Music Publishing in South Africa: R630.2m Allocated, R419.7m Actually Paid

South Africa’s performing-rights society allocated R 630,167,000 in royalties to composers, authors and publishers in the year to 30 June 2025, and paid R 419,718,000 out in cash to members and sister societies. The difference, R 210,449,000, is 33.4 per cent of the year’s allocation, and it is not a one-year event: the same gap was R 193,062,000 the year before, or 31.0 per cent.

Those figures are transcribed from SAMRO’s audited summary consolidated annual financial statements for the years ended 30 June 2024 and 30 June 2025, signed off by an independent registered auditor. Company-level figures are used throughout, not group, because licence revenue is a company line and the group also consolidates two property subsidiaries whose business is the society’s own building in Braamfontein, Johannesburg, not music licensing. This measures performing rights only. Mechanical rights in South Africa sit with CAPASSO and recorded-music neighbouring rights with SAMPRA, and neither is in this measurement.

#	Financial year to 30 June	Total company licence revenue, ZAR	Royalties accrued to creators and publishers, ZAR	Paid in cash to members and sister societies, ZAR	Accrued but not paid in cash that year, ZAR	Share of the year's accrual not paid in cash, %	Share of the year's accrual paid in cash that year, %
1	FY2025	680.35m	630.17m	419.72m	210.45m	33.40	66.60
2	FY2024	683.84m	622.22m	429.15m	193.06m	31	69

Source: SAMRO (Southern African Music Rights Organisation NPC) Integrated Report 2025, audited summary consolidated annual financial statements for the year ended 30 June 2025, audited by SizweNtsalubaGobodo Grant Thornton Inc., retrieved 22 Aug 2026 · download the data

Sponsored

The allocation went up while the cash went down. Royalties accrued to rightsholders rose from R 622,215,000 to R 630,167,000. Cash paid to members and sister societies fell from R 429,153,000 to R 419,718,000. Licence revenue itself barely moved, from R 683,843,000 to R 680,350,000.

One qualifier belongs in the body rather than in a footnote. The R 419,718,000 is not money paid to South African writers. It is money paid to members and to sister societies, meaning collecting societies abroad owed for foreign music played in South Africa. The accrual of R 630,167,000 is not domestic-only either. The society does not disclose either split, so what is measured here is a blended rate across domestic and foreign rightsholders, and 33.4 per cent is not the figure for a South African composer specifically. That is the honest limit, and it cuts both ways: the domestic rate could be better or worse than the blend depending on a split nobody outside the organisation can see. The society’s own account of why the money is stuck, which is works that cannot be matched to a named person, points at the domestic side rather than at the foreign one, but this desk has not measured that and does not claim it.

What music publishing in South Africa collects, and from whom#

The money does not mostly come from streaming. Broadcasting, meaning radio and television, is 55 per cent of licence revenue. General licensing, which is clubs, bars, malls, restaurants and live venues, is 26 per cent. Digital, which is every streaming and online platform combined, is 15 per cent. Foreign royalty income, collected by sister societies abroad under bilateral agreements, is 4 per cent.

That 4 per cent is worth naming in rand, because it is the export line. Foreign royalty income was R 31,200,000 in FY2025, down from R 35,800,000 the year before. Local licensing was R 649,200,000 against R 648,000,000. A society representing 152,565 composer and publisher members brought back R 31,200,000 from every other country on earth combined, and that number is falling. It fits what this desk found when it measured South Africa’s own Spotify chart against its income: the country’s rights holders are not, on these measures, capturing much abroad.

The bigger number is on the balance sheet. Distributions payable and in progress, the liability for royalties owed to rightsholders but not yet paid, stood at R 1,286,855,000 at 30 June 2025, up from R 1,150,569,000 a year earlier. That is 189.1 per cent of a full year’s licence revenue, up from 168.3 per cent. Against it the society held cash of R 837,405,000, which covers 65.1 per cent of the liability.

The society is explicit about why the pile exists. It is undocumented works: royalties that have been earned and allocated but cannot be matched to a named rightsholder, because the work was never properly registered or the registration does not match the usage data coming back from broadcasters and platforms. Its own disclosure puts the undocumented balance at 46 per cent of revenue against a 2028 target of 15 per cent, and it discloses that a 2023 forensic investigation into undocumented works has been followed by a further forensic investigation still under way.

The counter-reading#

The obvious objection is timing, and it is a fair one. An accrual and a cash payment in the same twelve months are not the same money. The R 630,167,000 allocated in FY2025 is paid out over subsequent distribution cycles, and the R 419,718,000 paid during FY2025 discharges obligations accrued in earlier years as well. So the gap is not money withheld. It is the amount by which the year’s allocation outran the year’s cash out of the door.

That objection would kill the finding if the gap were a single year’s artefact. Dropping the larger of the two years from the two-year total of R 403,511,000 takes 52.2 per cent off the sum, which is what dropping one of two similar rows always does. The mean barely moves: from R 201,755,500 to R 193,062,000. The two years are close to the same size, so the gap is not a spike in one of them.

Be clear about how weak that test is on its own. Two years is not a series. A robustness check across two rows can tell you that one of them is not carrying the result, and it cannot tell you anything about a trend. The word structural is doing work here that the robustness table cannot support by itself, and it is the balance sheet, not the two-year mean, that earns it.

The balance sheet settles it independently of any single year. If allocations were converting to cash on a normal cycle, the payable balance would not be climbing from R 1,150,569,000 to R 1,286,855,000 while licence revenue fell from R 683,843,000 to R 680,350,000. A backlog that grows while the income feeding it shrinks is not a cycle. It is an accumulation.

The second objection is that this is an efficiency problem, and it is not. The society’s cost-to-income ratio improved from 22.9 per cent to 21.6 per cent, the best in a decade against 29.1 per cent in 2016 and a peak of 40.0 per cent in 2018. Costs are coming down. The money is still not arriving.

Two further limits belong on the page, beyond the sister-societies split already noted. This is one organisation, and one organisation is not an industry: CAPASSO’s mechanical collections and SAMPRA’s neighbouring rights are not measured here, and a songwriter’s total South African income runs through more than one of them. And the intermediate years on the cost-to-income series, 2017 to 2023, are read from the report’s own ten-year chart rather than restated in its prose, which is why that dataset is marked medium confidence while the rest is marked high.

What this means if you write or publish in South Africa#

Do not book an allocation as cash. If you are a publisher modelling South African performing-rights income, the measured conversion of a year’s allocation into a year’s cash was 66.6 per cent in FY2025 and 69.0 per cent in FY2024. A model that assumes an allocation lands in the same financial year has overstated cash by roughly a third, two years running, and the direction of travel is against you. Both conversion rates are blended across domestic and foreign rightsholders, because the society publishes no split, so treat them as the rate for the pool rather than the rate for your own catalogue.

The lever is documentation, not negotiation. The unpaid pile is not a rate dispute and it is not a deduction anyone is arguing about. It is works that cannot be matched to a person. Registration data, split sheets agreed before release rather than after, and consistent title and writer credits across every platform are the only things that move a royalty from the undocumented balance into a payment run. That is unglamorous administrative work, and on these figures it is worth more per hour than any rate renegotiation available to a South African writer.

Weigh the streaming line correctly. Digital is 15 per cent of what this society collects, against 55 per cent for broadcasting. A South African writer optimising purely for streaming is optimising for the smallest of the three domestic streams, and the per-stream economics of that line are thin before anyone else touches them. Radio play still pays the larger share of the publishing pool here, and it is logged from cue sheets and station returns that a writer can check.

And treat foreign income as a target rather than an assumption. R 31,200,000 across the whole membership, falling year on year, is what South African music publishing currently brings home from the rest of the world. Any deal memo that prices meaningful foreign performing-rights collection into a South African signing is pricing something the aggregate data does not yet show.