Two pertinent questions arise with the Bill. First, is the proposed measure rational? Does it fall within the constitutional confines of section 9(2) of the Constitution, the equality provision, which allows “legislative measures” (i.e. affirmative action measures) to be taken to “protect, or advance persons, or categories of persons disadvantaged by unfair discrimination”. Second, will such sectoral numerical targets address substantive equality – is this the way forward? 

The Bill requires “designated employers” (employers required to comply with the EE Act’s affirmative action chapter) to ensure that the “numerical goals” in their employment equity plan comply with the sectoral numerical target set. “Numerical goals” in the EE Act mean that if there is an “under-representation of people from designated groups”, an employer must establish “numerical goals of suitably qualified people of designated groups” at all occupational levels in their employment equity plan. 

To answer the first question, the CFCR applied the three-step test determined in Minister of Finance and Others v Van Heerden (Van Heerden test) to the proposed measure. If it succeeds the Van Heerden test, such a measure falls within the confines of section 9(2) of the Constitution and does not constitute unfair discrimination.* 

From considering the Annexure to the Bill, substantiating the need for the proposal, the following becomes clear. There is unhappiness with the lack of “equitable representation” of specific “designated groups”, specifically at “middle to upper occupational levels” in the private sector. Specific reference is made to representation of African and Coloured persons, and persons with disabilities, whose representation at “middle-to-upper occupational levels” in relation to their Economically Active Population demographics (EAP) is low. Data from the Commission for Employment Equity’s Annual Reports (CEE reports) is used as support. 

If one solely focuses on Top and Senior Management levels in the private sector and only on the African and Coloured population groups, their representation is undisputedly low. This is according to the 2017-18 CEE report, with reference to the National EAP (roughly being 78% African, 10% Coloured, 9% White and 3% Indian). However, at Middle Management level, the “Professionally Qualified level” and lower levels, the results show a different angle. This information was not included in the Annexure.

The 2016-17 CEE report specifically noted “a positive trend towards equitable representation…for the first time at the Professionally Qualified/ Middle Management level”. The CEE itself recognised this trend as “positive towards reaching an equitable representation at Senior Management level. The African Population Group’s representation at the “Professionally Qualified level” in the private sector according to the 2017-18 CEE report is higher than that of the White population group. 

Critically, this shows that affirmative action measures are effective at these lower levels. This is without mentioning the “Skilled” and “Semi-skilled levels”. The trend in Middle Management level would logically feed through in time to Senior and Top Management levels – which requires sustainably acquired skills and experience. 

Therefore, the facts do not support that such measures be introduced at all occupational levels. The 2017-18 CEE report also specifically noted that the Indian Population group’s representation at Senior and Top Management level in the private sector is much higher than their National EAP of 3%. One would assume this is positive, considering that the Indian population group falls clearly within “designated groups”, which experienced previous unfair discrimination – but apparently not.

It is on the second and third requirement of the Van Heerden test that the proposed measure falters. If one asks – as per the test – what future outcome the measure is directed at – it cannot, based on the facts, be said that such future outcome of “equitable representation” would not naturally evolve without the proposed intervention. The Van Heerden test’s last requirement provides that it should “promote the achievement of equality” which may “come at a price for the previously advantaged”. The irony is that it this instance it may also come at a price to those who suffered previous discrimination, but do not have the numbers in society. This cannot be what our Constitution strives to achieve – and the proposed measure is a far cry from section 9(2) of the Constitution.

Finally, there is the question of whether the sectoral numerical targets will give effect to substantive equality? Does it practically address the deeper disadvantage and intersectional discrimination to ensure equal enjoyment of opportunities? Key to addressing these disparities is ensuring that the data on which it is based reflects this contemporary reality. This is a flaw in South Africa’s current affirmative action measures. In a 2016 UN Committee report observing South Africa’s compliance with the International Convention on the Elimination of All Forms of Racial Discrimination (ICERD) (which refers to affirmative action measures as special measures), concern was raised about the lack of “comprehensively disaggregated data on the impact of special measures”.  The recent South African Human Rights Commission (SAHRC) Equality Reportalso recommended that the EE Act be amended to “target more nuanced groups on the basis of need and taking into account social and economic indicators”.

Section 9(2) of the Constitution provides the mechanism to achieve substantive equality, but to do so, redress measures need to speak to real socio-economic disparities and intersectional discrimination. There is an urgent need to amend the EE Act to provide for a more nuanced approach in applying affirmative action measures, specifically considering socio-economic indicators. The proposed sectoral numerical targets provision is a far cry from the nuanced approach so desperately needed. 

*The Van Heerden three-step test has been criticised by some academics. The test however has been confirmed in subsequent court cases and is currently the yardstick to use in this analysis.

By Ms Christine Botha: Acting Director, Centre for Constitutional Rights
15 December 2018