Why enforcing VAT on imported electronic services should come before a digital services tax in South Africa

Introduction

A South African household can now watch a series, rent cloud storage, buy an app and sit an online course in a single evening without dealing with a single company that has an office, an employee or a warehouse in the country. That is the basic problem. A value-added tax built around supplies made here by people who are here does not catch much of this, and a foreign platform earning rands from Johannesburg users without a branch in Sandton also sits outside most of the income tax net, at least under the rules currently on the books (Ngidi, 2022; Mpofu, 2022a). The scale is not trivial either. In a sample of 22 countries, imports of digital services grew by a weighted average of nearly 60 per cent between 2010 and the IMF’s 2021 guidance note (Brondolo, 2021).

South Africa moved early. Foreign suppliers of electronic services have had to register and charge VAT since 1 June 2014, and the regime has been reworked several times since, most recently through regulations that took effect on 1 April 2025 (SARS, 2025; Republic of South Africa, 2025). Meanwhile a good part of the developing-country debate has drifted somewhere else, towards digital services taxes, significant economic presence rules and the OECD two-pillar deal (Quiñones et al., 2025; Mpoha, 2022). It is easy to see why. Those instruments promise a share of profits, not just a share of consumption.

This paper argues that South Africa should keep its attention on making the VAT regime for imported electronic services work, and that the weak point is enforcement far more than the rules on paper. The 2025 amendments narrow the base and ease some burdens, which may be sensible (Loffstadt et al., 2023), but they also lean harder on data matching and on platforms and payment channels doing part of the collecting (Brondolo, 2021). A digital services tax looks like a second front opened before the first has been shown to hold (Mpofu, 2022a; Mpoha, 2022). The sections below set out the current legal position, ask how well registration-based collection suits a developing economy, look at what the 2025 and 2026 changes actually do, and then weigh the direct tax alternatives.

Where the law stands in 2026

The basic scheme sits in section 1(1) of the Value-Added Tax Act 89 of 1991. A foreign supplier of electronic services is treated as running an enterprise in South Africa where two of three things are true: the customer is a South African resident, the payment comes from a bank registered or authorised here, or the customer has a business, residential or postal address in the country (SARS, 2025). Once taxable supplies pass the registration threshold in section 23(1A), the supplier has to register and charge VAT. This is the vendor collection model that, by the IMF’s count, more than 60 countries use in one form or another (Brondolo, 2021). The threshold began at R50,000 in 2014 and moved to R1 million on 1 April 2019, when foreign intermediaries were pulled into the net as well (SARS, 2025).

The next step needs a hedge. SARS’s current guide says the section 23(1A) threshold rises to R2.3 million from 1 April 2026, but the same guide carries a stray reference to R120,000 for that date, which is probably the voluntary registration figure and is not explained (SARS, 2025). SARS’s Budget 2026 questions and answers, which deal with the general VAT threshold, add that the amendments had not yet been promulgated when they were written, even though applications were already being processed at the higher figure (SARS, 2026). So the R2.3 million figure is the working position, and anyone relying on it for a particular supplier would want to check the Act as amended.

The second layer is the regulations. The Minister of Finance replaced the older prescription of electronic services with new regulations published in Government Gazette 52293 on 14 March 2025, in force from 1 April 2025. They describe electronic services as services supplied for payment through an electronic agent, electronic communication or the internet, and they carve out four categories: certain regulated educational services, telecommunications services, some intra-group services developed for a resident group company’s exclusive use, and services supplied solely to South African VAT vendors (Republic of South Africa, 2025). That last exclusion matters. Because it says solely, a foreign supplier with even a handful of consumer customers appears to stay inside the net (Republic of South Africa, 2025; SARS, 2025).

There is also a procedural change that gets less attention than it deserves. Since 24 December 2024 a foreign supplier or intermediary may appoint a representative vendor, and from 1 April 2025 the appointment is compulsory, whether the representative is resident or not. Non-residents from treaty countries that meet set conditions no longer need a South African bank account (SARS, 2025). From 5 January 2023 a supplier that crosses the threshold only because of temporary abnormal circumstances need not register (SARS, 2025). Taken together these are mostly administrative tweaks, but Loffstadt, Ndlovu and Padia (2023) found the VAT Act broadly in line with international practice on intermediaries, which suggests the design is not where the difficulty lies.

Why registration-based collection fits, and where it strains

The case for vendor collection starts from a failure. Consumer self-assessment, where households would declare and pay VAT on their own imports, has proved unworkable in most places, and intangible services cannot be stopped at a border post the way a parcel can (Brondolo, 2021). Pushing the duty onto the supplier is the fallback. In Australia, New Zealand, Norway and Singapore most overseas suppliers reportedly complied voluntarily, helped by a threshold and active outreach by the tax authority (Brondolo, 2021). A review of 55 studies on African consumption taxes found the same attractions in using existing VAT rules, namely higher revenue and efficiency gains, alongside unclear legislation, limited capacity and gaps in tax knowledge (Mpofu, 2022b). South Africa’s regime follows that model closely enough that a comparison with OECD guidance and with the New Zealand and Australian rules found only a few differences (Loffstadt et al., 2023).

The worry is that the model was built for places where it basically works. Habaasa (2023) argues that the OECD approach suits countries with high e-commerce volumes but not most African ones, where foreign businesses have little reason to register and little to fear from ignoring the rules. That is a pointed claim and it deserves a straight reading. It concerns incentives, and incentives are hard to observe. None of the studies used here reports what share of liable foreign suppliers actually register in South Africa, so the question stays open. South Africa probably differs from the typical case in one respect: the enterprise definition leans on payments from banks registered here (SARS, 2025), and the IMF notes that revenue authorities can use payment data to find customers and can instruct banks to redirect money owed by non-compliant suppliers (Brondolo, 2021).

Regional comparison adds a second layer. Diouf and Niesten (2025) find wide variation across Africa in how reforms are designed and carried out. Senegal’s 2024 reform, their case study, uses simplified registration and invoicing and does not require a local representative, and they list the absence of simplified compliance regimes and of thresholds for non-resident operators among the common gaps. South Africa has a threshold and, since April 2025, a compulsory representative vendor (SARS, 2025), which is a heavier demand on foreign suppliers than Senegal makes. Diouf and Niesten also flag the burden on small foreign businesses as an unresolved challenge. Uganda began collecting VAT from non-resident providers of electronic services to non-taxable persons on 1 July 2022 (Kabagumya, 2022). If every African market adds its own registration, representative and invoicing rules, the cost for a small foreign supplier multiplies, which is presumably why Diouf and Niesten (2025) press for harmonised regimes.

Put these together and the picture is mixed. The design is mainstream (Loffstadt et al., 2023), the legal basis is settled and the payment hooks are real. What nobody seems to have shown, at least in the sources reviewed here, is how well it performs, and Habaasa (2023) gives some grounds for doubt. Enforcement capacity is where Brondolo (2021) is most cautious, estimating that a reform takes roughly two years, with at least 12 months needed just to get systems and payment channels ready. South Africa’s regime is more than a decade old, so it should be past that stage. The open question is whether the back-office work behind it, the default assessments, third-party data and marketplace reporting, has kept up with the legal changes of 2025.

What the 2025 and 2026 changes do

The vendor-only exclusion looks like an answer to something Loffstadt, Ndlovu and Padia (2023) recommended, namely a clearer distinction between business-to-business and business-to-consumer sales in the VAT Act. The logic is familiar from destination-based VAT. A registered business can account for the tax itself, so asking a foreign supplier to register for sales that raise little net revenue adds cost without much return (Republic of South Africa, 2025; Loffstadt et al., 2023). Still, the word solely creates a cliff. A supplier whose customers are almost all registered vendors but includes a few consumers must treat everything as in scope, and presumably needs some way of checking each customer’s registration status. Whether that is easier or harder than registering is hard to say without supplier-side evidence.

The threshold increase is the more debatable change. There is a respectable argument for it. The IMF generally supports a registration threshold set near the domestic one so that foreign suppliers are not treated worse than local ones (Brondolo, 2021), and a higher figure keeps small foreign sellers out of a system that costs them more than it yields. But Diouf and Niesten (2025) list the absence of thresholds for non-resident operators among the common gaps in African regimes, which suggests a threshold is treated as normal good practice and the real argument is about where to set it. On the revenue side, a higher threshold could matter more than it first looks if many small suppliers sell through a few large platforms, because the IMF notes that treating marketplaces as deemed suppliers keeps the number of registrants manageable, up to about 1,000 entities, and improves compliance efficiency (Brondolo, 2021). South Africa has brought intermediaries into scope since 2019 (SARS, 2025), so the platform route is already there in principle.

Then there is the distributional worry. Mpofu (2022b) cautions that VAT on digital services could raise the cost of services, limit access and deepen inequality, with knock-on effects for digital financial inclusion and poverty reduction. It is a fair warning, and it needs weighing against the alternative. The IMF’s own account of why countries tax imported digital services is partly about fairness to domestic businesses that must already charge VAT on comparable supplies (Brondolo, 2021). Whether the incidence worry bites hard in South Africa is open to doubt, since the services concerned are mostly paid for by cards and bank accounts, but that should not be pushed too far without household data. What the cautious reading does support is a sense of proportion. VAT on a streaming subscription is a modest sum and a modest design problem, and it should not be dressed up as a revenue breakthrough.

Direct taxes and the digital services tax option

The income tax side has a different problem, which is nexus. Ngidi (2022) argues that South Africa’s permanent establishment definition, which follows the OECD model and enters the Income Tax Act through section 1, effectively demands physical presence, so a business serving South African customers remotely can earn here without being taxable here. The proposals are a digital establishment rule in treaties, plus amendments to the permanent establishment definition in section 1 and the source rules in section 9, but Ngidi also warns against uncoordinated unilateral action that could discourage foreign investment. The article says outright that its focus is corporate income tax, so it tells us little about VAT, which is where Loffstadt et al. (2023) and Brondolo (2021) do their work. That is worth keeping in mind, because it means the income tax and VAT literatures are largely talking past each other.

Digital services taxes sit in between. Mpofu (2022a), in a critical review, finds mixed results for direct digital service taxes in Africa, with a real prospect of revenue and growth set against the risk of harmful side effects where policy is badly designed, and notes that African digital tax laws are still at an early stage. The most optimistic numbers come from Quiñones et al. (2025), who report that a 3 per cent digital services tax could raise an estimated €7.7 to €11.8 billion for South Centre member states, against about €6.9 billion from the OECD’s Amount A. They also record rates ranging from 1.5 to 7 per cent, a Kenyan digital services tax from 2019 followed by a significant economic presence tax in December 2024, and a Nigerian presence tax from 2020. Those figures are estimates, produced by a paper that openly argues for taxing digital income at source, so they are best read as one side of an argument.

The UN route is not a clean escape either. Mpoha (2022) examines Article 12B of the UN Model, a simplified provision on automated digital services, and concludes that its weaknesses outweigh its strengths for developing countries, so that other alternatives should be considered. As far as the official material consulted for this paper shows, South Africa has not adopted a digital services tax, and the Budget 2026 questions and answers talk about a proposed levy on online gambling revenue rather than a general digital levy (SARS, 2026). Nothing in the literature suggests that a South African digital services tax would be easy to design, and Quiñones et al. (2025) themselves treat design and administration as matters for peer learning and harmonisation. A second regime applied to the same foreign firms that already face VAT registration would add compliance cost on top of an enforcement question nobody has answered.

None of this makes direct taxation unimportant. Profit taxation is where the large sums are, and Ngidi’s (2022) point about the permanent establishment definition is probably right in its own terms. The sensible reading is a matter of sequence. VAT is already law, SARS has a decade of experience with it, and the evidence from the IMF suggests the collection tools are known (Brondolo, 2021). Direct tax reform needs treaty renegotiation and international agreement, which is slow by nature (Ngidi, 2022; Mpoha, 2022). It makes more sense to push on the first while the second plays out.

Conclusion

South Africa has a VAT regime for imported electronic services that is mainstream in design, anchored in the Value-Added Tax Act and the March 2025 regulations, and consistent with international guidance on most points that comparative studies have checked (Loffstadt et al., 2023; Brondolo, 2021). Its weak point is the one Habaasa (2023) identifies for Africa generally, which is whether foreign suppliers have enough reason to register and enough reason to fear being caught if they do not. The 2025 changes narrow the base, make a representative vendor compulsory and, on SARS’s current guidance, lift the threshold, and each can be defended while together shifting more weight onto enforcement tools that nobody has yet shown to be working.

Three practical points follow. SARS could publish collection and registration figures for foreign electronic services suppliers, since the gap in evidence is the thing that most limits any assessment. It could lean on payment data and platform reporting, the levers the IMF treats as most useful (Brondolo, 2021). And it could keep an eye on what neighbours do, because small foreign suppliers face a patchwork unless regimes converge (Diouf and Niesten, 2025). A digital services tax, in the meantime, can wait. The treaty and permanent establishment work Ngidi (2022) describes looks the more useful long-run investment, and the case for a unilateral levy has not been made well enough, on the evidence available, to justify the extra layer.

Reference list

Brondolo, J.D. (2021) Administering the value-added tax on imported digital services and low-value imported goods. Technical Notes and Manuals 2021/004. Washington, DC: International Monetary Fund. doi:https://doi.org/10.5089/9781513576480.005

Diouf, A. and Niesten, H. (2025) Tracking and taxing cross-border digital transactions: the case of VAT in Senegal. ICTD African Tax Administration Paper 42. Brighton: Institute of Development Studies. doi:https://doi.org/10.19088/ICTD.2025.043

Habaasa, A. (2023) VAT and the cross-border digital economy in Africa: an issue of non-compliance or design? International VAT Monitor, 34(6). doi:https://doi.org/10.59403/3nrdb56

Kabagumya, F. (2022) Tax on tech giants for electronic services. International Transfer Pricing Journal, 29(6). doi:https://doi.org/10.59403/1c4sedw

Loffstadt, A., Ndlovu, J. and Padia, M. (2023) Do South Africa’s e-commerce VAT rules measure up to international trends and OECD guidelines? Journal of Economic and Financial Sciences, 16(1). doi:https://doi.org/10.4102/jef.v16i1.815

Mpofu, F.Y. (2022a) Taxation of the digital economy and direct digital service taxes: opportunities, challenges, and implications for African countries. Economies, 10(9), article 219. doi:https://doi.org/10.3390/economies10090219

Mpofu, F.Y. (2022b) Taxing the digital economy through consumption taxes (VAT) in African countries: possibilities, constraints and implications. International Journal of Financial Studies, 10(3), article 65. doi:https://doi.org/10.3390/ijfs10030065

Mpoha, J.W. (2022) Article 12B of the UN Model (2021): a simplified solution for developing countries to tax income from the digital economy? Bulletin for International Taxation, 76(5). doi:https://doi.org/10.59403/366azk

Ngidi, M. (2022) Determining ‘permanent establishment’ in the digital economy epoch: a case for South Africa. Obiter, 43(3), pp. 509-535. doi:https://doi.org/10.17159/obiter.v43i3.14882

Quiñones, N. et al. (2025) The taxation of the digital economy in practice: digital services taxes and other measures. Research Paper 226. Geneva: South Centre. Available at: https://www.southcentre.int/wp-content/uploads/2025/11/RP226_The-Taxation-of-the-Digital-Economy-in-Practice_EN-1.pdf (Accessed: 18 October 2026).

Legislation and case law

Republic of South Africa (1991) Value-Added Tax Act 89 of 1991. Available at: https://www.sars.gov.za/legal-counsel/primary-legislation/ (Accessed: 18 October 2026).

Republic of South Africa (2025) Regulations prescribing electronic services for purposes of the definition of ‘electronic services’ in section 1(1) of the Value-Added Tax Act, 1991. Government Notice 5993, Government Gazette 52293, 14 March 2025. Available at: https://lawlibrary.org.za/akn/za/act/gn/2025/5993/eng@2025-03-14 (Accessed: 18 October 2026).

South African Revenue Service (SARS) (2025) Supply of electronic services by foreign suppliers and foreign intermediaries: external guide (VAT-REG-02-G02), Revision 07. Available at: https://www.sars.gov.za/vat-reg-02-g02-supply-of-electronic-services-by-foreign-suppliers-and-foreign-intermediaries-external-guide/ (Accessed: 18 October 2026).

South African Revenue Service (SARS) (2026) Budget 2026 frequently asked questions. Available at: https://www.sars.gov.za/about/sars-tax-and-customs-system/budget/budget-2026-frequently-asked-questions/ (Accessed: 18 October 2026).