8/2026: From Redress To Empowerment: An ISI Assessment Of South Africa’s Proposed Preferential Procurement Regime And The Case For An Outcomes-Based Pathway To Economic Normalisation

Read the full Occasional Paper by downloading the PDF below.

PUBLICATION INFORMATION

Title: From Redress to Empowerment

Authors: Daryl Swanepoel

Publication type: Occasional Paper

Publication date: September 2026

PUBLISHED BY

Inclusive Society Institute | NPO Registration: 235-515

5 High Cape House, 21 Ben Avenue, Vredehoek, Cape Town, 8001, South Africa

PO Box 12609, Mill Street, Cape Town, 8010, South Africa

Tel: +27 (0)21 201 1589

Email: admin@inclusivesociety.org.za | Website: www.inclusivesociety.org.za

COPYRIGHT

© 2026 Inclusive Society Institute.

Unless otherwise expressly indicated, copyright in this publication is owned exclusively by the Inclusive Society Institute. All rights reserved. No part of this publication may be reproduced, stored, distributed or transmitted in any form or by any means without the Institute’s prior written permission, except as permitted by applicable law.

Brief extracts may be quoted for criticism, review, research or educational purposes, provided that the authors, title and publisher are appropriately acknowledged. Permission requests should be directed to info@inclusivesociety.org.za.

SUGGESTED CITATION

Swanepoel, D. 2026. From Redress to Empowerment: An ISI Assessment of South Africa’s Proposed Preferential Procurement Regime and the Case for an Outcomes-Based Pathway to Economic Normalisation. Cape Town: Inclusive Society Institute.

DISCLAIMER

This publication has been prepared for research, discussion and public-information purposes. The views expressed are those of the authors and do not necessarily reflect the official position of the Inclusive Society Institute, its Board, staff, funders, partners or associated organisations.

While reasonable care has been taken to ensure the accuracy of the information at the time of publication, the Institute and the authors make no warranty regarding its completeness, accuracy or continued relevance. This publication does not constitute professional advice. To the fullest extent permitted by law, the Institute and the authors accept no liability for any loss or damage arising from reliance on its contents.

References to third-party organisations, products, services or websites do not constitute endorsement. The Institute is not responsible for the content, availability or accuracy of external websites.

Executive summary

Public discussion around South Africa’s Draft General Public Procurement Regulations, 2026 has become increasingly polarised. At one end are claims that white-owned businesses are to be barred from government contracts below R20 million. At the other is a tendency to treat criticism of the proposed regulations as opposition to transformation itself.

Neither position is satisfactory.

A close reading of the Public Procurement Act 28 of 2024 and the Draft General Public Procurement Regulations, 2026 shows that the claim that white South Africans are simply to be excluded from all state procurement below R20 million is incorrect. The proposed regime is considerably more complex. It provides for set-asides for a range of designated categories, some racial and some non-racial. These include black people and black women, but also women, people with disabilities, military veterans, small enterprises, co-operatives and, within the small-enterprise category, enterprises owned by youth. A white woman, a white person with a disability, a white youth who owns a qualifying small enterprise, or a white-owned small enterprise may therefore qualify under particular set-asides depending on the category selected.

Nor does the 30% target for black people mean that exactly 30% of contracts below R20 million are racially reserved while the remaining 70% are automatically open. The 30% is a minimum target expressed as a percentage of an institution’s annual procurement budget for that identified category. Other categories carry different targets, categories may overlap, and the regulations permit institutions to identify up to five categories. Conversely, the 30% is a minimum rather than a ceiling. There is therefore no legally guaranteed “70% open pool”. The actual effect will depend on the procurement profile of each institution, the categories it selects and the tenders to which set-asides can practically be applied. In the interests of good governance, it is critical that these criteria around set-asides be transparently disclosed in procurement policies and annual procurement plans when the budget of a public institution is tabled annually, and that there is oversight, both by the executive authority and the Procurement Office/provincial treasuries in relation to these set-asides to ensure that these are executed in accordance with the plan and compliant with regulations.

Correcting misinformation, however, should not shield the proposed system from scrutiny.

The requirement that enterprises in several designated categories be 100% owned by persons belonging to the selected category is unusually rigid. In a black-ownership set-aside, for example, a 99% black-owned enterprise with a 1% white shareholder would fail the ownership requirement and its bid would have to be disqualified. The same logic applies to other relevant designated categories, such as persons without disabilities who own businesses jointly with persons who are disabled, or youth as defined to be less than 35 years and non-youth aged older than 35. This raises questions about proportionality, economic incentives, partnership, investment and the increasingly elaborate administrative classification of citizens and businesses.

The Inclusive Society Institute’s assessment is therefore neither that redress has run its course nor that race should immediately disappear from economic policy. South Africa’s constitutional project expressly recognises the continuing legitimacy of measures designed to overcome disadvantage produced by unfair discrimination. The Constitutional Court has repeatedly affirmed this.

But remedial measures are directed towards an outcome. They are not constitutionally valuable merely because they exist. The Court has cautioned that they are not ends in themselves and that their ultimate purpose is a more equal, fair, non-racial, non-sexist and socially inclusive society and economy.

This paper proposes that South Africa begin moving towards an outcomes-based sunset framework for race- and gender-conscious economic redress. This is not an arbitrary date on which transformation ends. It is a system under which the intensity and form of remedial intervention change as measurable disadvantage changes.

Strong redress would remain where substantial disadvantage persists. As measurable progress is achieved, policy would move progressively from exclusionary or protective measures towards preference, enterprise development, access to finance, skills transfer, supplier development and ultimately needs- and circumstance-based assistance. Race- and gender-based intervention would diminish not because a particular number of years had passed, but because the conditions that justified it had demonstrably changed.

The central proposition is simple:

South Africa still requires redress. But successful redress should progressively make the most intrusive forms of racial intervention less necessary. This is particularly important given the growing administrative burden of monitoring and evaluating increasingly precise identity classifications, which become more difficult to sustain the further South Africa moves from apartheid-era systems of formal racial classification.

1. Introduction: between denial and permanence

South Africa’s debate about economic transformation frequently becomes trapped between two unsatisfactory propositions.

The first is that more than three decades have passed since the establishment of democracy and that race-conscious redress has therefore exceeded its legitimate lifespan. The weakness in this proposition is obvious. Time elapsed is not an adequate measurement of disadvantage overcome. If the material consequences of past racial exclusion remain evident in ownership, access to capital, management, skills, opportunity or participation in particular sectors, the fact that thirty or forty years have elapsed does not by itself eliminate the constitutional or policy rationale for remedial measures.

The second proposition is that because apartheid created profound racial inequality, race-conscious remedial measures may continue indefinitely in substantially their present form. That position is equally unsatisfactory. If an intervention exists to overcome disadvantage, there must ultimately be some means of assessing whether the disadvantage has diminished and whether the intervention should consequently change.

The Constitution itself contains both sides of this equation. Section 1 identifies both the achievement of equality and non-racialism as founding values (Republic of South Africa, 1996:3). Section 9 recognises formal equality, but also expressly permits legislative and other measures designed to protect or advance persons or categories of persons disadvantaged by unfair discrimination in order to promote the achievement of equality (Republic of South Africa, 1996:5–6).

These principles are complementary, rather than contradictory. South Africa is constitutionally committed both to redressing the racialised inequality inherited from its past and to building a society in which race ultimately matters less, rather than more.

The difficulty is that public policy has never adequately defined the bridge between those two objectives.

The controversy surrounding the Public Procurement Act 28 of 2024 and the Draft General Public Procurement Regulations, 2026 provides an opportunity to address that omission. It also demonstrates why accurate description matters. Public debate has increasingly characterised the proposed procurement regime as one in which white businesses are to be excluded from all state contracts below R20 million. That is not what the legislation and draft regulations say.

Correcting that exaggeration should not, however, become a defence of every aspect of the proposed system. There are legitimate concerns about the rigidity of the proposed ownership requirements, the administrative burden of increasingly precise identity classification, the implications for partnerships and investment, and the relationship between transformation, competition and value for money.

An evidence-based position must therefore be capable of doing two things at once: rebut misinformation and critically assess the policy that actually exists.

2. The constitutional starting point: redress is legitimate

Any serious assessment must begin by rejecting the proposition that the Constitution requires government to be colour-blind in procurement.

It does not.

Section 217(1) requires state procurement to operate through a system that is “fair, equitable, transparent, competitive and cost-effective” (Republic of South Africa, 1996:112). But section 217(2) immediately confirms that this does not prevent organs of state from implementing procurement policies providing for categories of preference and for the protection or advancement of persons or categories of persons disadvantaged by unfair discrimination (Republic of South Africa, 1996:113).

The constitutional framework therefore expressly accommodates preferential procurement.

The Constitutional Court has similarly rejected a purely formal conception of equality. In Minister of Finance v Van Heerden, Moseneke J formulated the now-established three-part test for remedial measures. The measure must target persons or categories of persons disadvantaged by unfair discrimination; it must be designed to protect or advance them; and it must promote the achievement of equality (Minister of Finance v Van Heerden, 2004:23).

Importantly, the Court did not require government to establish that its chosen remedial measure was the only possible or strictly necessary means of achieving the objective. It recognised that policy operates prospectively and that exact outcomes cannot always be predicted (Minister of Finance v Van Heerden, 2004:25–26).

This makes an important point for the contemporary debate. An argument that preferential procurement is illegitimate merely because it treats historically disadvantaged categories differently is unlikely to survive serious constitutional scrutiny.

But Van Heerden also supplies the limitation. Remedial measures are directed towards an “envisaged future outcome”. They must be reasonably capable of attaining that outcome. Measures that are arbitrary, capricious or amount to naked preference cannot satisfy the constitutional purpose (Minister of Finance v Van Heerden, 2004:25).

The issue is consequently not whether government may pursue redress. It may. The issue is whether the measures chosen remain appropriately connected to their constitutional objective and whether their design and implementation continue to promote equality.

3. What the proposed procurement regime actually

does

The Public Procurement Act expressly requires procuring institutions to develop and implement preferential procurement policies (Republic of South Africa, 2024:26). Where prescribed thresholds and conditions are met, section 17 requires a procuring institution to set aside a bid for a designated category of persons (Republic of South Africa, 2024:28).

The categories are important, because they immediately expose the weakness in describing the system simply as racial exclusion.

Section 17(3) includes black people and black women, but it separately includes women, black people with disabilities, people with disabilities and military veterans. It also provides for small enterprises owned by various categories, including black people, black women, women, people with disabilities, black youth and youth. Small enterprises generally, co-operatives and geographically defined categories are also included (Republic of South Africa, 2024:28, 30).

The distinction between “black youth” and “youth” is deliberate. Likewise, “black people with disabilities” and “people with disabilities” are separate categories, as are “black women” and “women”.

The legislation therefore establishes a designated-category procurement system, not a system in which every set-aside is necessarily racial.

This matters materially.

A white South African woman may fall within the category “women”. A white South African person with a disability may fall within the category “people with disabilities”. A qualifying small enterprise owned by white South African youth may fall within the “small enterprises owned by youth” category. A small enterprise owned by white South Africans may fall within the general small-enterprise category.

Race is an important component of the regime, but it is not the sole organising principle.

4. The R20 million claim: what is true and what is not

The draft regulations provide that a procuring institution must identify up to five categories from those contemplated in section 17(3). In doing so, it must consider its strategic mandate, the relevant sector or supply market and the availability of potential suppliers in the selected category (National Treasury, 2026:57).

For a set-aside to operate, the estimated value of the contract or specifically unbundled portion may not exceed R20 million, and there must be at least three potentially qualifying suppliers in the identified category (National Treasury, 2026:58).

Where those threshold and supplier conditions are satisfied, the institution must set aside the bid for one or more of its identified categories (National Treasury, 2026:58).

The R20 million figure therefore determines the upper value of a contract, or an unbundled portion of a contract specifically apportioned for set-asides, to which the set-aside mechanism may apply. It does not say that every contract worth R20 million or less must be reserved for black-owned enterprises.

That distinction has been substantially blurred in parts of the public debate.

NEASA, for example, has publicly described the proposals as requiring government to “set aside all tenders that are R20 million and below” for companies 100% owned by designated categories, and has presented the mechanism primarily as one that will completely disqualify white-male-owned businesses (NEASA, 2026).

This interpretation has also found its way into mainstream commentary. Writing in Business Day, Lael Bethlehem stated that the regulations “set aside any procurement below R20m for suppliers that are 100% black-owned” and went on to suggest that every supplier of every service below that threshold would have to be 100% black-owned (Bethlehem, 2026). That is not what the draft regulations provide. The R20 million threshold applies to the set-aside mechanism, while the Act provides for multiple designated categories, including non-racial ones. An institution must select up to five such categories with reference to its mandate, the relevant supply market and supplier availability. The black-ownership category is therefore one possible category within the system, not a universal eligibility condition governing every procurement below R20 million.

The same interpretation subsequently reached a wider readership through The Citizen. Reporting on Mavuso’s criticism of the regulations, Koteli repeated the proposition that “[a]ll tenders under R20m are set aside exclusively for 100% black-owned suppliers” (Koteli, 2026). The significance is that a contested interpretation of the regulations was thereby being presented to readers as a straightforward description of what the regulations propose. As the analysis above demonstrates, that formulation conflates one designated category with the set-aside regime as a whole and therefore materially overstates the extent of racial exclusion.

Across these accounts, the same important qualifications in the actual regulations are lost.

First, procuring institutions identify particular categories, not one universal racial category.

Second, some of the legislated categories are non-racial.

Third, the regulation expressly requires attention to the nature of the market and the availability of qualifying suppliers.

Fourth, at least three potentially qualifying suppliers must exist before the set-aside condition is satisfied.

Finally, where qualifying bids are not received, the institution must cancel and re-advertise, potentially using another identified category, prequalification, or ultimately no preferential procurement after the required reporting process (National Treasury, 2026:58).

It is therefore inaccurate to state that a white-owned business is, merely because it is white-owned, prohibited from competing for every government contract below R20 million.

It may be excluded from a particular set-aside whose ownership criterion it does not satisfy. That is a materially different proposition.

This distinction does not mean that all media criticism of the proposals is misplaced. Some commentary has described the mechanism more carefully. Business Day, for example, reported Sakeliga’s concern that race would become a condition of entry for particular tenders and that the 100% ownership requirement could discourage interracial business partnerships (Khumalo, 2026). More recently, Ann Bernstein argued in the same newspaper that “certain tenders” up to R20 million may be set aside for designated categories, while questioning whether the resulting restrictions would reduce competition and increase procurement costs (Bernstein, 2026). Those are legitimate policy questions. They differ materially from the claim that white-owned businesses are simply barred from all state contracts below R20 million.

Moreover, a reasonable premium might be a justifiable business development cost, but there should be bench marking between set aside unit costs vs open market unit costs.

5. The 30% question, and why there is no simple 30/70

split

Annexure 2 of the regulations introduces minimum set-aside targets expressed as percentages of a procuring institution’s annual procurement budget.

For “black people”, the minimum is 30%. For black women it is 15%; women 18%; people with disabilities 4%; military veterans 2%; small enterprises owned by youth 4%; small enterprises generally 30%; and various other categories carry their own targets (National Treasury, 2026:90).

The regulations state that a procuring institution must set aside the minimum percentage of its annual procurement budget in accordance with Annexure 2 in respect of its identified categories (National Treasury, 2026:58).

This requires careful interpretation.

It is incorrect to say simply that “30% of all contracts below R20 million are reserved for black businesses”.

The target is expressed against the annual procurement budget, not against the number of contracts below R20 million.

It is equally incorrect to say that the remaining 70% of a department’s total procurement budget is automatically an unrestricted pool.

There are several reasons.

First, the 30% figure is a minimum, not a ceiling.

Second, the department may identify up to five categories.

Third, the categories overlap. A black woman with a disability who owns a small enterprise could potentially fall within multiple categories. The percentages therefore cannot simply be added arithmetically as though each represents a separate portion of the budget.

Fourth, not all designated categories are racial.

Fifth, which parts of a department’s procurement programme can practically be set aside depends upon its actual procurement profile and the existence of the required supplier base.

The correct conclusion is therefore more nuanced.

There is no guaranteed 70% open pool written into the regulations. At the same time, it is entirely possible that considerably more than 70% of a particular institution’s real procurement opportunities will remain accessible to white-owned firms, depending on the institution’s procurement mix and the categories chosen.

A white-owned business might itself qualify under a women, disability, youth-owned small-enterprise, general small-enterprise or other non-racial set-aside. It may also compete outside set-asides, albeit within the broader preferential procurement framework.

The public debate loses this complexity when it converts “R20 million” and “30%” into a claim of blanket racial exclusion.

6. The 100% rule: the more serious concern

Correcting misinformation does not dispose of the substantive policy concern.

Regulation 57 provides that businesses falling under section 17(3)(a) to (h) must be 100% owned by members of the relevant selected category. It further provides that bids failing the relevant ownership and eligibility requirements must be disqualified (National Treasury, 2026:58–59).

The consequence can be stark.

Where the selected category is black people, a 100% black-owned company qualifies. A 99% black-owned company with a 1% white shareholder does not.

Where the category is women, a company that is 100% women-owned can qualify, but one containing an ownership interest held by a man would fail the relevant 100% category test.

Where the category is a qualifying small enterprise owned by youth, all relevant ownership would have to satisfy the youth requirement.

The question is not whether such distinctions are automatically unconstitutional. The Constitution permits targeted remedial action, and Van Heerden specifically cautions against imposing an unnecessarily strict necessity test on government’s choice of remedial means.

The better question is whether 100% ownership is a rational and productive policy instrument.

A black entrepreneur may need investment, technical knowledge, networks or experience from a partner who happens to be white. A 100% rule creates a potential incentive not to enter that partnership.

The same principle can apply across other categories. A policy ostensibly intended to foster enterprise development may inadvertently discourage the pooling of capital, knowledge and experience.

That becomes particularly significant because Van Heerden requires the measure to be reasonably capable of achieving its intended outcome and cautions against arbitrary, capricious or naked preference (Minister of Finance v Van Heerden, 2004:25).

The issue deserving scrutiny is therefore not simply whether preference is permitted. It is whether the absolute ownership threshold advances empowerment better than a less rigid mechanism would.

7. Procurement has more than one constitutional

purpose

Transformation is not the only constitutional value governing procurement.

Section 217(1) requires fairness, equity, transparency, competitiveness and cost-effectiveness (Republic of South Africa, 1996:112). Section 217(2) authorises preference and redress (Republic of South Africa, 1996:113).

The Constitution requires the two to coexist.

In AllPay, the Constitutional Court emphasised that the purpose of a tender is to obtain the best solution through a process that is fair, equitable, transparent, cost-effective and competitive (AllPay Consolidated Investment Holdings v CEO of SASSA, 2013:52).

That does not mean preference must disappear whenever it narrows competition. If that were the test, section 217(2) would have little practical meaning.

But neither can transformation make competitiveness and cost-effectiveness irrelevant.

The draft regulations themselves recognise this tension. They define value for money by reference to economical acquisition, efficiency and effectiveness, while expressly including transformation imperatives among the outcomes against which effectiveness may be assessed (National Treasury, 2026:7–8).

The proper policy task is consequently one of balance.

A set-aside that creates sustainable new suppliers, develops capacity and broadens future competition may advance both transformation and long-term value for money.

A set-aside that merely reduces the available supplier pool without building sustainable productive capability may do neither.

That distinction is central to moving from redistributing contracts to building economic capability.

8. Afribusiness: another area in which the public

debate requires care

The 2022 Constitutional Court judgment in Minister of Finance v Afribusiness NPC is frequently invoked as though the Court held that racial prequalification in public procurement is inherently unconstitutional.

That is not an accurate statement of the majority judgment.

The majority’s central finding concerned the powers of the Minister under the then applicable Preferential Procurement Policy Framework Act. Madlanga J held that the Minister could not use regulations to appropriate a policy-making power that the legislation had vested in individual organs of state (Minister of Finance v Afribusiness NPC, 2022:40–47).

Of particular relevance today, the Court expressly observed that the Minister could instead introduce legislation in Parliament so that the Act itself contained the desired preferential procurement policy, while adding the crucial qualification that the content of such legislation would still have to pass constitutional muster (Minister of Finance v Afribusiness NPC, 2022:46).

That is substantially what has now occurred.

The Public Procurement Act itself expressly provides for set-asides, prequalification and related preferential procurement mechanisms (Republic of South Africa, 2024:28, 30).

This does not immunise the new system from constitutional challenge. It changes the nature of the challenge.

The straightforward ultra vires objection that prevailed in Afribusiness cannot simply be transposed onto a new Act in which Parliament itself has expressly authorised the mechanisms concerned.

Critics are entitled to challenge whether the new Act or its regulations comply substantively with sections 9 and 217. But it is misleading to claim that the Constitutional Court has already definitively ruled the new policy unconstitutional.

It has not.

9. The administrative problem: when empowerment

becomes classification

There is nevertheless an uncomfortable feature of the proposed system that deserves more attention.

The more precise the ownership gateway becomes, the more precise the state’s classification machinery must become.

The draft regulations require prospective suppliers to provide identity information and ownership information, including information concerning directors, members, owners and beneficial owners. Procuring institutions must verify submitted information against available data sources and databases (National Treasury, 2026:12).

That is understandable in any procurement system. Government must know who owns and controls its suppliers.

But a 100% category rule changes the consequences of classification.

When race is merely one element of a broader score, uncertainty at the margin affects points.

When 100% racial ownership determines whether a business may compete at all for a particular opportunity, classification becomes a gateway to market participation.

The draft regulations examined for this paper do not themselves establish a comprehensive new mechanism explaining how disputed racial identity is to be determined. They prescribe ownership information and verification, but the intensity of the 100% rule gives the underlying classification question greater significance.

This is the sense in which the system risks becoming administratively Orwellian, not because the objective of redress is illegitimate, but because the state may progressively find itself constructing more elaborate systems for determining which citizen belongs in which category before deciding who may participate in particular economic opportunities. This places ever more onerous administrative burdens on procurement processes, the need for internal controls, the reporting by procurement entities and the auditing by the Auditor-General.

That should make South Africans uncomfortable even while accepting that redress remains necessary.

A constitutional democracy committed simultaneously to equality and non-racialism should be asking not only how to administer such classifications more effectively, but how successful transformation can progressively make them less necessary.

10. From contract allocation to empowerment

There is a more constructive route.

South Africa should not abandon redress before the conditions that justify it have been overcome. But it should increasingly distinguish between protected access to opportunity and the creation of lasting economic capability.

The B-BBEE framework itself points in this direction.

Significantly, this concern is not confined to critics of B-BBEE. Reporting on a February 2026 meeting between the Minister of Trade, Industry and Competition and the sector charter councils, NewsDay recorded Minister Parks Tau as arguing that the focus should increasingly move beyond compliance towards measurable outcomes, asking what had actually changed in ownership, management, skills, enterprise growth and industrial capability (Kersten, 2026). That is substantially the distinction this paper seeks to develop: transformation should ultimately be judged not merely by whether prescribed inputs have been satisfied, but by whether lasting economic capability has been created.

The Generic Code does not conceptualise empowerment solely as ownership. It includes ownership, management control, skills development, enterprise and supplier development and socio-economic development (Department of Trade and Industry, 2013:10–11).

Enterprise and supplier development is explicitly intended to assist and accelerate the growth and sustainability of black enterprises (Department of Trade and Industry, 2013:11).

That broader conception of empowerment is important.

An emerging entrepreneur who wins a protected government contract has received access to a market. That may be valuable redress.

But sustainable empowerment requires more: capital, skills, technology, management capability, networks, productive assets, access to private as well as public markets, and an ability eventually to compete without protection.

A transformation system should therefore ask whether a beneficiary remains dependent upon preferential procurement or whether preferential procurement has helped create an enterprise capable of competing independently. The state should guard against endless cycles in which the same entities repeatedly benefit from preferential procurement. There should therefore be limits on repeat access to set-asides once an enterprise has demonstrably developed the capability to compete independently. Moreover, success in employment creation should be demonstrated. Material or repeated failure to deliver on a set-aside, without reasonable justification, should result in debarment from accessing other set-asides until the underlying shortcomings are demonstrably remedied.

The principle of broadening rather than concentrating the benefits of preferential procurement should also apply within each fiscal year. There should be a reasonable limit on the number of set-aside contracts that a single entity may receive during that period. The purpose of set-asides should be to broaden economic participation and develop a wider pool of sustainable enterprises, rather than repeatedly concentrating protected procurement opportunities among the same beneficiaries. National Treasury should therefore establish appropriate annual limits, taking account of contract value, sector and the availability of qualifying suppliers.

The distinction can be expressed simply: Redress creates access. Empowerment creates capability, for example, experiential learning associated with location within local and global value chains and market access.

Public policy needs both, but over time the balance should increasingly move towards capability.

This could include substantially stronger development-finance mechanisms, working-capital guarantees, technical support, apprenticeship and skills programmes, supplier-development partnerships, technology transfer, access to export markets and incentives for genuine joint ventures.

It should encourage productive interracial and intergenerational partnerships, rather than inadvertently penalising them.

A transformation policy that makes a black entrepreneur less eligible for public work because that entrepreneur has brought an experienced white investor into the business may be achieving formal ownership purity at the expense of genuine empowerment.

11. The missing element: what does success look like?

The deepest weakness in South Africa’s transformation debate is not that there are remedial measures. It is that the country has never adequately agreed on what successful completion of different stages of transformation looks like.

The result is predictable.

Critics ask: “It has been more than thirty years. When does this end?”

Supporters respond: “Inequality remains. Redress is still necessary.”

Both may be correct, but neither proposition supplies a policy mechanism for moving forward.

The passage of thirty years cannot be the test. If material disadvantage remains, the calendar does not cure it.

But historical injustice alone cannot tell policymakers whether a particular intervention should remain unchanged for another thirty years.

The correct variable is effect.

• Has the measure reduced the disadvantage it was designed to address?

• If not, why not?

• If it has, by how much?

• Has the affected group developed meaningful productive capacity?

• Has ownership broadened?

• Has access to finance improved?

• Have management and professional participation changed?

• Are enterprises surviving and growing outside protected procurement?

• Are disparities narrowing?

• And at what point does the evidence justify moving to a less intrusive form of intervention?

Without answers to those questions, transformation risks becoming a process without an agreed destination.

12. An outcomes-based sunset framework

The Institute proposes consideration of an outcomes-based sunset framework for race-conscious economic redress.

This should not be confused with an arbitrary sunset date.

A provision stating that B-BBEE or preferential procurement ends in 2035, 2040 or 2050 would be intellectually weak. It would substitute chronology for evidence.

Instead, the intensity and character of interventions should be linked to measurable outcomes.

Alternatively, Parliament could establish a provisional sunset date, subject to renewal on clearly defined and evidence-based grounds. Such a date would operate as a statutory decision point rather than as an automatic termination of transformation policy. If the specified outcomes have not been achieved, the relevant measures could be renewed for a further defined period, with their scope and intensity determined by the evidence and, where appropriate, by sector-specific circumstances. This would preserve the outcomes-based character of the framework while providing greater policy certainty and potentially generating less opposition.

The framework could operate through four broad phases.

Phase 1: Intensive redress

Where credible data show severe continuing exclusion, strong remedial mechanisms remain available. These could include targeted set-asides, enhanced preference, development finance and sector-specific interventions.

The justification would be transparent: the relevant disparity remains sufficiently severe to warrant intensive intervention.

Phase 2: Preference and capability building

As measurable participation improves, the system should move away from absolute exclusion and towards preference combined with capability-building.

Ownership thresholds could become less absolute. Greater weight could be placed on skills transfer, management participation, investment, enterprise development, employment creation and supplier development.

Phase 3: Empowerment based increasingly on circumstances

As racial disparities narrow materially, support should shift progressively towards indicators such as socioeconomic disadvantage, geography, access to capital, enterprise size, youth, disability and exclusion from opportunity.

Race may remain relevant where evidence shows that race remains a predictor of disadvantage, but it should become less determinative as that relationship weakens.

Phase 4: Normalised competition

Once the measurable disadvantage that justified a particular race-based intervention has substantially been overcome, that intervention should fall away.

Government may continue supporting small firms, poor communities, new entrants, youth, persons with disabilities or economically marginalised regions. But assistance would increasingly be based on present disadvantage rather than historical racial classification.

This is not the abandonment of transformation.

It is the completion of successful transformation.

13. How should the sunset be measured?

The measurements should be established in advance rather than improvised when political pressure arises.

They could include inter alia the following sector-specific indicators of improved empowerment of the respective identified historically disadvantaged classes:

• Transformative ownership of productive assets;

• Improved participation in senior and executive management;

• Improved access to development and commercial finance;

• Improved entry and survival rates of black-owned and other identified enterprises;

• Growth of enterprises beyond protected public procurement;

• More advanced professional and technical skills participation;

• Improved employment and apprenticeship creation;

• Greater participation in private-sector supply chains;

• Reduced income and wealth disparities;

• Less geographic concentration of disadvantage; and

• Evidence of reduction in continuing barriers to entry.

No single indicator should determine the outcome.

Nor should a national average automatically determine policy for every sector. Transformation may progress faster in one industry than another. A sector with substantial and sustainable black participation may warrant a different policy intensity from one in which historical exclusion remains entrenched.

The model should therefore be evidence-based, sector-sensitive and periodically reviewed.

Importantly, the Public Procurement Act already recognises the need for review. Section 68 requires the Minister, within 24 months after publication of the Act, to review its implementation and the need for amendments, consult stakeholders, and publish and submit a report to Parliament within 27 months (Republic of South Africa, 2024:72).

That is valuable, but it is a review mechanism rather than an outcomes-based sunset mechanism.

The proposed next step is to require government not merely to ask whether the law is functioning, but whether the underlying disadvantage is changing sufficiently to alter the intervention itself.

14. Why this approach strengthens rather than weakens

transformation

An outcomes-based approach answers two different anxieties in South African society.

To historically disadvantaged South Africans, it says: Redress will not disappear simply because an arbitrary number of years has passed. Where measurable disadvantage remains, there remains a legitimate case for intervention.

To citizens who fear that racial classification has become permanent, it says: Race-conscious measures are not intended to organise the economy forever. As their objective is achieved, their intensity will diminish.

That distinction also improves certainty for businesses and investors.

Economic actors can plan more rationally when the state provides not merely present rules but an intelligible policy trajectory.

A business would know which transformation outcomes are being pursued, how progress will be measured, when rules will be reviewed and what improvement could trigger a transition to a different form of support.

Transformation thereby becomes an increasingly progressive pathway, rather than an indefinite condition.

15. The Constitutional Court already points towards this

destination

The Constitutional Court’s jurisprudence strongly supports the idea that redress must remain connected to an ultimate constitutional outcome.

In Barnard, Moseneke ACJ emphasised that substantive equality remains necessary, because past disadvantage continues. But he immediately added that remedial measures must operate within constitutional discipline, must not unduly invade dignity and are “not an end in themselves”. Their ultimate goal is a more equal, fair, non-racial, non-sexist and socially inclusive society (South African Police Service v Solidarity obo Barnard, 2014:16).

The Court further observed that restitution measures, important though they are, cannot do all the work required to achieve social equity (South African Police Service v Solidarity obo Barnard, 2014:17).

In a separate judgment, Van der Westhuizen J made an equally relevant observation: assessing whether equality is promoted must involve attention to the actual effect and impact of the measure, and the enquiry must remain alive to shifting circumstances and the changing distribution of privilege and under-privilege (South African Police Service v Solidarity obo Barnard, 2014:72–73).

That is precisely the logic of an outcomes-based sunset mechanism.

If changing circumstances matter in judging whether a remedial measure remains justifiable, public policy should measure those changing circumstances systematically rather than waiting for litigation.

16. An ISI assessment of the 2026 procurement

proposals

The Inclusive Society Institute’s assessment can therefore be summarised in four propositions.

First, the claim that white businesses are being barred from all government contracts below R20 million is materially misleading.

The legislation creates a designated-category set-aside system. Some categories are race-based; others are not. The R20 million figure is a threshold for the mechanism, not a blanket prohibition on white participation. The 30% black-people target is a minimum annual procurement-budget target for that category, not a rule that all sub-R20 million contracts are black-only.

Second, the existence of preferential procurement and set-asides is not inherently unconstitutional.

Sections 9(2) and 217(2) expressly recognise remedial and preferential measures. The constitutional argument must concern their design, effect and implementation,  rather than the mere fact that preference exists.

Third, parts of the proposed design deserve serious reconsideration.

The 100% ownership rule is particularly rigid. Government should be required to demonstrate why absolute category ownership produces better empowerment outcomes than a model permitting meaningful majority ownership, partnership, investment and skills transfer.

The administrative implications of absolute identity-based eligibility also require scrutiny. Legislative clarity on defining and validating historically advantaged and disadvantaged classes should be urgently obtained.

Fourth, South Africa needs to move the transformation debate over time from duration to outcomes.

Thirty years is neither proof that redress is no longer required nor justification for its indefinite continuation unchanged.

The proper question is what measurable disadvantage remains and whether the current intervention is reducing it.

17. Recommendations

The Institute proposes that government and Parliament consider the following approach.

The proposed procurement framework should retain legitimate preferential mechanisms where material disadvantage remains, but the 100% ownership requirement should be reviewed against evidence demonstrating whether such an absolute threshold is necessary to produce superior empowerment outcomes.

National Treasury should publish a clear rationale for each set-aside target and ownership threshold, including the economic and supplier-development outcomes expected from the measure.

Transformation policy should increasingly reward measurable capability-building: enterprise growth, skills, jobs, capital formation, supplier development, technology transfer, management participation and sustainable access to private markets.

Government should explicitly encourage genuine partnerships between established and emerging entrepreneurs rather than create incentives that may discourage mixed ownership or external investment.

The procurement system should establish transparent indicators against which the effect of race-conscious interventions is periodically assessed.

Finally, Parliament should develop an outcomes-based sunset framework under which the intensity of race-conscious measures diminishes as independently measured disadvantage diminishes. This could operate either through continuous outcomes-based adjustment or through a provisional sunset date subject to renewal on clearly defined, evidence-based grounds.

Neither approach should predetermine when transformation ends merely through the passage of time. Both should establish how South Africa will determine whether a particular form of redress remains necessary, should be modified, or has succeeded.

18. Conclusion: successful redress should make itself

less necessary

South Africa should resist two temptations.

The first is to allow legitimate frustration with transformation policy to become denial of the continuing consequences of historical exclusion.

The second is to allow legitimate commitment to redress to turn temporary remedial categories into permanent features of citizenship and economic participation.

The current procurement controversy demonstrates both dangers.

Claims that whites are simply being expelled from all state procurement below R20 million do not survive careful reading of the legislation and regulations. Such misinformation inflames racial anxiety without improving the policy debate.

But the exaggeration of critics must not become an excuse for dismissing legitimate concerns. A system that insists upon 100% category ownership, disqualifies businesses that fall marginally outside that requirement and relies increasingly on precise identity classification deserves careful constitutional, administrative and economic scrutiny.

South Africa still needs transformation.

But transformation should have a destination.

That destination cannot sensibly be determined by saying that thirty years, forty years or fifty years is enough. Time does not measure equality.

Neither can the destination remain undefined.

A democratic state should be able to explain what disadvantage an intervention seeks to overcome, how progress will be measured, whether the intervention is working, and what evidence will justify changing or ending it.

The objective should be a deliberate progression: from redress, to empowerment, to capability, to normalised opportunity.

The most successful transformation policy will not be the one that perfects the permanent administration of racial categories.

It will be the one that progressively creates the conditions in which those categories are needed less.

That is not a retreat from redress. It is what successful redress should ultimately achieve.

References

AllPay Consolidated Investment Holdings (Pty) Ltd and Others v Chief Executive Officer of the South African Social Security Agency and Others [2013] ZACC 42.

Department of Trade and Industry. 2013. Amended Codes of Good Practice under the Broad-Based Black Economic Empowerment Act, Government Gazette No. 36928, 11 October 2013. Pretoria: Government Printer.

Minister of Finance and Another v Van Heerden [2004] ZACC 3.

Minister of Finance v Afribusiness NPC [2022] ZACC 4.

National Employers’ Association of South Africa (NEASA). 2026. ‘Exclusionary race-based tendering rejected: Public Procurement Regulations 2026: NEASA submits comments’, 16 July 2026. Online publication.

National Treasury. 2026. Draft General Public Procurement Regulations, 2026. Pretoria: National Treasury.

Republic of South Africa. 1996. Constitution of the Republic of South Africa, 1996, as amended. Pretoria: Government Printer.

Republic of South Africa. 2024. Public Procurement Act 28 of 2024. Government Gazette No. 50967, 23 July 2024. Cape Town: Government Printer.

South African Police Service v Solidarity obo Barnard [2014] ZACC 23.

Bernstein, A. 2026. ‘Procurement reforms risk making the state less effective’. Business Day, 16 July 2026. Available at: https://www.businessday.co.za/opinion/2026-07-16-ann-bernstein-procurement-reforms-risk-making-the-state-less-effective/

Bethlehem, L. 2026. ‘Procurement overhaul could sideline many qualified suppliers’. Business Day, 7 May 2026. Available at: https://www.businessday.co.za/opinion/columnists/2026-05-07-lael-bethlehem-procurement-overhaul-could-sideline-many-qualified-suppliers/

Khumalo, K. 2026. ‘Public procurement rules polarise business groups’. Business Day, 24 April 2026. Available at: https://www.businessday.co.za/news/2026-04-24-public-procurement-rules-polarise-business-lobby-groups-along-racial-lines/

Kersten, K. 2026. ‘R100 billion spent on BEE skills development and nothing to show for it’. NewsDay, 28 February 2026.

Koteli, T.C. 2026. ‘BLSA CEO questions whether changes to BEE will grow South Africa’. The Citizen, 12 May 2026.