GOVERNANCE WATCH
By InsidEntity Editorial Desk · Sep 25, 2026 · 10 min read
Series 3 | 25 September 2026
One ordinary-looking number, taken apart until the number that matters is on the table.
7% of the shares. 43% of the votes.
Remgro’s largest disclosed holder of ordinary shares is the Public Investment Corporation, with 16% at the last published register date, 30 June 2025. In August 2026 Ninety One’s clients crossed 5%. Neither number tells you who decides.
Underneath sits a class of shares that never trades. Rupert Beleggings owns all 39,056,987 unlisted B ordinary shares. They are about 7% of Remgro’s shares outstanding, excluding treasury shares, and carry 43% of the votes.
Three numbers, three different stories. 16% is the biggest holding. 43% is the voting power attached to the B shares. 75.6% is how much of the voting ordinary-share pool would have to vote before the B shares stop being a majority of the votes cast.
At a glance
| The number | What it is |
|---|---|
| 16.0% | PIC’s share of the issued ordinary shares, the largest disclosed ordinary holding |
| 7.0% → 43.0% | B shares: economic interest → share of the votes |
| 15.3% → 9.4% | PIC: economic interest → share of the votes |
| 75.6% | Ordinary turnout needed before the B shares stop being a majority of votes cast |
The number
16.04%. The Public Investment Corporation’s holding of Remgro’s 529,217,007 issued ordinary shares at 30 June 2025: 84,876,994 shares. It was the only holder of listed ordinary shares above 5% that year, down from 17.31% a year earlier. The issued ordinary share count did not change between the two dates, so the fall reflects about 6.7 million fewer shares held.
On 17 August 2026 Remgro announced that Ninety One’s clients now hold 5.0156% of the issued ordinary shares. That is roughly 26.5 million shares, held for clients rather than on Ninety One’s own balance sheet.
The question
How much voting influence do those holdings actually create?
Remgro has two classes of shares. Each ordinary share carries one vote. Each unlisted B ordinary share carries ten. The dividend is the same per share on both: the FY2026 final dividend of 422 cents and special dividend of 550 cents were declared on the ordinary and the B shares alike. The difference between the classes is not money. It is votes.
Number 1: 7.0% of the economic interest, 43.0% of the votes
Remgro holds some of its own ordinary shares in treasury, and those carry no vote. Take them out and the register at 30 June 2025 splits like this.
| Share class, 30 June 2025 | Shares | Votes per share | Votes | % of shares | % of votes |
|---|---|---|---|---|---|
| Ordinary, excluding 12,409,667 in treasury | 516,807,340 | 1 | 516,807,340 | 93.0% | 57.0% |
| Unlisted B ordinary (Rupert Beleggings) | 39,056,987 | 10 | 390,569,870 | 7.0% | 43.0% |
| Total | 555,864,327 | 907,377,210 | 100% | 100% |
Each B share votes like ten ordinary shares, so a 7.0% slice of the economics becomes 43.0% of the votes, about six times its weight. An associate company of which Johann Rupert and Anton Rupert are directors also holds 7,551,005 ordinary shares. Add those and the combined position is 8.4% of the economic interest and 43.9% of the votes.
None of this is hidden. Remgro prints the ten-to-one ratio on its shareholder page every year, and the issued ordinary and B share counts have not changed since at least 2022; only the treasury balance has moved. The structure is fully disclosed. The distinction is between economic ownership and voting power.
The 43% also drifts. Treasury shares carry no vote, so an increase in the treasury balance shrinks the ordinary voting pool and lifts the B shares’ share of the votes. Remgro reported the B shares at 42.62% of the total votes at 30 June 2020; on the June 2025 register the same arithmetic gives 43.04%.
Number 2: 15.3% of the economic interest, 9.4% of the votes
The 16.04% is measured against issued ordinary shares. Measured against all shares outstanding, excluding treasury shares, the PIC’s 84,876,994 ordinary shares are 15.3% of the economic interest and 9.4% of the votes.
| Holder | Economic interest | Share of votes |
|---|---|---|
| Rupert Beleggings (B shares) | 7.0% | 43.0% |
| Public Investment Corporation (30 June 2025) | 15.3% | 9.4% |
| Ninety One clients (August 2026, about 26.5 million shares) | 4.8% | 2.9% |
The Ninety One stake is from Remgro’s August 2026 announcement; every other figure is from the 30 June 2025 register. The Ninety One row is indicative until the 2026 register is published.
The PIC owns 2.2 times as many shares as the B class. Its votes are 9.4% of the total, against the B shares’ 43.0%. Even taken together, which they are not, the two largest listed holders would have about 20% of the economic interest and 12.3% of the votes.
Number 3: 75.6%
A South African ordinary resolution passes with more than half of the votes cast. A special resolution needs at least 75%.
If every share votes, the B shares hold 43.0% of the votes. On an ordinary resolution they cannot carry the result alone. On a special resolution, 43% voting against would be enough to stop it, provided the B shares are entitled to vote on that resolution.
But not every share votes. At the 30 June 2025 capital structure, and assuming all B shares vote, the B shares would be a majority of the votes cast unless more than 75.6% of the voting ordinary shares also voted. If the 7.55 million ordinary shares held by the Rupert associate are also assumed to vote with the B shares, the threshold rises to about 78.2% of the remaining ordinary voting pool.
The last two AGMs show why that line matters.
| AGM | Ordinary shares voted | % of issued ordinary | % of voteable ordinary | B shares voted | B shares’ share of votes cast |
|---|---|---|---|---|---|
| 27 November 2025 | 369,118,627 | 69.75% | 71.4% | 39,056,987 | 51.4% |
| 28 November 2024 | 361,288,956 | 68.27% | 70.0% | 39,056,987 | 51.9% |
Shares voted exclude abstentions. The last two columns are InsidEntity calculations at ten votes per B share, on Ordinary Resolution 4 in 2025 and Ordinary Resolution 1 in 2024.
At both meetings, the B-share votes were sufficient for the resolutions to pass despite a majority of the ordinary-share votes cast being against the respective director candidates. In 2025, ordinary shareholders voted 52.53% against Ordinary Resolution 4, the election of Mr P J Moleketi, and 52.30% against Ordinary Resolution 5, the election of Mr M Morobe. In 2024, they voted 53.00% against Ordinary Resolution 6, the election of Mr F Robertson. The B shares voted 100% in favour of all three.
Remgro reported every resolution at both meetings as approved by the requisite majority. On InsidEntity’s calculation, each of the three elections was carried with roughly three quarters of the combined votes cast.
The B shares do not vote on everything. At both AGMs, all 39,056,987 B shares abstained on Special Resolution 2, the general authority to repurchase shares. The announcements record the abstention but do not say why. What the register does show is the direction such an authority points: shares bought back, whether cancelled or held in treasury, stop voting, so a repurchase of ordinary shares would raise the B shares’ share of the votes.
That is why “43%” should not be read as 43% of every decision. Three separate numbers are at work:
- Entitlement. The B shares carry 43.0% of the total votes.
- Votes cast. Their share of the votes actually cast depends on turnout and on whether they vote at all: 51.4% on the 2025 director election, none on the repurchase authority.
- Effect on the outcome. That depends on the type of resolution, the turnout, and how the ordinary votes split.
The ratings
Remgro’s Company Risk Rating is 4.57, in the Excellent band, and its Financial Stability Rating is 4.3.
| CRR pillar | Score |
|---|---|
| Director Independence | 5.00 |
| Director Capacity | 3.27 |
| Auditor Independence | 5.00 |
| Shareholder Influence | 5.00 |
| Company Risk Rating | 4.57 |
The CRR is the average of the four pillar scores: (5.00 + 3.27 + 5.00 + 5.00) ÷ 4 = 4.57. Director Independence is carried at 5.00 as a placeholder until Remgro’s directors complete InsidEntity’s independence questionnaires; it is not an assessment of the board.
Auditor Independence reads 5.00 because EY has audited Remgro since the financial year ended 30 June 2024; Remgro’s FY2025 financial statements record it as auditor for two years, so FY2026 is its third, in the one-to-three-year band. Until 25 September 2026 the platform still recorded PwC, Remgro’s auditor for 55 years before EY, which scored the pillar at 0.00 and the CRR at 3.32. The record has been corrected.
Shareholder Influence is its own number behind the number. The pillar counts shareholders and tests whether any holder reaches 20%. On holdings, nobody does: the PIC holds 16.04% of the ordinary shares and Rupert Beleggings about 7% of all shares. So the pillar scores 5.00, the Benchmark band.
The 43% of the votes attached to the B shares does not enter it. The 5.00 therefore reflects the pillar’s defined, holding-based test. It should not be read as a conclusion that voting influence at Remgro is dispersed. The pillar is doing what it says; what it does not measure is the gap this whole piece is about. On InsidEntity’s own governance rating, the size of a holding stands in for the size of a voice, and at Remgro the two are not the same.
The InsidEntity takeaway
The headline number is true, but incomplete. 16%, 7% and 43% are all accurate, and they describe different things.
On Remgro’s register, the size of a holding and the size of a voice are separate measurements. The largest disclosed ordinary shareholder holds 9.4% of the total votes, against 43.0% attached to the B-share block, although that block has less than half as many shares. Every shareholder receives the same dividend per share. The difference between the two classes is the voting power attached to each share.
That difference is not only arithmetic. In two consecutive years, director elections have passed on the strength of the B shares’ votes despite a majority of the ordinary-share votes cast against them.
The number to watch next is 75.6%. That is the ordinary-share turnout required, on the 30 June 2025 capital structure and assuming all B shares vote, before the B shares stop being a majority of the votes cast. At the 2024 and 2025 AGMs, ordinary turnout was about 70% of the voteable ordinary shares. The 2026 register will move the threshold slightly.
The next test is the AGM on 30 November 2026.
Independent research. Not financial advice. No allegation of wrongdoing is made against any individual or entity named.
Know more. Risk less. Decide better.
Sources: Remgro shareholder information: major shareholders, shares in issue, treasury shares, directors’ interests and votes per share, at 30 June 2025 and 2024; Remgro at a glance, 2020, for the B shares’ 42.62% of total votes at 30 June 2020; Remgro report on proceedings at the 2024 annual general meeting, JSE SENS, 29 November 2024; Remgro report on proceedings at the 2025 annual general meeting, JSE SENS, 27 November 2025; Remgro Audit and Risk Committee Reports 2022 and 2023, for PwC’s tenure and the appointment of EY from the year ending 30 June 2024; Remgro annual financial statements 2025, independent auditor’s report, dated 22 September 2025, stating that Ernst & Young Inc. “has been the auditor of Remgro Limited for two years”; Remgro annual financial statements 2026; Remgro AGM details for 2026; Remgro summary of audited results for the year ended 30 June 2026, JSE SENS; Remgro SENS announcement of Ninety One’s holding, 17 August 2026; InsidEntity Remgro report.
Note on figures: Ratings are as read from the platform on 25 September 2026. Register figures are at 30 June 2025, the last published register date; the 2026 register will be published with the AGM documents at the end of October 2026. Percentages of shares and votes are InsidEntity’s arithmetic on those figures, at ten votes per B share and excluding treasury shares, which carry no vote. The Ninety One share count is derived from its 5.0156% of the issued ordinary shares. The combined election results are derived from the ordinary-share voting split reported in each AGM announcement and the B-share count.
Related reading
One Response