to be regarded as wholly unobjectionable, which is why it is seldom reported on (and very rarely critically) ... In a sense it is both surprising and not surprising that a large part of civil society would go along with the idea that a racial majority government is entitled to intervene to limit the share of minorities in all occupations and fields (and property ownership as well) to their percentage of the population. It is surprising as this principle has a deeply nasty history. It is not surprising as if one traces this idea back to the moment it took fire in 1930s Europe it is clear that it exerts a truly terrible power over the modern democratic imagination. 77 And here we find an oblique reference to what may be one reason for the singular lack of engagement with the issues raised here in the mainstream legal literature, the quest for political correctness or a fear of rocking the boat in respect of issues that can raise such emotional debate. This, however, cannot justify the failure to critically 76 Myburgh 2014 http://www.politicsweb.co.za/news-and-analysis/race-quotas-the-terrible-powerof-demographic-repr. 77 Myburgh 2014 http://www.politicsweb.co.za/news-and-analysis/race-quotas-the-terrible-powerof-demographic-repr.
interrogate the nature of this beast, especially in the light of its central role in this new South Africa.
The guidance we require, I would submit, is to be found in a very important critical piece by Malan, which focuses on the trend of the post-apartheid South African legislature to embed (demographic) representivity in legislation. In a piece entitled "Observations on Representivity, Democracy and Homogenisation", 78 the author makes a convincing case for the unconstitutionality of this practice. Nevertheless, it is extremely wide-spread and pervasive; the concept has found its way into at least 47 separate statutes which deal with representivity in the composition of the boards of public bodies (charged with overseeing fields ranging from libraries to quantity surveyors and property valuers, to weather services). 79 In fact, I would suggest that this is indicative of the fact that Malan is absolutely correct when he declares that: ...
[t]he notion of racial and gender representivity has mushroomed into one of the foremost principles in terms of which the public order in South Africa is organised. If transformation has developed into the master concept of our post-1994 public order, representivity is the principal instrument for achieving transformation ... [In South Africa] there is in all probability no other legal principle that is so virulently and unrelentingly pursued. 80 As we have seen, the EEA reeks of it, and we need to interrogate this phenomenon much more closely in considering the constitutionality of the Act.
In the light of the uncertainty arising from the courts' apparent unwillingness to grapple more meaningfully with the concept of representivity and its links, if any, with equality, it is helpful to consider Malan's definition and his evaluation of its purported interaction with equality: Representivity is the norm in terms of which institutions and organised spheres of people are required to be composed in such a manner that they reflect the national population profile, particularly the racial profile of the national population. 81 78 Malan 2010 TSAR. 79 Malan 2010 TSAR note 7 (for a non-exhaustive list of such statutes). 80 Malan 2010 TSAR 427, 432. 81 Malan 2010 TSAR 427. Representivity is viewed as a prerequisite for achieving equality. This is evident from some of the most prominent legislative instruments for the achievement of transformation, namely the Employment Equity Act and the Broad Based Black Economic Empowerment Act. Both these acts use representivity as a pivotal instrument for achieving transformation and both acts assert that they are there to achieve equality. 83 This is, however, a sham. As will be explained later there is no clear link between representivity and equality (and Malan argues that representivity, in fact, leads to inequality  86 Malan's following conclusion regarding the effects of the enforcement of representivity in the sphere of civil society (specifically, in respect of minority communities and their organisations) might seem alarmist, or it might just seem paranoid: "Representivity mops up the organised spheres of minority communities and functions as a scheme through which organised spheres of minorities are transformed into organised spheres for the majority. It establishes a system whereby organised spheres of the minorities are annexed for and put to the service of the majority. Once representivity has taken its course, the erstwhile organised sphere of minority communities will no longer reflect the minority community in question. Neither will it be under the control of members of that minority community. Instead, it will reflect the composition of the national population and will be under the control of the numerically dominant section of the population. In short, representivity then is a strategy of systemic annexation through which minority-organised spheres are dismantled and placed under the control of the majority -a strategy of totalitarian majority domination of not only the state sector, but also of civil society in general. " Malan 2010 TSAR 445. institutions which serve us the public (which, in turn, points to a very real potential for injustice to be visited upon or further exacerbated in respect of those excluded by such policies): Representativeness as a measuring criteria for the success of the transformation of the South African public service, is derived from the White Paper on the Transformation of the Public Service [1995], which states that representativeness "is one of the main foundations of a nonracist, non-sexist and democratic society, and as such is one of the key principles of the new Government". Representativeness needs to be counterbalanced by the rights of everyone in the country, as promoted and protected by those very same public institutions which are supposed to be characterised by their representativeness. 87 But for me the most problematic aspect of its use in the context of equality (where, according to the Constitution, the affirmative action discourse resides) is the extremely tenuous link in logic between representivity and equality, which I hope to explain more clearly in the section that follows.

The conceptual problems with the numbers game 88
Having pointed to the all-pervasive pursuit of the political ideology of demographic representivity in respect of so many aspects of our society (and much more will be said on this below), it remains to consider its validity more critically. In a majoritarian democracy such as ours, such a programme of social engineering may very well be viewed as being untouchable in the light of its apparent majority support at the ballot box. But we have a sovereign Constitution, which contains a Bill of Rights, which in turn contains a fundamental equality guarantee for all the citizens of our country. This may have, at least initially, posed a problem for the architects of this programme of demographic representivity in aligning it to the aims of the ruling party's National Democratic Revolution (or NDR), in full view of the world (and our pesky judges): The NDR charged the ANC with using state power to deracialise the economy. This predisposed the ANC to regard the parastatals as "sites of transformation" where the capacity of the existing order to resist rapid black upward mobility would be at its weakest. In contrast, the "transformation" of the private sector presented more 87 Wessels 2008 Politeia 29.

88
I previously examined these issues in another piece (Louw 2006 SA Merc LJ), and will include some of that discussion here.
problems. For a start, private property was protected by the constitution and the close scrutiny of the government's economic policies by global markets. 89 So, after targeting the public service -an easy target under majority rule -for its highly racialised vision of "transformation" in the new South Africa, those faceless architects were hard-pressed to find a way to both extend its reach into the private domain and also tether this ideology (which is, prima facie, unconstitutional in the light of the foundational value system) to the bedrock of the Constitution. They chose to place it within the equality paradigm (readers will note, in the extracts from government publications and other documents reproduced in the footnotes, the frequent references to equality and to "equity" -the EEA, of course, is a prime example of this last). In doing so, they have exposed this programme to constitutional scrutiny, and I would suggest that our critical consideration of its continued legitimacy is indispensable for the future relevance of our equality jurisprudence. In this section I will attempt to do so not by talking politics 90 but with reference to (what I believe to be) quite simple precepts of logic. I plan to expose the irrationality of the Employment Equity Act's organising principle, the pursuit of (demographic) representivity, when it is considered specifically within this context of the constitutional pursuit of substantive equality. I will do so by focusing on the following aspects: provide clearly for affirmative action to bridge the gap between the great wealth accumulated by the few under apartheid and colonisation and the grinding poverty of the toiling masses for whom life is a constant struggle for mere existence. Since South Africa still has some poor whites, they too will benefit through affirmative action programmes. But the main beneficiaries of affirmative action will, for demographic reasons, be blacks. Failure to narrow the gap between black and white economic circumstances will leave black South Africans hostage to the illegally acquired wealth of white South Africans." (Masemola 1989(Masemola -1990 Colum Hum Rts L Rev 54.) Roger Southall argues that the "racial bargain" was re-negotiated in the late 1990s with the introduction of aggressive black economic empowerment policies (and note, this was also the time of the introduction of the EEA) -see Southall 2007 Transformation.
the lack of a logical (or moral) link between representivity and equality; how the pursuit of demographic representivity blurs the line between illegitimate quotas and legitimate numerical goals in the application of affirmative action measures under the Act; and the fact that the pursuit of demographic representivity in fact contributes to inequality.
Let me take each in turn.

No logical link between representivity and equality
When demographic representivity is used as a proxy for equality (as it is in the EEA - But this is all very far removed from logic. I would submit that no workplace in any sector or industry can ever be totally representative of the whole -unless such representation has actively (and artificially) been engineered to achieve that objective.
The argument linking representivity and equality seems to assume that, in any workplace or context where one selects a number of persons for a particular purpose (eg a sports team), and the group selected contains members from each of the racial groups in the same proportion as such groups are represented in society, this would equate to all such persons (as well as those who were not selected) having been treated equally, and thus fairly, in terms of section 9 of the Constitution. But this is absurd. Such an inference cannot follow naturally merely from the racial make-up of the group. In theory, a number of persons not selected may have been excluded because of unfairly discriminatory selection practices. Does the fact that the eventual demographic make-up of the group is "representative" of society then negate or justify such discriminatory selection? Of course it can't.
Also, surely we cannot rationally expect all occupations and workplaces to mirror the demographic make-up of our population -traditionally, women are more likely to pursue careers in the fashion industry (this is not a sexist stereotype but merely a reflection of reality) 93 and not all persons aspire to be bankers, engineers or insurance salesmen. In fact, international studies conducted on the subject have invariably found this to be accurate -one such study, which examined "proportional representation" of ethnic groups in order to search for some form of equality in the representation of groups in occupations, institutions and income groups, found that "few, if any societies have ever approximated this description". 94 The "representivity" of the group can never be viewed as the proof or consequence of equal, non-discriminatory treatment.
And that is precisely the problem with the use of race-based quotas. Such a policy elevates the group above the individual in the assessment of selection criteria for participation. As such, it holds no guarantees for the promotion of equality and the eradication of unfair discrimination, and serves merely to ensure, reminiscent of tokenism, that the group eventually selected is "representative". This is suspect not only on the basis of logic, but also ethically and constitutionally.
At this point I need to note that I am willing to admit that there is one particular (but rather limited) way in which representivity might be relevant in respect of equality, specifically in respect of benchmarking its achievement.  The reason, I would suggest, why such factors were originally included in s 42 of the EEA (namely, the pool of suitably qualified persons from designated groups from which the employer may reasonably be expected to promote or appoint employees; economic and financial factors relevant to the sector in which the employer operates; the present and anticipated economic and financial circumstances of the employer; the number of present and planned vacancies that exist in the various categories and levels, and the employer's labour turnover). Note, however, that the amendment of the Act by means of the Employment Equity Amendment Act 47 of 2013 has removed these factors from s 42, leaving demographic statistics as the main determinant of an employer's achievement of employment equity as an affirmative action measure. structure and its representation in the greater (economically active) population? It should also be noted that mere reference to "diversity'" in this manner (as an apparent catch-all for the measurement of the achievement of equity -remember: these terms are not distinguished or explained in the Act) is not really helpful, and may in fact be meaningless. 98 And yet, despite all these inaccuracies, uncertainties and obfuscations, the concept of equitable representation (based, as it is so frequently in practice, on population statistics) has become the be-all-and-end-all of affirmative action under the Act. Even the foremost proponent of the Act must admit that both the rationality and proportionality of this system of equitable (demographic) representation, which is so central and fundamental to the scheme of this legislation, must be in question.

As for the ethics and constitutionality of it all, emeritus professor of ethics, Martin
Prozesky, points to the flawed nature of this ideology of demographic representivity.
He explains that the misuse of demographic representivity flows from errors of logic arising from "some muddled thinking". This, he contends, presents in the following ways: 99 98 I have observed previously (see Louw 2006 SA Merc LJ 352) that one must bear in mind that the ideal of achieving diversity is not necessarily an objective one. There is an element of arbitrariness in defining the meaning of this term, which depends heavily on where its proponents are coming from. In the American context, John Kekes (Kekes "Injustice of Affirmative Action" 200) has observed the following: "[A]rbitrariness pervades the attempt to justify preferential treatment by appeal to the benefits of diversity … No effort is made to contribute to diversity by according preferential treatment to religious fundamentalists, anti-feminists, political conservatives, defenders of the desirability of American primacy in international affairs, or to those who advocate research into genetic racial differences. As it now stands, diversity is a code word for individuals or views that find favor with left wing academics." "Diversity", as the justification for affirmative action, it seems, is in the eye of the beholder. The concept or ideal of diversity involves a value judgment which is at least political in nature, and is inspired by the "accepted" public opinion at any given time. Diversity as a yardstick for equity is highly suspect.

99
Such muddled thinking, apparently, riddles government discourse on the subject. Compare the following, from the White Paper 22 (note the misperception/misconstruing of the words of the Constitution iro "broad" representation): "Achieving broad representation: 1.7 The Constitution requires, amongst other things, that the composition of the Public Service be broadly representative of South African Society as a whole and the White Paper on the Transformation of the Public Service sets out a number of specific targets to achieve this. 1.8 In 1995 the composition of the population of South Africa stood at 75% African, 13% White, 9% Coloured and 3% Indian people. Women who comprised just over half (51%) of the entire population were made up of 76% African women, 12% White women, 9% Coloured women and 3.7% Indian women. At this point figures on disability are not very reliable, however, it is generally accepted that people with disabilities comprise 5% of the population. The Constitution requires therefore that the Public Service will strive to reflect these proportions in its staffing in order for it to be representative." Firstly, when those involved fail to distinguish properly between activities where freedom of choice plays a key role and ones where it cannot; secondly, when they disregard the vital role of choice in the workplace, and thirdly, when they apply the logic of activities where choice cannot be a key factor, to activities where it is -like employment. Think of activities where choice is not a key factor -like anything required of us all by law: such as having identity documents ... Since about 75% of our people are black Africans, it is logical to expect that around 75% of all ID documents should be held by those of us who are black Africans. The same goes for school attendance in the years when the law requires it of all our learners. This shows that demographic representivity can logically be expected of any activity that is necessitated by law for all members of the groups concerned. Quotas can then be set both as a just ultimate goal and as a just way of moving towards it. Things are quite different when we are dealing with voluntary activities like careers, religious affiliation, party membership and sport. For example, there is no logical reason to demand that a certain percentage of Catholics, based on national or regional demographics, must come from our previously disadvantaged communities, because people are free to choose in matters of religion, and the results of genuine choice are unpredictable. Maybe about 75% of Catholics will be black Africans, but maybe not. And if not, that does not make the Catholic Church guilty of inequity. Similarly with careers. 100 Prozesky believes that such demographic representivity objectives are also ethically flawed: Though the work we do is obviously affected by the availability of jobs, it is to a significant extent a voluntary matter, so there is no advance way of stipulating what percentage of participants from our ethnic communities, in any given career or workplace, would count as equitable. Yet this is exactly what some of our politicians and others seem to demand when they call for demographic representivity and quotas based on them for the workplace. Their commitment to justice in the workplace is clear ... but are they really thinking about it logically? It seems not. What about ethics? Demanding quotas based on demographic representivity for voluntary activities is ethically flawed because it results in unfair discrimination. For example, it is unfairly discriminatory to make people lose their jobs in the cities and towns where they work because of their skin colour, and be told to go somewhere else where people of their colour are allegedly needed. What is that but outright racism and thus one of the To add to this, I can honestly find little reason for disagreement with Myburgh's following view on demographic representivity: "I do not know whether the principle of 'demographic representivity', in the judiciary or elsewhere, is constitutional or not. What I do know is that it is a morally abhorrent principle, which drags behind it an odious history. What is unsettling and depressing … is that no moral stigma is currently attached to its invocation." (Myburgh 2009 http://www.moneyweb.co.za/ moneyweb-soapbox/demographic-representivitys-nasty-history).
Malan also points to the constitutional implications of such an ideology as the pursuit of demographic representivity (which he calls a programme of homogenisation). 102 He explains that demographic representivity can perform a legitimate role only in a rather limited context, namely where the state deals with certain interests that might be regarded as equal stake common interests of the entire national population, which interests "are not peculiar to a particular community (culturally, linguistically, religiously etc), in contradistinction with the rest, which have a discernibly higher stake than any other community". 103 The position is very different, however, when it comes to the pursuit of demographic representivity in the composition of organisations which deal with the specific interests of (minority) groups, and especially organisations of civil society, particularly workplaces in the private sector. In this context demographic representivity not only makes no logical sense, but it may also constitute a rather sinister mechanism for a majoritarian democratic state to hi-jack the interests of minorities in the interests of homogenisation. 104 In this respect, the pursuit of demographic representivity may hold very significant constitutional implications. And yet this does not seem to bother our Constitutional Court judges. We find implicit (and sometimes explicit) endorsement of the pursuit of (demographic) representivity in a number of the judgments in Barnard (CC). 105 While it can be expected that contentious issues such as this will always be, well, contentious, it is not only disappointing but actually worrying to see the extent of the 102 He explains: "Assimilation and homogenisation programmes are ordinarily construed on the basis of preference for a specific dominant community -usually a majority of the national populationwhich is identified with the idea of national identity or the national community. All non-dominant communities are then forced or manipulated to adopt the character of the dominant community, thus to be assimilated into one homogeneous national community that bears all the characteristics of the dominant majority." Malan 2010 TSAR 435.

103
He explains: "[T]he application of the representivity principle in the case of equal stake common interests also reflects a communitarian perspective of society. This is so because it recognises communities (by requiring representation for them) not only in the case where specific community interests are being dealt with, but also in cases where interests are the same and do not run along community lines. From a communitarian point of view the principle of representivity can therefore be regarded as commendable when it is applied to equal stake common interests. fundamental disagreement between, for example, academic writers and these judges of our highest court: Jafta J refers to the "eloquent judgment of the Labour Appeal Court" 106 in Barnard (LAC), while Malan calls the Labour Appeal Court's judgment a "dreadful miscarriage of justice". 107 While the reader can decide which view he or she subscribes to, I will simply note that at least Malan has examined this issue of the role of representivity and its legitimacy under our Constitution. Is it too much to require our Constitutional Court judges to do the same before basing their decisions on such a flimsy justification as this (which is so potentially fraught with the risk of, in fact, perpetuating injustice)? 108 Finally, I am reminded of the scathing academic criticism that met a recent, controversial scholarly publication 109 regarding the measurement of the achievement of equity (in the meaning of equality/(racial) representivity) at higher education institutions in South Africa. One such critic (a mathematician) of this work noted the following in the process of a critique of the maths used by the authors, which I believe is apt here whilst we are considering this issue of the formulaic application of population statistics in order to construe some form of a notion of "equality" (or equity): Many processes may be required to eliminate injustice and promote more rapid access to better life circumstances. Elimination of injustice cannot be adjudicated by evidence only from a mere calculation. Both the legitimacy and role of any arithmetic have to be firmly clarified. Otherwise the invocation of one or more indices becomes a vehicle of bureaucratic self-gratification, rather than a series of ordinal indicators, each indicative of only one possible objective at a time. This position does not 106 Barnard (CC) para 230. 107 Malan 2014 De Jure 125. 108 This is especially poignant if one considers the concurring judgment of Cameron J, Froneman J and Majiedt AJ. These three judges specifically mentioned that their task was to determine what essentially comes down to, this very issue of the proper role for representivity in respect of an affirmative action measure applied in terms of the Act: "The difficulty Ms Barnard faces is that the Act says '[i]t is not unfair discrimination to take affirmative action measures consistent with the purpose of this Act'. 81 The SAPS defended the National Commissioner's decision on this basis. But, crucially, this defence avails the SAPS only if the National Commissioner's decision is 'consistent with the purpose of [the] Act'. So the Court's task here is to understand what that purpose is, and to determine whether the National Commissioner's decision was consistent with it." (Barnard (CC) para 86). Disappointingly then, these three judges end up basing their finding of the fairness of the National Commissioner's decision primarily on the role of representivity (without engaging with its meaning or with any of the potential inconsistencies and illogicality of its application in this context).

AM LOUW
PER / PELJ 2015(18)3 632 exonerate universities from accountability. It affirms a collective obligation of an examination of conscience in robust debate. However it also claims that true transformation is a matter of the heart and an issue of complexity, which warrants authentic scholarship rather than fumbling mathematical conjuring. 110 Much the same can be said of the drafting of employment equity plans under the EEA and other instruments which regulate "transformation" (especially in the public sector, but even on our sports fields). 111 A colleague recently asked me, if one were to reject these formulas and the juggling of numerical permutations, how one could go about assessing the achievement of the objective(s) of affirmative action? I am not sure that I know the answer. What I do know is that this cold and clinical head-count exercise (the "Excel spreadsheet approach", for want of a better term) is not the way to go; not under our Constitution. As O'Regan J observed, "the broad goals of transformation can be achieved in a myriad of ways. There is not one simple formula for transformation". 112 Surely a more nuanced approach is both possible and feasible, one that could still, to an extent, employ race as an important denominator for the allocation of preference (even in the light of the problems with race and its continued use by the legislature), 113 but which would not reduce persons to numbers and which could focus on other, more germane factors (such as personal circumstances, socioeconomic conditions, the real need for protection and advancement, etc). It is, after all, not necessarily true that formulas need to be used, divorced from the reality and circumstances of the scenario in which affirmative measures are applied. In the employment context, specifically, there may well be little reason to require 110 Dunne 2014 SAJS 6.

111
The South African Rugby Union (SARU)'s latest transformation plan, which is being enforced in rugby, seeks to ensure that by 2019 at least half the Springbok (national men's) side consists of players of colour, with 60% of those required to be black African. Moves are currently afoot to pose a legal challenge to such transformation efforts, which its critics claim make use of illegitimate race quotas. I have previously written about the fact that the application of such measures in professional rugby is fraught with risks, seeing that professional rugby players in South Africa qualify as employees under the labour legislation. Accordingly, the EEA, and its prohibition of quotas (as recently confirmed by the Constitutional Court in Barnard (CC)) applies to any such programmes or measures. Those pushing the transformation agenda in professional sport (including, and especially, government's department of sport, Sport and Recreation South Africa) have never satisfactorily dealt with this objection in respect of the use of racial quotas in professional sports teams (which, in any event, offends a number of fundamental principles of international sports law -see the articles referred to above in fn 8 consideration of the application of affirmative measures in the abstract (and based on race-based and numbers-driven assumptions). As Brassey has observed, 114 when it comes to appointments and promotion and the like in the workplace, the opportunity must surely always be there to apply such measures with specific consideration of the circumstances, history, etc of the relevant applicant. None of this needs to happen in vacuo (and within the very cold embrace of impersonal tick-boxes and numerical permutations). This, I would suggest, highlights the apparent irrationality and arbitrary nature of the EEA's scheme of affirmative action. Whatever the practicalities, at its heart the preferred approach to affirmative action should not be aimed at ensuring the achievement of that most artificial of constructs, an ideal of a racially representative workforce in every single one of our workplaces. We require a "concrete, contextualised approach" to the various issues that abound when we are faced with disadvantage and inequality, rather than a "formulaic, mechanistic approach". 115 As it has been put, quite succinctly: "Equality is simply too complex to be achieved by means of a calculator and a simple racial ideology." 116 defined in the Act, and one must assume that the words should bear their usual meaning. "Equitable" is defined by Dictionary.com as "characterised by equity or fairness; just and right; fair; reasonable". "Equity", in turn, is defined as "the quality of being fair or impartial; fairness; impartiality". Accordingly, the meaning envisaged by the EEA for the term "the equitable representation of designated groups" appears at first glance to be rather clear, If we decide that the mere equal and fair treatment of all persons is not sufficient to ensure equity, we have already made a value judgement as to what equity means, in anticipando. We have in fact decided that equality of treatment is not enough;

Representivity turns (legitimate) numerical goals into (illegitimate) quotas
something more is needed. We have already decided that, in order to address the existing inequalities through a substantive view of equality, we must for example treat Africans preferentially in order to ensure an outcome for them in the workplace that is 'equal' to that enjoyed by white males, for example. Therefore, we have succeeded in removing the determination of equitableness from the specific workplace and the (reasons for the) existing representation of each of the groups, and have decided to rather base our judgement on an external standard, which is coloured by a value judgement as to which facts are relevant and should enjoy preference in this evaluation of what is equitable.
In this light, even though our courts have expressed a clear preference for numerical goals as opposed to rigid quotas, such goals are themselves not immune from scrutiny.
They are not mandatory, but appear to be guidelines based on a preconceived notion of what would be "equitable". In this we find a paradox. Surely this goal is no longer a guideline and must in fact function as a mandatory quota if we have already determined it to be the be-all-and-end-all of equity? Any failure to achieve this target will fall short of achieving the equitable representation of the group in question. Also, goals are distinguished from quotas as being grounded in reality and the due consideration of objectively verifiable facts that influence the meaning of what is equitable in the circumstances of any given case. But how can this be, when such a meaning has in fact been predetermined in the motivation of the very target-setting exercise, as a product of a value judgement based on the objective of the equality of outcomes? In reality, and on this reading of the objective of chapter III of the EEA, there may not be much difference between numerical goals and quotas. The SCA recognised this in its judgment in Barnard, in warning against the rigid application of numerical goals based on demographics 120 (although the judges of the Constitutional Court seemed less concerned).
By way of summary, therefore, a strong objection to the setting of numerical goals based on demographics is the following: (1) employers may legitimately apply affirmative action only in cases where a group is not equitably represented in the workplace; (2) this determination must be made on some reasonable and rational basis -what is the basis for a finding that the existing representation of such a group is not equitable?
(3) only once inequitable representation is established may steps be taken to address this, which should proceed from the basis of the setting of a goal for the representation of such a group that would in fact be equitable; (4) as such a goal must itself be equitable in order to address the existing inequality, the setting of the goal must itself also involve a rational and reasonable exercise; (5) this can be achieved only by considering objectively verifiable facts, as opposed to a value judgement; (6) the EEA's preference for an objective of equality of the outcomes imposes exactly such a value judgement as to which facts should sway the scales, at the very outset in the target-setting exercise.
So, are we any closer to identifying the practical difference between numerical goals and quotas? Let's consider an example. If an employer employs 100 persons in the workplace, of whom three persons are African, this group is clearly underrepresented in terms of national and regional demographics. If the regional demographics of the province within which the employer operates (eg the Western Cape) shows that 20% of the economically active population is African, the employer may legitimately set a numerical goal of 20 Africans to be employed by a certain future date, and affirmative action measures implemented in order to achieve such a goal would also be acceptable. However, should it transpire that the pool of suitably qualified Africans applying for vacancies during the period in question is very limited (eg only five Africans apply), or that the employer's labour turnover is very low (eg only three vacancies occur), an aggressive pursuit of such a goal that eventually culminates in the appointment of 20 Africans could never be justified. Obviously, this would constitute illegitimate measures reminiscent of tokenism, or the appointment of persons from a certain group for the sole purpose of achieving a set goal for their representation in the workplace, with no regard for the actual circumstances of such workplace and the requirements of the job in question.
So we see that the deciding factor is not only the goal set or the criteria considered in the target-setting exercise, but also the employer's conduct in pursuit of the goal. Had the circumstances been conducive for achieving the goal, which we have seen appears to be perfectly legitimate, there could be no valid criticism of an affirmative action measure employed in this regard. However, once it becomes clear that the goal or its achievement is not realistic in the light of specific circumstances, which may be beyond the control of the employer, the employer's conduct in implementing affirmative action measures in pursuit of the otherwise legitimate goal becomes illegitimate. The goal, in being accorded supremacy over the realities of the workplace, has now come to function as a quota.
In Please note that I am referring here to the content of s 42 of the Act prior to its recent amendment. I will briefly discuss the amendments to this section of the Act in s 5 in the text below.
conflict between the terms "under-representation" and "equitable representation" employed in Chapter III. Under-representation of a designated group as a result of one or more of the job-or workplace-related factors (eg a limited pool or, low labour turnover) can surely not be marked as inequitable, if the reason for such underrepresentation is not due to inequitable treatment or discrimination by an employer.
And why should under-representation in terms of the first of the factors mentioned in section 41(a)(i) (the national or regional demographics) necessarily be viewed as inequitable? As stated earlier, we surely cannot rationally expect all occupations and workplaces to mirror the demographic make-up of our population. The following is surely logically sound: [T]here is no prima facie reason to suppose that members of different racial and ethnic minorities would be equally likely to want to go into [a specific profession] and, on the contrary, many reasons to expect that they would not. To the extent that ethnic and racial groups form at least partially self-contained communities (and they do), members of one community will value different sorts of character traits, encourage the acquisition of different skills, and have different ideas about what sorts of jobs carry the most prestige. Most arguments in favour of affirmative action in fact suppose that racial and ethnic groups differ in these sorts of ways; if they did not then bringing in a wider variety of such groups would not contribute to diversity. 122 We need to urgently reassess the real value of target setting based on demographics and the EEA's extremely crude head-count approach to the achievement of "equity" in employment, if only on the basis of logic. 123 One of the distinguishing features of 122 Wolf-Devine "Proportional Representation" 136. 123 Van Wyk and Hofmeyer 1997 SAJLR 6: "[M]any politicians and trade unions seem to argue that the workforce composition of an employer will be equitable only once it reflects or approaches the proportions found in the general population. The reasoning is straightforward: but for apartheid, the constituent parts of our population would have had equal absorption rates into the labour market … The historical justification is of course a counterfactual hypothesis and there is no way of testing its historical accuracy, in the sense that one could accurately estimate, after having factored out historical discrimination, 'equitable proportions' … This is not to deny the pernicious effects of historical discrimination -it only serves to point out that the counterfactual historical construction is not the full story, and that proportionality may not be the appropriate or best employment equity standard (or that it should be adjusted to take into account other factors that could have caused disproportionate representation in society even if our society had been characterized from its inception by the absence of discrimination) … [H]ead count analysis assumes that one factor, and one factor only, namely past discrimination, explains the disproportional representation in the present labour force. However, perhaps other factors, such as certain demographic characteristics of a particular group, should also be included as explanatory variables, while acknowledging that past discrimination in all likelihood plays the dominant part in our present situation … [W]e should question the unreflective assumption that a head count analysis is unchallengeably the only way to justify the crucial task of promoting greater social justice." quotas is that they are not based on realities regarding the workplace, the job or the labour market. In short, while the EEA purports to pursue the objective of the "equitable representation" of designated groups, quotas emphasise the representation of such groups whereas numerical goals (appear to) focus on the equitableness of such representation in the light of these surrounding realities. The one is ostensibly reasonable, realistic and morally justifiable as a means of addressing the legacies of institutional discrimination. The other has no moral justification. It is easy to identify the one most in line with the underlying ethos of our Constitution, but less easy to distinguish a real difference between the two.
What is quite worrying is the fact that, even though some of the judges in Barnard course, the use of demographics is truly aimed at promoting equality) -because it in fact leads to inequality. Malan has done so, by explaining why there is no merit in the view that equality remains unattained as long as there is not full compliance with the representivity principle: [R]epresentivity does not promote equality: it instead entrenches systemic inequality. Representivity denies minorities the organised spheres that they need as the infrastructure to stabilise and organise themselves. It affords such organised spheres exclusively to the majority. This … is the effect of annexation and the homogenising effects of representivity. Representivity enables the majority to capture the organised spheres of the minorities. All organised spheres are therefore those of the majority, while owing to representivity there is nothing left for the minorities. The majority can use all these organised spheres to promote their interests and to deny the interests of the minorities. The minorities have none of this. In accordance with the representivity principle individual members of minority communities are present in all organised spheres, but once again in accordance with representivity they are but a small outnumbered minority in each. No organised sphere is theirs, where they can fully exercise community identity with their fellow community members and no organised sphere is under their control. In this way representivity acts as a strategy of entrenching the totalitarian control of the majority -the tyranny of the majorityin the famous words of Alexis de Tocqueville. 126 In the light of the Constitutional Court's persistent and strident rejection of formal equality in favour of substantive equality (in cases such as Van Heerden and others), it is interesting to note that Malan argues that the representivity principle in fact is much more in line with the paradigm of the former than with that of the latter: The closest this gets to equality is equality in the formal sense, which is no less than a flimsy cover-up for systemic inequality and repression of minorities. It works like this: the same principle, namely that of national representivity, is applicable to all. However, since various people and groups find themselves in different positions they are differently -sometimes vastly differently -affected by the application of the very same principle. This is in fact what is occurring with the application of representivity. It is highly beneficial to the majority and glaringly disadvantageous to the minorities. It institutionalises a system of majority domination and minority repression. Hence, representivity is not a strategy of equality but a strategy of systemic substantive inequality to the detriment of the minorities. In fact, the working of representivity is a textbook example of the distinction between formal and substantive equality and how harmful the effects of formal equality can be. 127

126
Malan 2010 TSAR 447. 127 Malan 2010 TSAR 447. Pretorius also makes the point (with reference to the work of Sandra Fredman) that a formal notion of equality ignores the need for the Constitution to deal with differences: "[S]ince formal equality in principle commits to an individualist uniformity or sameness of treatment, it is conceptually averse to forms of differential treatment necessary to affirm or accommodate difference. Understanding equality in this way could result in 'collapsing the principle of equality into one of sameness, devaluing difference and endorsing assimilation and conformity'.
Dave Stewart, of the FW de Klerk foundation (yes, another of the "usual suspects" who persistently criticise the drive for representivity) made the point of the singular inefficacy (or even counter-productive nature) of demographic representivity in more practical terms in a speech delivered a week after the CC's judgment in Barnard: The approach of those who support demographic representivity is based on fundamental fallacies. The first fallacy is that the measures adopted in terms of section 9(2) have promoted equality. They have not. South Africa is now a more unequal society than it was in 1994. The overall GINI coefficient has increased not only nationally -but within all our communities as well. Affirmative action measures have done nothing to advance the equality of the most disadvantaged 85% of the black population. 128 In the light of these and other strident criticisms expressed by the opponents of the ideology of demographic representivity, it is even more surprising to note the deafening silence of our courts in respect of the critical interrogation of the role and legitimacy of its pursuit (even if just to refute such views regarding its potential effects on the achievement of (substantive) equality). One should find it extremely disconcerting, for example, that the unequivocal finding by Cameron J, Froneman J and Majiedt AJ that the SAPS measure applied in Barnard was not unfair towards the complainant was based expressly on the apparent legitimacy of the pursuit of racial representivity within the service: We conclude that the facts show that the National Commissioner's decision passes the fairness standard. While we find this a close call, what has proved determinative to us is the pronounced over-representation of white women at the salary level to which Ms Barnard was applying. This was not just by one or two, but by many. There was thus greater justification for prioritising racial representivity over other considerations. 129 Therefore, one of the central features of the Constitutional Court's equality jurisprudence is the 'recognition of difference as a positive feature of society'". See Pretorius 2013 PELJ 289. It is submitted that the ideology of demographic representivity envisages just this, the establishment of a "sameness" which offends against the Constitution's calls for the pursuit of substantive equality (its guarantee of equality before the law for all -including minorities). It also, of course, offends against the notion of embracing diversity (which is itself referred to in the Preamble of the EEA).
Pretorius, in this same piece (Pretorius 2013 PELJ 290), also observes that "[t]he right and value of equality, when informed by the imperative of affording all persons equal respect and concern, functions as a bulwark against enforced uniformity" (and see his reference to National Coalition for Gay and Lesbian Equality v Minister of Justice 1999 1 SA 6 (CC) para 132). 128 Stewart 2014 http://www.biznews.com/undictated/2014/09/11/stewart-affirmative-actionconcourt-judgement-threatens-rainbow-nation/.
Pared down to its essence, representiviy in this context concerns itself with nothing more than the "statistical racial balance" 130 of the workforce, surely a concept that is at least in theory antithetical to the underlying value system of our Constitution. 131 This treats human beings (who have rights under our Constitution) as little more than numbers. And it is a patently arbitrary basis on which to distinguish or prefer persons for any purpose (but also in employment). When one considers the grounds of unfair discrimination listed in section 9(3) of the Bill of Rights it is clear that the demographic representation of one's race group, for example, in the general population would fit in well amongst these. It is a number, a statistic, and it is no less arbitrary a ground for the differentiation between individuals. When considering its treatment in Barnard (CC), it is even more worrying that Jafta J in this same case expressed his approval of race-based job reservation -something found in our recent past and which many would have hoped would not survive our transition to a constitutional democracy -on this very same ground of justification: representivity (and its apparent although never explained fairness). 132 This echoed the similar affirmation of the claimed constitutionality of the practice by Mlambo J in Barnard (LAC). 133 130 In the words of Pretorius in his evaluation of the Labour Court judgment in Barnard -see Pretorius 2013 SALJ 37. 131 Even though some declare that this is to be found in the Constitution, this is simply not true. Compare the approach of Mlambo J in Barnard (LAC) para 25: "The contextual importance in this case is the reality with which the appellant was confronted as to how the designated and nondesignated groups were represented in its workforce. In this regard it is common cause that white employees were overrepresented in level 9. It is also common cause that appellant's Employment Equity Plan was cognisant of this factual dynamic and made specific provision for the creation of posts calling for the appointment of persons from designated groups ie blacks to achieve equitable representation demographically. Contextually therefore the reality in the appellant's workforce required corrective intervention as decreed in the Constitution. The appellant had adopted an Employment Equity Plan to achieve this." [My emphasis] 132 Barnard (CC) para 227: "By not appointing Ms Barnard and reserving the post for black officers, the National Commissioner sought to achieve representivity and equity in the Police Service. This accords with its Employment Equity Plan and is consistent with the purpose of the Act. Therefore, the National Commissioner's decision cannot constitute unfair discrimination nor can it be taken to be unfair." [My emphasis] These glaring logical and other inconsistencies underlying the rationale of demographic representivity in terms of the EEA serve to further expose Van Heerden's "rationality test" for affirmative action measures to critical scrutiny. It is doubtful that the use of the "light brush of reasonableness" 134 can ever really manage to paper over these cracks, not to mention the fact that the Constitutional Court's emphasis on pursuing the rationality standard of testing of affirmative action measures in terms of the EEA poses an additional threat to members of minority groups, possibly to the point of adding insult to injury for those who do happen to end up in court. 135

Why the EEA's affirmative action scheme is unconstitutional
In the previous section I highlighted the conceptual problems with the EEA's aggressive pursuit of the equitable representation of groups, which so often in its implementation in practice comes down to little more than a nonsensical numbers game and rather odious head-count in the pursuit of demographic representation in the workplace. In focusing on the conceptual poverty of the putative link between the redress of past disadvantage and the promotion of equality with such a starkly statistics-based approach, I believe that the unconstitutionality of the Act's affirmative action scheme becomes clear for all to see. The Constitutional Court has unequivocally stated that any affirmative action measure such as those pursued in terms of the Act, which bases its qualification for benefit on race and sex (and disability), must conform to the internal test set in section 9(2) of the equality right in order to avoid the presumptive unfairness that attaches to differentiation on such arbitrary grounds in insidious consequences of our unhappy past. White people were advantaged over other races especially in the public service. This advantage was perpetuated by the transfer of skills, some critical, to the same white race to the exclusion of others, especially blacks. The over representivity of whites in level 9 is a stark reminder of our past and indeed the present and yet another wake up call to decisively break from these practices. These are practices that can be effectively broken by embracing the restitutionary spirit of the Constitution." [My emphasis] 134 Liebenberg and Goldblatt 2007 SAJHR 349. 135 As Pretorius observes: "[C]onstructing a contextual frame of reference most conducive to a deliberatively or discursively inclusive process of adjudicating rights disputes ... is particularly relevant when those detrimentally affected are also disadvantaged or have little or no real prospect of participating equally in the political process. Langa P recognised the special importance of judicial review in such circumstances in City Council of Pretoria v Walker [where he declared that] '[i]t is precisely individuals who are members of such minorities who are vulnerable to discriminatory treatment and who, in a very special sense, must look to the Bill of Rights for protection'." (Pretorius 2013 SALJ 42).
terms of sections 9(3) and 9(5). And a court is empowered to consider the legitimacy of any restitution measure for this purpose to determine if it complies with the requirements of section 9(2). 136 This requires it to tick all three of the following boxes: 1) such a measure must target persons or categories of persons disadvantaged by past unfair discrimination (or, as Moseneke ACJ -interestingly -rephrased this category of potential beneficiaries in Barnard (CC): "a particular class of people who have been susceptible to unfair discrimination"); 2) such a measure must be designed to protect or advance those classes of persons; and 3) it must promote the achievement of (substantive) equality. 137 When we consider the scheme of the EEA's brand of affirmative action against this (2) and explained in van Heerden], it is neither unfair nor presumed to be unfair. This is so because the Constitution says so. It says measures of this order may be taken. Section 6(2) of the [Employment Equity] Act, whose object is to echo section 9(2) of the Constitution, is quite explicit that affirmative action measures are not unfair. This however, does not oust the court's power to interrogate whether the measure is a legitimate restitution measure within the scope of the empowering section 9(2)." that I believe the Act's affirmative action scheme falls far short of the constitutional demand that any such measure must, ultimately, promote equality.
Please note that I am not objecting to the EEA's identification and use of designated groups (based on disadvantage caused by past discrimination), or of the use of preferential measures ("affirmation") in order to advance such persons in employment.
What I am objecting to is the implicit assumption that advancing such groups in order to achieve their representation in the workplace on a par with the demographic representation of their respective groups in the broader population equates to the achievement of equality (and/or equity). Some -like Partington and van der Waltfind this less objectionable (although I would question the authors' reasoning). 139 With the lack of a clear and rational link between representivity and equality one is forced to conclude that the advancement of persons in terms of a numerical goal based on demographic representivity does little more than promote, only, representivity. There is no guarantee that the achievement of a numerical target based on population statistics would in any way serve to redress past disadvantage (or, in fact, contribute anything more of real value beyond short-term window-dressing of our workplaces). 140 Yes, you might say that this would automatically follow in the light of the fact that the 139 Partington and Van der Walt 2005 Obiter 597 examine s 42 of the EEA (prior to its 2014 amendment -as briefly discussed in s 5 in the text below) in the specific context of this third leg of the Van Heerden test, and come to the following conclusion: "[I]t is apparent that in setting numerical goals the EEA is sensitive to an employer's particular circumstances. The argument could therefore be raised that 'equitable representation' as a goal of affirmative action is in conflict with the constitutional goal of 'the achievement of equality'. It is submitted, however, that the goals are reconcilable. Employment is the most effective way in which to achieve the full and equal enjoyment of all rights and freedoms (in a substantive sense). Equitable representation can therefore be viewed as an intermediate goal: a precondition for the achievement of equality. Accordingly, especially given the duty to interpret legislation in conformity with the Constitution to the extent possible, no doubt should arise as to the compatibility of the Constitution with the EEA's vision of affirmative action." Intriguingly, however, the authors' above conclusion is made with reference to the workplace-related factors in s 42, and they do not even mention the first of the listed factors in the section (namely, the demographic representation of designated groups). Also, following the recent amendments to the Act and specifically this section (see s 5 in the text below), these workplace-related factors no longer form part of s 42, and only the factor of demographic representation as per the pre-amendment wording of the section has been retained. 140 The following, observed in respect of affirmative action, generally, is in my opinion especially apt in respect of affirmative action aimed at the objective of the representivity of groups in the workplace: "While affirmative action may change the colour and composition in a workplace in the short to medium term, its ability to bring about fundamental change to the structure of decisionmaking within the workplace will be more difficult and may take longer. of the treatment of members of such a group -this time not on the basis of their race, but rather the size of their group and its relative statistical place in the population. As ready said, this last is not a listed ground of prohibited unfair discrimination, but it is hardly any less arbitrary than those that did make it into section 9(3). In fact, it is probably even more arbitrary. (While the listed grounds relate mainly to immutable characteristics, the demographic representation of a group in the population has nothing to do with the individual). And here we need to consider the words of Moseneke J in Van Heerden regarding this third leg of the internal test contained in section 9(2). After noting that the use of a remedial measure may lead to disadvantage for those who do not benefit from it, the learned judge felt it necessary to pose the following warning: [I]t is also clear that the long-term goal of our society is a non-racial, non-sexist society in which each person will be recognised and treated as a human being of equal worth and dignity. Central to this vision is the recognition that ours is a diverse society, comprised of people of different races, different language groups, different religions and both sexes. This diversity, and our equality as citizens within it, is something our Constitution celebrates and protects. In assessing therefore whether 141 Or, even worse, of the fact that the complainant had been doubly affected by past discrimination, as an Indian and a woman. Compare the following observations by Cameron J et al in Barnard (CC) para 114 -if this is true for a Barnard it must be even more so for a Naidoo: "But Ms Barnard is both white and a woman. We must be judicious about grouping these elements of identity together. As a white person, Ms Barnard is a member of a group that has been historically advantaged. But as a woman, Ms Barnard is a member of a group that has faced a history of discrimination. As explained above, women are one of the Act's designated groups. The Act requires employers to implement affirmative action measures to redress disadvantage to women." a measure will in the long-term promote equality, we must bear in mind this constitutional vision. In particular, a measure should not constitute an abuse of power or impose such substantial and undue harm on those excluded from its benefits that our long-term constitutional goal would be threatened. 142 It should be noted that the third leg of the enquiry under the Van Heerden test focuses on (or demands a real consideration of) the implementation of a restitutionary measure: The first two prongs test whether the measure itself, in its design, is rationally connected to the end it aims to achieve, in accordance with the wording of section 9(2) which provides for measures designed to protect or advance persons, or categories of persons, disadvantaged by unfair discrimination. The focus of the third prong is somewhat different. It is on the measure, but also on its implementation. The word "achievement" implies some effect or impact. This could hardly be tested without contemplating some action taken in terms of the measure. Van Heerden thus acknowledges some distinction between a measure and its implementation. A decision or other action taken in terms of an affirmative measure, as well as the measure itself, must be constitutionally compliant. 143 Even if we consider that substantive equality is something closer to the pursuit of equality of outcomes, the eventual achievement of a demographically representative workforce does not and cannot equate to an assumption that equal outcomes have been achieved. There is no reason in logic to believe that, absent past unfair discrimination against certain groups, the make-up of the workforce would in any event have been representative in this way. There are simply too many reasons (many of them intangible; some of them based on history, culture, genetic disposition, interests, personal dispositions and preferences, personality, natural skills, learned roles, "lottery of life" decisions, etc, virtually ad nauseum) for the racial and gender make-up of different workforces in our and all other societies. There really is a huge blend of factors to consider (the ones relating to history should not be read to refer only to apartheid; and others may include factors ranging from geography to even a lack of horses!). 144 Tailoring representation through the EEA's brand of demographics-142 Van Heerden para 44. 143 Per Van der Westhuizen J in Barnard (CC) para 143. Also see the following in para 145: "Something more is needed though when a measure as well as its implementation are evaluated. A measure might be legitimate in form, but its application may be unlawful. The main judgment recognises this and states that, 'a validly adopted Employment Equity Plan must be put to use lawfully'. Section 15(4) of the Act also focuses some attention on decisions concerning an employment policy or practice. Once the measure is found to fall within section 9(2) and is thus not unfair discrimination under section 9(3), the effect and impact of its implementation must be evaluated." 144 As succinctly explained by Sowell: "The enormous variety of geographic, cultural, demographic, and other variables makes an even, random, or equal distribution of skills, values, and based numerical goals is highly artificial and arbitrary; one can drive a rather large truck through the holes in the hypothesis that, absent past unfair discrimination, our workplaces would all have been demographically representative of our population.
And, I would submit, the lawfulness of such an approach is therefore highly suspect, especially in the light of its implications under the third leg of the Van Heerden test when one considers (as referred to in the previous section) that representivity may in fact lead to inequality. I pointed earlier to the fact that the court in Van Heerden stated unequivocally that the purpose of affirmative action in terms of the constitutionallyendorsed notion of substantive equality is remedial or restitutionary. I then pointed to the fact that the EEA, in its section 2 which describes the purpose of the Act, adds another objective; it declares that redressing past disadvantage is apparently just a means to achieving another end, namely equitable representation. And this is problematic in the light of the equivocal lawfulness of such a measure implemented under the Act, as Moseneke ACJ himself said in Barnard (CC): [A] validly adopted Employment Equity Plan must be put to use lawfully. It may not be harnessed beyond its lawful limits or applied capriciously or for an ulterior or impermissible purpose. As a bare minimum, the principle of legality would require that the implementation of a legitimate restitution measure must be rationally related to the terms and objects of the measure. It must be applied to advance its legitimate purpose and nothing else. Ordinarily, irrational conduct in implementing a lawful project attracts unlawfulness. Therefore, implementation of corrective measures must be rational. 145  Van Heerden para 33, where the following was said: "It seems to me plain that if restitutionary measures, even based on any of the grounds of discrimination listed in section 9(3), pass muster under section 9(2), they cannot be presumed to be unfairly discriminatory. To hold otherwise would mean that the scheme of section 9 is internally inconsistent or that the provisions of section 9(2) are a mere interpretative aid or even surplusage. I cannot accept that our Constitution at once authorises measures aimed at redress of past inequality and disadvantage but also labels them as presumptively unfair. Such an approach, at the outset, tags section 9 (2)  encompass its application as well as the context of such application, and, accordingly, its impact (both on its beneficiaries and those excluded from benefits). And this inevitably implicates fairness (and proportionality), because this whole determination must take place within the equality paradigm (where constitutionally-compliant affirmative action must resort): It is difficult to see how fairness considerations can be purged from the notion of substantive equality underlying the whole of section 9. The reasoning in the Naidoo case demonstrates that fairness and proportionality considerations cannot be excluded from the evaluation of affirmative action, if justice is to be done to the third Van Heerden criterion, namely that the measure must promote the achievement of substantive equality. It seems hardly contestable that unfair or unreasonably disproportional forms of affirmative action would be irreconcilable with realising the long-term ideal of equality based on the affirmation of equal worth and respect. 154 Surely, the opposite view -that affirmative action and unfair discrimination are simply, and dismissively, viewed as being mutually exclusive based purely on the label (s) attached, would mean that the drafters of our Constitution created a sacred cow that is for all intents and purposes immune from constitutional scrutiny and adjudication, and they decided (for some unexplained reason, although the argument might go that such reason is the ILO Convention's provisions) to situate this sacred cow within one of the fundamental rights contained in the Bill of Rights and in the process immunise it from the working of and scrutiny in respect of all of the rest of the scheme of the Bill of Rights (specifically the limitations clause). This would not make sense to me, as this sacred cow is, at least potentially (to butcher what started out as a promising metaphor) susceptible to mad cow disease; it holds the potential to go on the rampage and trample on the rights and interests of certain of our citizens. By its very nature this animal should be kept under close watch, as we know that it has this potential for harm. But the argument would seem to go that the drafters of the Constitution decided to open the gate of the pen and decree that no-one may touch this cow, for the simple reason that it is a cow, and not a lion (irrespective of its (by definition) inherent "uncowlike" propensities). I can only hope that the lack of logic in such an argument will overshadow the weakness of the metaphor. The Convention provides as follows regarding the ring-fencing of affirmative measures: Any Member may, after consultation with representative employers' and workers' organisations, where such exist, determine that other special measures designed to meet the particular requirements of persons who, for reasons such as sex, age, disablement, family responsibilities or social or cultural status, are generally recognised to require special protection or assistance, shall not be deemed to be discrimination. 155 The question is thus whether the ILO would sanction a member (the South African government) to label the EEA's agenda of demographic representivity as not being "deemed to be discrimination". I strongly doubt this. The Convention famously defines "discrimination" as "any distinction, exclusion or preference made on the basis of race, colour, sex, religion, political opinion, national extraction or social origin, which has the effect of nullifying or impairing equality of opportunity or treatment in employment or occupation". 156 When faced with the de facto race-based job reservation policy of the SAPS under scrutiny in Barnard, and the Constitutional Court's apparent unequivocal rejection of a system of affirmative action as serving as an absolute barrier to employment or advancement in employment (remember: the wording of the EEA is much less clear on the subject), this so clearly constitutes a case of the nullifying or impairing of equality of opportunity or treatment in employment that it would be outrageous for the South African government to be allowed to determine that this does not constitute discrimination. Having said this, of course, the court reiterated this point of departure on the threshold for review of an "affirmative action measure" in its judgment in Barnard (CC), so it is probably a moot point. However, I prefer Malan's suggested approach: [T]he constitutionality of both measures for restitutionary equality under section 9(2), such as [employment equity plans], and of individual decisions purported to be taken in pursuance of such measures, cannot be assumed to be measures or decisions for restitutionary equality solely on the basis of what they purported. They have to be scrutinised in order to establish, with reference to the relevant evidence, whether they are in fact truly measures and decisions that promote equality for all. To qualify as such, both the interests of beneficiaries of such measures and decisions (those previously disadvantaged by unfair discrimination), as well as all others persons, such as those in Barnard's position, will have to be taken into account. 157

155
In A 5(2) ILO Discrimination (Employment and Occupation) Convention 111 of 1958. 156 In A 1(a) of the ILO Discrimination (Employment and Occupation) Convention 111 of 1958.

Pretorius makes the same point in explaining why the rationality test does not comply
with what is required of the judicial adjudication of affirmative action disputes in the context of a Bill of Rights that is very much all about the balancing of the rights and interests of citizens. 158 It is inconceivable that the Convention would allow the determination of the legitimacy of a purported affirmative action measure by a member state to be anything short of this level of rigour. If not, it would provide scant protection indeed. I would suggest that what we need is recognition not of a standard of "strict scrutiny" (as applied in the USA, and as rejected in Van Heerden), but of a standard of "stricter scrutiny" than the deferential rationality test and its highly suspect negation of fairness and proportionality (and the consideration of the impact on nonbeneficiaries).
Finally, it needs to be noted that the problems identified above regarding the application of the EEA's affirmative action provisions have become systemic. We are faced with what amounts to nothing less than a race-based job reservation system 159 -at least in vast swathes of the public sector -which is just too eerily reminiscent of what preceded our constitutional dispensation. We find here, I would submit, "the construction of [new] patterns of disadvantage such as has occurred only too visibly in our history". 160 Sachs J, in Van Heerden, made a point of cautioning us regarding the application of restitutionary measures in the name of the pursuit of substantive equality. He warned against sacrificing non-racialism and the inherent human dignity of all South Africans, by admonishing us to approach the reconciliation of the interests of those advantaged and those disadvantaged by affirmative action measures "in a manner that takes simultaneous and due account both of the severe degree of structured inequality with which we still live, and of the constitutional goal of achieving an egalitarian society based on non-racism and non-sexism". 161 In fact, he asked us 158 "Requiring a rational relationship between means and ends is a far less exacting standard than demanding that means should be proportional to ends … Rationality review does not express the same responsiveness to situations where the infringements of rights are unnecessarily intrusive.
To the extent that a rights-limiting act can be rational, even if disproportional or unfair, a mere rationality standard of justification demands no explanation for the disproportional or unfair invasion of rights." Pretorius 2013 SALJ 40. 159 As expressly acknowledged (and clearly condoned) by Jafta J in Barnard (CC) para 227. 160 As per Goldstone J in Harksen para 49. 161 Van Heerden para 136.
to ensure that in the process of achieving equity "the baby of non-racialism is not thrown out with the bath-water of remedial action". 162 I referred earlier to the fact that Moseneke J, in Barnard (CC), rather intriguingly rephrased his own words as used in Van Heerden to formulate the internal test for compliance of an affirmative action measure with the provisions of section 9(2) of the Bill of Rights. In respect of the first leg of such test, he changed the wording of the requirement that such a measure must be targeted at "persons or categories of persons who have been disadvantaged by unfair discrimination" 163 to the targeting of "a particular class of people who have been susceptible to unfair discrimination" 164 (my emphasis). Does anything turn on this? I would suggest that one should consider that the application of the EEA has been with us now for approximately 16 years. Job reservation systems such as those established in the public sector may very well have wrecked (or seriously impacted on) the lives of many of those "previously advantaged" under apartheid. I want to ask whether this time span is sufficient for our courts to start acknowledging that what Sachs J warned us about has, in fact, occurred. After all, Moseneke J in Van Heerden pertinently told us that a situation-sensitive approach to equality and to such cases involving remedial action is indispensable "because of shifting patterns of hurtful discrimination and stereotypical response in our evolving democratic society" 165  apartheid, are they newly disadvantaged but yet disadvantaged enough to qualify as Moseneke J's class of people who are (now) susceptible to unfair discrimination? Of course, this would require that such an underclass must be experiencing disadvantage as a result of unfair discrimination. As said above, the current understanding of the scope and approach of the Van Heerden test will not allow us to (ever?) classify affirmative action this way, so such an argument would probably be a non-starter. But the apparent signs in Barnard (CC) that the future may at some point at least hold the potential for reconsideration of the rationality standard of review of affirmative action measures might require reconsideration of this aspect. 168

Conclusion (to Part 1)
Having considered the constitutionality of the Employment Equity Act's affirmative action scheme in this contribution, Part 2 of this piece will undertake a critical analysis of the Constitutional Court's first encounter with this Act, in the Barnard case. It will also briefly consider the recent amendments to the Act, before providing detailed conclusions regarding the urgent need for the rejection of the EEA's numbers-based approach to the promotion of substantive equality in our workplaces. 168 Some of the judges of the Constitutional Court in Barnard (CC) provided a titillating taste of how affirmative action cases should in future be approached, one we can only hope will not only be seriously considered by judges in future cases but also, more pointedly, by the legislature in respect of the re-evaluation of the continued lifecycle of the affirmative action provisions as found in the EEA: "Assessing the fairness of the individual implementation of affirmative action measures is different to deciding whether those measures amount to unfair discrimination. The latter enquiry is at the general level of determining whether the formulation and content of a restitutionary measure are constitutionally compliant. The former enquiry examines whether a specific implementation of a measure that is constitutionally compliant in its general form is nevertheless in conflict with the provisions of the Act. We must insist that the specific implementation as well