State-to-state dispute settlement leaves South Africa’s own trade administration to carry the AfCFTA

Introduction

The Agreement Establishing the African Continental Free Trade Area came into force on 30 May 2019, and trade under it was approved to start on 1 January 2021, although the tariff schedules that make preferences usable took years longer (SARS, 2023; tralac, 2026). South Africa gave it local effect by a notice in the Government Gazette on 31 December 2020, which means that for a South African exporter the AfCFTA is, in the first place, a matter of customs rules and tariff schedules (Vinti, 2025). The Agreement and its protocols on trade in goods, trade in services and dispute settlement came first. The protocols on investment, competition policy and intellectual property were adopted in February 2023, and those on digital trade and on women and youth in trade followed in February 2024 (tralac, 2026).

This paper argues that the AfCFTA’s dispute settlement system, being open to State Parties only, cannot be what gives the Agreement its practical force in South Africa. That work falls to South African trade administration and the courts that review it, which are the International Trade Administration Commission (ITAC), the Minister’s powers under the International Trade Administration Act 71 of 2002, SARS and, ultimately, the High Courts. The argument matters because South African scholarship on the AfCFTA has spent much of its energy on continental design, and rather less on the unglamorous domestic procedures through which a trader’s rights are actually decided (Vinti, 2025; Ncube, 2022). The paper looks at how the Agreement entered South African law, what the dispute settlement protocol does and does not offer, how trade remedies show the domestic gaps, and how the softer protocols and the WTO relationship leave the same gaps open. The SADC Tribunal’s history (Shumba, 2022) is a reason for caution about forums that only governments can use, and a short conclusion takes a position.

How the AfCFTA entered South African law

The Constitution says an international agreement binds the Republic once the National Assembly and the National Council of Provinces approve it, but that its provisions become domestic law only when legislation enacts them, unless they are self-executing and approved by Parliament (Constitution of the Republic of South Africa, 1996, s 231). South Africa’s route for trade agreements is mostly administrative. SARS explains that preferential trade agreements listed in Schedule 10 of the Customs and Excise Act 91 of 1964 take effect in South Africa when published by notice in the Gazette, and the AfCFTA appeared as Notice R1433 in Government Gazette 44049 of 31 December 2020, effective from 1 January 2021 (SARS, 2023). Vinti (2025) reads that notice as showing the AfCFTA has been ratified and incorporated into municipal law and that it creates domestic rights.

That reading is plausible for the tariff and customs machinery the notice carries, but it should not be stretched much further without more. A Schedule 10 notice puts the tariff-and-origin commitments into the Customs and Excise Act framework, but nothing in the sources reviewed shows that a South African court has yet treated, say, a protocol provision on non-tariff barriers or trade remedies as directly enforceable by a private party. Vinti (2025) also notes that the WTO Agreement on Safeguards, which is binding on South Africa but not incorporated, is used by courts as an interpretive aid under section 233 of the Constitution. If the AfCFTA’s trade remedy rules are read the same way, they will shape decisions by ITAC and by judges without giving anyone a freestanding claim (Vinti, 2021; Tavuyanago and Vinti, 2023).

The Minister’s own policy powers add another layer. Section 5 of the International Trade Administration Act lets the Minister of Trade, Industry and Competition issue trade policy directives, and Vinti (2021) argues that the process has to meet section 195 of the Constitution, which means public participation and openness, so that policy is formulated with procedural rationality and avoids arbitrariness. That is a useful point for AfCFTA work. Tariff offers, origin rules and any trade remedy directed at African partners will be shaped through exactly these instruments (SARS, 2023), so the AfCFTA’s South African content is being written inside a domestic administrative law framework, not a continental one.

There is also a timing gap between incorporation and use. The Gazette notice took effect on 1 January 2021, but SARS reported in November 2021 that preferential trade was not yet possible because the tariff schedules were still being negotiated (SARS, 2023). Tralac records that South Africa published its AfCFTA schedule in the customs duty tariff book, dated 15 May 2026, with general notes dated 27 March 2026 (tralac, 2026). So the legal instrument and the published tariff schedule were about five years apart. Anyone asking what the AfCFTA meant for a South African business in that period was really asking what the tariff book and SARS said, a question no continental body was positioned to answer.

What the dispute settlement protocol gives and withholds

On paper the continental mechanism looks serious. Article 20 of the Agreement sets up a dispute settlement mechanism for disputes between State Parties, and the Protocol on Rules and Procedures on the Settlement of Disputes defines a dispute as a disagreement between State Parties about the interpretation or application of the Agreement (AfCFTA Agreement, 2018, art 20; Protocol on Rules and Procedures on the Settlement of Disputes, 2018, art 1). A complaining State must first ask for consultations, and the other side has 10 days to reply and 30 days to begin talking; if nothing is settled within 60 days a panel may be requested (Protocol, 2018, art 7). Panels of three or five are meant to report within five months, a standing Appellate Body of seven members hears appeals on points of law, and an appeal should take about 60 days and no more than 90 (Protocol, 2018, arts 15, 20 to 22).

Everything in that design assumes the claimant is a government. The visible text of the Protocol refers only to State Parties, panellists who are nationals of the disputing States are excluded, and a State that has used the Protocol on a matter may not take the same matter to another forum (Protocol, 2018, arts 3(4), 10(5)). Article 4(6) adds that the mechanism cannot add to or diminish the rights and obligations in the Agreement. For a South African exporter whose goods are held up at another State’s border over a dubious origin ruling, that means going to the Department of Trade, Industry and Competition and asking it to take up the case. Whether it does is a policy decision, and nothing in the Protocol gives the exporter a right to insist, which is the structural point critics made about the SADC Tribunal’s 2014 Protocol (Shumba, 2022). The implementation stage is no quicker, since the guideline reasonable period for compliance is 15 months and the cap is 18 (Protocol, 2018, art 24).

South Africans have seen this shape before. Shumba (2022) tells how the SADC Tribunal was suspended after it ruled against Zimbabwe in the Campbell land seizures case, and how the 2014 Protocol then confined it to disputes between states only, a change critics said weakened the rule of law and human rights in the region. In Law Society of South Africa v President of the Republic of South Africa ([2018] ZACC 51) the Constitutional Court held that the President’s part in suspending the Tribunal and in signing the 2014 Protocol was unconstitutional, unlawful and irrational, and ordered him to withdraw his signature. The AfCFTA is not the SADC Tribunal, and the two arrangements differ in origin and purpose. Even so, South African constitutional practice has already shown suspicion of arrangements where access to a regional forum depends on governments alone (Shumba, 2022).

Where the real law is being made: trade remedies

The best evidence for the argument is the AfCFTA’s trade remedy regime. The Protocol on Trade in Goods deals with anti-dumping, global safeguards and preferential safeguards in Annex 9, and Vinti (2025) shows that Annex 9 defers heavily to GATT Article XIX and the WTO Agreement on Safeguards. It says safeguards are applied in accordance with those rules, caps measures at four years with a possible extension of up to four more, and says nothing about the procedure for an extension. The AfCFTA Guidelines that would fill that gap had not been adopted when he wrote. So until they arrive, a South African trade remedy affecting African partners will be assessed against WTO law and South African procedure, not against any detailed continental code, which sits with Adams and Lenaghan’s (2022) wider point that the AfCFTA’s relationship with WTO law is still being worked out.

South African procedure has its own holes. ITAC’s extension guidelines, reissued in 2021 under section 60(1) of the International Trade Administration Act, do not require an investigation to finish before the measure lapses, do not say whether the duty continues while it runs, and do not bind ITAC or the courts, because section 60(2)(b) lets them be departed from (Vinti, 2025). The Macsteel episode shows what that can cost. A duty on flat hot-rolled steel that started at 12 per cent was due to expire on 10 August 2020, the application came about 17 days before lapse, and on 7 August the Deputy Minister of Finance purported to extend the duty at 8 per cent for a year, after which a correction notice recast the extension as a maintenance of the measure (Vinti, 2025). Vinti argues that there was no completed investigation, no progressive liberalisation and no mid-term review, and that duties paid should have been refunded. None of that would have been fixed by a continental panel.

The courts are the other half of the picture. Tavuyanago and Vinti (2023) look at the national security clause in section 18A of the Competition Act 89 of 1998 and argue that an aggrieved party can challenge a security-based merger prohibition in the higher courts even though the Act gives no explicit route, relying on legality, legal certainty and the courts’ inherent jurisdiction, and drawing on the WTO panel decision in Russia – Measures Concerning Traffic in Transit. The subject is competition, not the AfCFTA. Still, the reasoning travels, because it shows South African scholars treating domestic review as the safety net where a statute or treaty is silent. Add ITAC’s role as the de facto investigating body for the Southern African Customs Union while the SACU Tariff Board is not operating (Vinti, 2025), and a good deal of the AfCFTA’s South African future sits with one commission and a handful of High Court judges.

Soft text and the WTO shadow

The softer protocols repeat the pattern. Mupangavanhu (2024), working from the final draft of the AfCFTA Protocol on Intellectual Property Rights, points out that the word shall appears in every paragraph of article 11 except article 11(6), where the States merely agree to comply with their international obligations on access to published works for visually impaired persons. So the text falls short of requiring ratification of the Marrakesh Treaty, and it says nothing express about cross-border exchange of accessible copies. Ncube (2022) had suggested the IP protocol would follow the architecture of the Agreement and its principles, build on the Tripartite Free Trade Area’s phase two IP agenda and be broadened to address gender and development. The result, as Mupangavanhu reads it, balances public and private interests on paper while leaving States to decide what to do.

Investment is similar. Qumba (2021) argued that the investor-state dispute settlement system the AfCFTA is expected to adopt will be shaped by the legitimacy crisis in investment arbitration, and proposed that the Investment Protocol require investors to exhaust local remedies first. That proposal puts domestic courts at the front of the process, which fits the thrust of this paper, and it was made before the protocol was adopted. The tralac status page reports that the working procedures of the investment dispute annex were still under negotiation in 2026 (tralac, 2026). Antifon (2025) makes a related point about silence, noting that the Agreement has no explicit trade-climate provisions and that any climate role has to be drawn out through a purposive reading. A text that is silent or soft pushes interpretation back to the officials and judges applying it.

The WTO sits behind all of this. Adams and Lenaghan (2022) assessed the African Union’s proposed 0.2 per cent levy on imports from non-members and concluded that it probably breaches the most-favoured-nation rule in GATT Article I, noting that as of October 2020 the AfCFTA had not been notified to the WTO as a regional trade agreement, and that only 28 states had ratified it. They suggested seeking approval under the Enabling Clause in the short term and meeting the requirements of GATT Article XXIV in the longer term. Whether that has happened since is something the sources reviewed do not settle. What matters here is that the AfCFTA borrows WTO law as its baseline in trade remedies (Vinti, 2025) while its own WTO status was still unsettled when Adams and Lenaghan wrote, so officials applying Annex 9 are leaning on a framework whose relationship to the Agreement is not fully resolved.

Practical plumbing matters as much as legal text. Monye and Monye (2022) describe the proposed Pan-African Payment and Settlement System as a way to move cross-border payments in local currencies in real time, cutting the cost of correspondent banking and the effects of currency inconvertibility. A trader who cannot be paid cheaply is not helped by a clean origin certificate. Rules of origin deserve a similar mention. Tralac reports that the outstanding automotive and clothing and textile rules were adopted in February 2026, so that all product-specific rules are now agreed (tralac, 2026). That should matter to South Africa’s car exporters, but questions about whether a given vehicle qualifies will be decided first by customs authorities, SARS at home and its counterparts abroad, through certificates and verification, and not before the Dispute Settlement Body (SARS, 2023).

Conclusion

The AfCFTA is often described through its institutions in Addis Ababa and its panels and Appellate Body. That description is accurate as far as it goes, but it misses where most South African disputes will actually be decided. A State-only mechanism with a 60-day consultation phase, panels that take months and an implementation window of up to 18 months is a diplomatic tool, and the exporter or importer depends on the Department to use it (Shumba, 2022). Meanwhile the daily law of the AfCFTA in South Africa is being made by Gazette notices, ITAC guidelines, ministerial directives and customs decisions, all of them reviewable in South African courts (Vinti, 2025).

If that is right, reform energy should go where the law is working. Vinti (2025) recommends making ITAC’s guidelines binding or turning them into regulations, and setting out an extension procedure for both AfCFTA and non-AfCFTA trade, and those proposals look sensible and cheap. South Africa could also say more clearly which AfCFTA provisions private parties may rely on, and keep the public participation that Vinti (2021) ties to section 195 as tariff offers and directives are made. None of this makes the continental mechanism unimportant, only incomplete. The risk is that officials, scholars and traders wait for a panel to settle questions that a Gazette notice and a High Court will settle first.

Reference list

Adams, G. and Lenaghan, P.M. (2022) The African Union’s self-financing mechanism: a critical analysis in terms of the World Trade Organisation’s rules and regulations. Potchefstroom Electronic Law Journal, 25, article 9472. doi:https://doi.org/10.17159/1727-3781/2022/v25i0a9472

Antifon, V. (2025) The motivation of climate action through free trade agreements: what does the AfCFTA Agreement bring to the 2015 Paris Agreement on climate change. Potchefstroom Electronic Law Journal, 28, article 15060. doi:https://doi.org/10.17159/1727-3781/2025/v28i0a15060

Monye, O. and Monye, E. (2022) Regional integration in Africa: proposals for an Africa-wide payment system. Law, Democracy and Development, 26, pp. 1-27. doi:https://doi.org/10.17159/2077-4907/2021/ldd.v26.8

Mupangavanhu, Y. (2024) Copyright law and human rights: Blind SA v Minister of Trade, Industry and Competition and its potentially far-reaching implications. South African Journal on Human Rights, 40(3-4), pp. 313-326. doi:https://doi.org/10.1080/02587203.2024.2437025

Ncube, C.B. (2022) Intellectual property and the African Continental Free Trade Area: lessons and recommendations for the IP protocol. Journal of International Trade Law and Policy, 21(2), pp. 105-121. doi:https://doi.org/10.1108/JITLP-09-2021-0051

Qumba, M.F. (2021) The exhaustion of local judicial remedies in investor-state dispute settlement: a proposal for the African Continental Free Trade Agreement on Investment Protocol. Law, Democracy and Development, 25, article 6. doi:https://doi.org/10.17159/2077-4907/2021/ldd.v25.6

Shumba, T. (2022) Rising from its ruins? The Southern African Development Community (SADC) Tribunal. Law, Democracy and Development, 26, pp. 1-24. doi:https://doi.org/10.17159/2077-4907/2022/ldd.v26.11

Tavuyanago, S. and Vinti, C. (2023) Reflections on the justiciability of the ‘national security’ clause as stipulated by section 18A of the Competition Act 89 of 1998: lessons from Russia – Measures Concerning Traffic in Transit WTO panel decision. Law, Democracy and Development, 27, pp. 240-264. doi:https://doi.org/10.17159/2077-4907/2023/ldd.v27.9

Vinti, C. (2021) An exposition of trade policy formulation through the issuing of ‘Trade Policy Directives’ by the Minister of Trade, Industry and Competition under the International Trade Administration Act 71 of 2002. Law, Democracy and Development, 25, pp. 1-23. doi:https://doi.org/10.17159/2077-4907/2021/ldd.v25.10

Vinti, C. (2025) The extension of ‘safeguard measures’ in South Africa within and outside the framework of the Agreement Establishing the African Continental Free Trade Area. Potchefstroom Electronic Law Journal, 28, article 17953. doi:https://doi.org/10.17159/1727-3781/2025/v28i0a17953

Legislation and case law

Agreement Establishing the African Continental Free Trade Area (2018), including the Protocol on Rules and Procedures on the Settlement of Disputes. African Union consolidated text. Available at: https://au.int/sites/default/files/treaties/36437-treaty-consolidated_text_on_cfta_-_en.pdf (Accessed: 18 October 2026).

Constitution of the Republic of South Africa, 1996. Available at: https://www.gov.za/documents/constitution-republic-south-africa-1996 (Accessed: 18 October 2026).

International Trade Administration Act 71 of 2002. Available at: https://lawlibrary.org.za/akn/za/act/2002/71 (Accessed: 18 October 2026).

Law Society of South Africa and Others v President of the Republic of South Africa and Others [2018] ZACC 51 (CCT 67/18), Constitutional Court of South Africa, 11 December 2018. Available at: https://www.saflii.org/za/cases/ZACC/2018/51.html (Accessed: 18 October 2026).

South African Revenue Service (2023) AfCFTA. Available at: https://www.sars.gov.za/legal-counsel/international-treaties-agreements/trade-agreements/afcfta/ (Accessed: 18 October 2026).

tralac (2026) AfCFTA: status of negotiations and implementation. Trade Law Centre. Available at: https://www.tralac.org/resources/by-region/cfta.html (Accessed: 18 October 2026).