South Africa Transfer Pricing Policy

South Africa was one of the earliest African jurisdictions to legislate against transfer mispricing, and the provision has been rewritten more than once since — from a discretionary adjustment power into a self-assessment obligation that puts the arm’s-length test on the taxpayer before a return is filed rather than on SARS afterwards.

The practical consequence is that the question SARS asks is not whether a policy exists but whether it was applied. A group can hold a defensible policy document and still lose the argument because the intercompany invoices for the year do not follow it, and that gap between the written policy and the ledger is the most common finding on a South African review.

South Africa’s treaty network carries the associated-enterprises article, so the arm’s-length principle reaches the same transactions from two directions: domestic law and the treaty. Where a South African adjustment triggers double taxation, the treaty route is the one that resolves it, and the domestic file is the evidence that route depends on.

The statutory position

• Section 31 of the Income Tax Act 58 of 1962 is South Africa’s transfer pricing provision, and it applies to transactions between connected persons as that term is defined in section 1 of the Act.
• SARS Practice Note 7 of 1999 remains the authority’s foundational guidance on how section 31 is applied, and it states expressly that section 31 imposes no hierarchy of methods — while noting that in practice some methods give a more reliable result than others depending on the quality of the available data, and that the comparable uncontrolled price method is preferred where it can be used because it looks directly at the property or service transferred.
• SARS Interpretation Note 127, dated 17 January 2023, deals specifically with the determination of taxable income on intra-group loans.

Those references were read in the primary source — sars.gov.za — on 2026-08-19, and they are quoted because they were read. Legislation is amended and guidance is reissued, so confirm the current text before relying on any of it for a filing position.

SARS expects a company within the reach of section 31 to hold an approved transfer pricing policy that complies with the arm’s-length principle, to be able to demonstrate that compliance, and to show that the policy was actually implemented. That third limb is the one groups underestimate: implementation evidence means invoices, intercompany agreements and management accounts that agree with each other.

Documentation follows the three-tier shape the OECD recommends — a master file describing the group, a local file describing the South African entity’s own controlled transactions, and a country-by-country report where the group is large enough to owe one. Which of the three a particular group owes, and when each is due, turns on the group’s size and where its ultimate parent sits, so we confirm the year’s position against SARS guidance rather than assume it carries over.

Intra-group financing is documented separately in practice. Interpretation Note 127 exists because a loan raises questions a goods-and-services file does not answer — whether the borrower could have carried the debt at all, on what terms an independent lender would have advanced it, and what the arm’s-length consequence is where it could not.

The three-tier documentation shape

The documentation shape is close to universal because BEPS Action 13 designed it that way. A master file sets out the group — structure, intangibles, financing, how it makes its money. A local file covers this entity’s own controlled transactions and the reasoning behind each price. A country-by-country report shows administrations, jurisdiction by jurisdiction, where the revenue, the profit, the tax and the people sit. That last report is owed only above a size threshold, and the OECD’s agreed figure is EUR 750 million of consolidated group revenue, or a near equivalent amount in domestic currency as of January 2015 (OECD, Action 13 country-by-country reporting (oecd.org)). Each jurisdiction legislated its own local-currency equivalent; those are not repeated on this page, because they were not read from the source.

What we confirm before a South Africa file is signed off

• Which documentation tiers this group actually owes for this period, tested against SARS’s current requirements rather than last year’s.
• Whether the transaction list is complete. The goods and the service fee are always on it; the guarantee, the secondment, the shareholder loan and the one-off recharge are the ones that are not.
• That the intercompany agreements, the invoices and the management accounts agree with each other and with the policy the file describes.
• That nothing here contradicts what the group has already filed about South Africa elsewhere. Administrations exchange that data, so the inconsistency arrives before the auditor does.

Because section 31 sets no hierarchy, method selection in South Africa is an evidence argument rather than a compliance box. SARS’s own guidance says the comparable uncontrolled price method is preferred where the data supports it, so a file that reaches straight for a net-margin comparison should say why a direct price comparison was not available — that sentence is the difference between a method choice and an assumption.

The five methods in the OECD framework

Comparable uncontrolled price (CUP)

Takes the price independent parties actually charged each other for the same thing and applies it. Nothing beats it for directness, which is why several administrations state a preference for it, and nothing is harder to find: product, volume, market and contract terms all have to line up before the comparison holds.

Resale price

Derives the transfer price by subtracting an arm’s-length gross margin from the resale price charged to an independent customer. Appropriate where the reseller does not transform the product, and unreliable where the comparison set performs a different mix of marketing, warranty or inventory functions.

Cost plus

Builds the price up from the supplier’s costs plus an arm’s-length mark-up. Used for contract manufacturing, routine assembly and intra-group services. The argument is almost never about the mark-up; it is about which costs belong in the base and whether the cost accounting is consistent year to year.

Transactional net margin (TNMM)

Benchmarks the tested party’s net profit indicator against independent companies doing broadly similar work. It carries more documentation worldwide than the other four combined, precisely because it forgives product differences — and that is also its weakness, because the comparable search then does all the work.

Profit split

Divides the combined profit of the controlled transaction between the parties by reference to their relative contributions. Reached for where both sides make unique and valuable contributions, where the operations are so integrated that neither can be priced on its own, or where each shares in economically significant risk.

No method is correct in the abstract. The one that holds up in South Africa is the one whose selection is reasoned on the record — with the alternatives named and rejected for stated reasons — because a method presented without alternatives reads as a default rather than a choice.

The question a South Africa review turns on is rarely "is this margin in the range". It is "does this describe what the business did". So the work starts with the operating facts and the agreements, not with a comparables database.

What a South Africa engagement covers

• Mapping and delineating the transactions. We list every related-party flow through the South Africa entity and tie it back to the ledger first, because an analysis of an incomplete list is an incomplete analysis.
• Functional and risk analysis. What the South Africa operation actually does, who makes the decisions, and which risks it is genuinely in a position to control — the step that decides whether the rest of the file can stand.
• A reasoned method for each tested transaction, written up with the alternatives that were considered and why they were not used. A method with no alternatives on the page reads as a default.
• Benchmarking with the search shown. Comparable selection, screening decisions and comparability adjustments documented so a reviewer can reproduce them instead of re-running them.
• A file in the form SARS expects, reconciled to the group master file and to whatever has already been said about South Africa in another jurisdiction’s filing.
• Where the numbers come from. Every threshold, deadline and rate that ends up in the file is confirmed against the South African Revenue Service (SARS) for the period concerned, and anything we cannot confirm is described as a mechanism instead of guessed at.

Advance certainty and dispute resolution

South Africa resolves cross-border transfer pricing disputes through the mutual agreement procedure in its treaties, and SARS operates the domestic objection and appeal process for the adjustment itself. Both are slow and both are documentary, which is the argument for getting the file right in the year of the transaction rather than in the year of the audit.

How we work

The fee is agreed in writing before any work starts, and you review the result before anything is filed. Where your South Africa entity sits relative to our offices makes no difference to how the file gets built. If you want to talk it through first, the helpline answers 24 hours a day and there is no obligation attached to a call — contact us or read our transfer pricing service page for what a full engagement includes.

Three things have changed the shape of transfer pricing work everywhere, and they are all visible in South Africa. Tax administrations now receive structured group data before they open a file, so risk is selected on evidence rather than sampled. Intangibles and intra-group financing have displaced goods pricing as the main source of adjustment. And documentation timing has become a substantive issue in its own right, separate from whether the price was right.

Where challenges concentrate in South Africa

• Intra-group debt is the single most audited transfer pricing item in South Africa, and the issue is rarely the interest rate in isolation. It is the debt capacity question underneath it: whether an independent lender would have advanced that amount to that borrower at all, which is an analysis of the borrower’s own cash flows rather than a search for a comparable spread.
• Management and head-office charges recharged into a South African subsidiary attract attention where the benefit to the South African entity cannot be shown. A cost allocation key is not evidence of benefit, and a service that duplicates something the local team already does is the classic disallowance.
• Currency movement makes a policy that was arm’s length at the start of the year fail the test at the end of it. A margin fixed in rand behaves very differently from a margin fixed in the functional currency of the counterparty, and groups that never revisit the policy mid-year end up defending an outcome nobody chose.

What it costs when the file is not there

The cost of a thin file in South Africa is rarely just the tax. It is the adjustment, the charge or penalty attached to it, the years it stays open, the management time an audit consumes, and the corresponding position in the counterparty jurisdiction that may not be relieved. Priced against that, the documentation is the cheap part.

By business size

Owner-managed group

One South African operating company and one offshore holding or IP entity is enough to bring section 31 into play. The obligation is real at this size and the file can be short — what it cannot be is absent.

Mid-market group

Once there are three or four related counterparties the file stops being one analysis and becomes a set of them, each with its own method and its own comparables, tied together by a policy that says how the group is meant to work.

Multinational group

At the size where a country-by-country report is owed, the South African local file has to be consistent with what the group has already told every other authority. Inconsistency between the two is itself the audit trigger.

By industry

Mining and resources

Commodity offtake through a related marketing entity puts the pricing point, the quality adjustment and the marketing function all in issue at once — and the file has to explain what the offshore marketer actually does that a trader would charge for.

Financial services

Intra-group funding, guarantee fees and shared treasury functions each need a separate arm’s-length story, and the guarantee fee is the one most often missing entirely.

Retail and consumer brands

Royalties for a brand developed offshore and deployed in South Africa need a benefit case that survives the question of who built the local brand equity.

Industry is a useful first cut and a poor final answer. What decides a South Africa file is the shape of the group and what the intercompany agreements actually say, which is why two competitors of identical size can face entirely different analyses. Use the sector to know where to start looking.